View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Helsingin yliopiston digitaalinen arkisto

Synthesis of lignano-9,9'-lactones and rearrangement studies

Barbara Raffaelli

University of Helsinki Faculty of Science Department of Chemistry Laboratory of Organic Chemistry Finland

Academic dissertation

To be presented, with the permission of the Faculty of Science of the University of Helsinki, for public criticism in Auditorium A110 of the Department of Chemistry, A.I. Virtasen aukio 1, on November 3rd, 2012, at 12 o’clock noon.

Helsinki 2012 Supervisor Professor Kristiina Wähälä Laboratory of Organic Chemistry Department of Chemistry University of Helsinki Finland

Reviewers Professor Emeritus Erkki Kolehmainen Department of Chemistry University of Jyväskylä Finland

Professor George W. Francis Department of Chemistry University of Bergen Norway

Opponent Professor Scott A. Snyder Department of Chemistry Columbia University NY, USA

ISBN 978-952-10-8280-1 (paperback) ISBN 978-952-10-8281-8 (PDF)

Helsinki 2012 Unigrafia Oy

2 Abstract

Lignans are found as naturally occurring compounds in plants, and also as metabolites in mammals. Due to their biological activity, lignans have attracted the interest of scientists from different areas like food chemistry, synthetic chemistry, and clinical chemistry. The research is very active, as evident by the number of publications related to this subject that are published annually.

The present work focuses on the chemistry of certain classes of lignan, namely lignano-9,9'- lactones and their rearranged products. The literature review is based on the synthesis of lignano-9,9'-lactones and 7'-OH-lignano-9,9'-lactones. Both racemic and asymmetric protocols appeared as from year 2000 have been extensively reviewed. In addition, the literature concerning the rearrangement reaction of 7'-OH-lignano-9,9'-lactones has been discussed.

The experimental section presents the development of a stereoselective synthesis to obtain 7'-hydroxylignanolactones. Subsequently, the known rearrangement reactions affecting hydroxylignanolactones are discussed. The stereochemistry of certain lignanolactones is a core part of this work and the findings of this study allowed the revision of certain controversial data found in the literature. For the first time, the X-ray structures of 7'-OH- lignano-9,9'-lactones were obtained. In addition, the synthesis of possible lignan metabolites is also presented. A summary of the available 1H NMR data for selected (7'S)- and (7'R)-OH- lignano-9,9'-lactones and for various lignano-9,7'-lactone stereoisomers has been made available in the Appendix section.

3 Acknowledgements

This study was carried out in the Laboratory of Organic Chemistry, Department of Chemistry, University of Helsinki.

I am greatly indebted to my supervisor, Professor Kristiina Wähälä, for introducing me to the challenging field of lignans and for believing in this project more than I often did. Her support, criticism, and encouragement have never waned during these years.

I wish to express my deepest gratitude to Professor Emeritus Tapio Hase for the stimulating discussions and precious advice. As we were sharing an office for many years, he was always near-at-hand for guidance through the countless problems encountered in this project.

Thanks to Professor Christine Cardin and Dr. Yu Gan for the flawless work done on my crystals.

I am grateful to the reviewers of my thesis, Professor Emeritus Erkki Kolehmainen and Professor George W. Francis, and to the language editor, Dr. Jennifer Rowland, for their careful and most useful comments.

I wish to thank Professor Scott A. Snyder for agreeing to act as my opponent at the public defence of this dissertation.

Several friends and colleagues deserve recognition for contributing to this thesis. I want to thank Monika for the invaluable help in the laboratory but particularly at arm's length. Her collaboration has also been crucial while writing the articles and the thesis. Thanks to Paula for her criticism and suggestions, for lending me the desilylation protocol, and also for visiting me in Belgium; to Gudrun, for her priceless assistance in the laboratory and the long chats; to my co-authors Eija, Clément, and Antti; to Sampo, for helping me with the X-ray figures. Thanks also to Sari (office- and life-experiences-mate), Emilia, Tuija, Somdatta, Ullastiina, Kirsti, and all the other present and past members or sympathizers of the ‘Phyto-Syn’ group, who definitely made my life more interesting, both in and out of the laboratory. A very special thank to Taina for being such a good friend and trusting me beyond the misunderstandings.

Many thanks to the staff of the Laboratory of Organic Chemistry, in particular to Dr. Jorma Matikainen, Seppo Kaltia, and Dr. Sami Heikkinen for their help with the MS and NMR spectroscopy; to Dr. Jorma Koskimies and Dr. Leena Kaisalo for their assistance in various chemistry-related matters; and Matti Keinänen for building the best ever TLC-chamber lids.

4 Many thanks to my dearest friends of the ‘Rompighiaccio’ group for being there, and to my friends and colleagues in Belgium. I feel lucky to have met you all in these years.

Financial support from the Graduate School of Bioorganic Chemistry, the Research Foundation of the University of Helsinki, the Orion-Farmos Research Foundation, the Chancellor of the University of Helsinki, the Finnish Chemical Society, the Kemian Päivät Foundation, the Magnus Ehnroth Foundation, and the Emil Aaltonen Foundation is gratefully acknowledged.

Vorrei ringraziare i miei genitori Silvana e Giovanni, per avermi dato sempre il loro supporto e la sicurezza di sapere che: “quando hai bisogno chiama!” Grazie a tutti gli altri componenti della mia famiglia in Italia e in particolare alla mia cara sorella Alice. Tutti voi sapete il bene che vi voglio e quanto siete importanti per me.

Kiitos myös mieheni perheelle, erityisesti Marjalle, joka on avuliaasti hoitanut lapsiamme.

The person I need to thank the most is my husband Mikko. Without him this thesis would have never been realized, and I mean it literally. His endless patience, support, and encouragement have substantially contributed to the work. He is always reminding me to look on the bright side and, at the same time, to keep my feet on the ground. There are not enough words to express my love and gratitude, also for taking care of our kids.

Finally, the most special thanks to my daughters Linda and Lisa. I am proud of you for being so patient and accepting to give up some of the time I could have spent playing with you. You are the blessing of my life.

Vantaa, October 2012

Barbara Raffaelli

5

6

“All truths are easy to understand once they are discovered; the point is to discover them.”

Galileo Galilei

A Linda e Lisa, le mie pucciole

7 List of original publications

I. Barbara Raffaelli, Kristiina Wähälä and Tapio Hase. Asymmetric synthesis, stereochemistry and rearrangement reactions of naturally occurring 7'-hydroxylignano-9,9'-lactones. Org. Bio. Chem. 2006, 4, 331–341.

II. Barbara Raffaelli, Monika Pohjoispää, Tapio Hase, Christine J. Cardin, Yu Gan and Kristiina Wähälä. Stereochemistry and rearrangement reactions of hydroxylignanolactones. Org. Bio. Chem. 2008, 6, 2619–2627.

III. Barbara Raffaelli, Yu Gan, Christine Cardin, Monika Pohjoispää and Kristiina Wähälä. Crystal structures of two hydroxylignanolactone regioisomers. Manuscript (to be submitted) 2012.

IV. Barbara Raffaelli, Eija Leppälä, Clément Chappuis and Kristiina Wähälä. New potential mammalian lignan metabolites of environmental phytoestrogens. Env. Chem. Lett. 2006, 4, 1–9.

V. Barbara Raffaelli, Antti Hoikkala, Eija Leppälä and Kristiina Wähälä. Enterolignans. J. Chromatogr. B 2002, 777, 29–43.

The published articles are reproduced in the printed version of this thesis with the kind permission of the publishers.

8 Abbreviations

Ac ...... acetyl hr ...... hour(s) acac ...... acetylacetonate IUPAC ...... International Union of Pure ACCN ...... azobis(cyclohexanecarbonitri- and Applied Chemistry le) LC ...... liquid chromatography AIBN ...... 2,2'-azobisisobutyronitrile LDA ...... lithium diisopropylamide Ar ...... aryl LHMDS ...... lithium hexamethyldisilazide ATC ...... atom transfer carbonylation LN ...... lithium naphthalenide BBN ...... borabicyclo[3.3.1]nonane Me ...... methyl BINOL ...... 1,1-binaphthol MOM ...... methoxymethyl Bn ...... benzyl MS ...... mass spectrometry Bu, n-Bu ...... normal butyl NADH ...... reduced nicotinamide adenine t-Bu ...... tert butyl dinucleotide BV ...... Baeyer-Villiger NADPH ...... reduced nicotinamide adenine BVMO ...... Baeyer-Villiger monooxygena- dinucleotide phosphate se NMM ...... N-methylmorpholine CHMO ...... cyclohexanone NMO ...... N-methylmorpholine-N-oxide monooxygenase NMR ...... nuclear magnetic resonance CHP ...... cumene hydroperoxyde NOESY ...... nuclear overhauser effect CPMO ...... cyclopentanone spectroscopy monooxygenase PCC ...... pyridinium chlorochromate COSY ...... correlation spectroscopy PDC ...... pyridinium dichromate DABCO ...... 1,4-diazabicyclo[2.2.2]octane PFA ...... paraformaldehyde dba ...... dibenzylideneacetone Ph ...... phenyl DBU ...... 1,8-diazabicyclo[5.4.0]undec- Piv ...... pivaloyl 7-ene PMHS ...... polymethylhydrosiloxane de ...... diastereoselective excess PPTS ...... pyridinium p-toluenesulfonate DEAD...... diethyl azodicarboxylate Pr ...... propyl DG ...... diglucoside PTDH ...... phosphite dehydrogenase DIBAL ...... diisobutylaluminum hydride psi ...... pound per square inch DMAP ...... 4-N,N-dimethylaminopyridine py ...... pyridine DME ...... 1,2-dimethoxyethane rt ...... room temperature DMF ...... N,N-dimethylformamide SIM ...... selected ion monitoring DMI 1,3-Dimethyl-2-imidazolidino- TBACl ...... tetra-n-butylammonium ne chloride DMPU ...... 1,3-dimethyl-3,4,5,6- TBAF...... tetra-n-butylammonium tetrahydro-2(1H)-pyrimidinone fluoride DMSO ...... dimethylsulfoxide TBS ...... tert-butylsilyl d.r...... diastereomeric ratio TBDMS ...... tert-butyldimethylsilyl ee ...... enantiomeric excess TBDPS ...... tert-butyldiphenylsilyl E. coli ...... Escherichia coli TFA ...... trifluoroacetic acid EI ...... electron impact THF ...... tetrahydrofuran eq...... equivalent TIPS ...... triisopropylsilyl Et ...... ethyl TLC ...... thin layer chromatography GC ...... gas chromatography TMS ...... tetramethylsilane HPLC ...... high performance liquid chro- TPAP ...... tetrapropylammonium matography perruthenate HMDS ...... hexamethyldisilazane Tf ...... trifyl HMPA ...... hexamethylphosphoramide Tr ...... trityl HAPMO ...... hydroxyacetophenone mono- Ts ...... tosyl oxygenase p-TSA, TSA ...... p -toluenesulfonic acid

9 Contents Abstract ...... 3 Acknowledgements ...... 4 List of original publications ...... 8 Abbreviations ...... 9 Summary ...... 11 1 Introduction to lignans ...... 12 1.1 Classification and nomenclature of lignans ...... 14 1.1.1 Classification ...... 14 1.1.2 Numbering ...... 15 1.1.3 Nomenclature ...... 16 1.1.4 Presentation and description of stereochemistry ...... 17 1.2 Ecological significance and biological activity ...... 18 1.2.1 Mammalian lignans (enterolignans) ...... 19 2 Stereochemistry assignment ...... 23 3 Syntheses of lignano-9,9'-lactones: The state of the art as from the year 2000 ...... 29 3.1 Racemic syntheses ...... 31 3.1.1 Catalytic mechanisms ...... 31 3.1.2 Radical mechanisms ...... 34 3.1.3 Miscellaneous mechanisms ...... 37 3.2 Stereoselective syntheses ...... 38 3.2.1 Catalytic mechanisms ...... 39 3.2.2 Use of chiral auxiliaries ...... 45 3.2.3 Radical mechanisms ...... 48 3.2.4 Enzymatic transformations ...... 52 3.2.5 Miscellaneous mechanisms ...... 56 3.3 Syntheses of 7'-OH- or 7'-OR-lignano-9,9'-lactones ...... 58 3.3.1 Catalytic mechanisms ...... 58 3.3.2 Radical mechanisms ...... 61 3.3.3 Use of chiral auxiliaries ...... 63 3.3.4 Miscellaneous mechanisms ...... 67 4 Rearrangement reactions of hydroxylated lignanolactones under basic conditions ..... 70 5 Aim of the study ...... 74 5.1 Stereoselective synthesis of 7'-hydroxylignano-9,9'-lactones ...... 75 5.1.1 The construction of the (8R,8'R)-lignano-9,9'-lactone framework ...... 75 5.1.2 The removal of the dithiane moiety and the reduction with L-Selectride ...... 76 5.2 Rearrangement and stereochemistry studies ...... 81 5.2.1 Translactonization of 18O-labeled substrates ...... 82 5.2.2 Translactonization studies with NaH/DMF and KOH/MeOH ...... 84 5.2.3 Translactonization studies with aq. NaOH/EtOH ...... 86 5.2.4 Translactonization studies of (±)-8-methylparabenzlactone ...... 87 5.2.5 Translactonization studies of Ar-O-silylated (7'S)-7'-hydroxymatairesinol ..... 88 5.3 Synthesis of potential mammalian metabolites of lignans ...... 90 6 Conclusions ...... 93 7 Experimental part ...... 95 References ...... 97 Appendix ...... 109 Summary

Lignans are naturally occurring compounds, found extensively in the plant kingdom, and also as metabolites in mammals. Scientific interest in lignans includes multidisciplinary activities ranging from the field of organic synthesis, to biology and biochemistry. As a result of this, lignan-related research is vast and heterogeneous and it would be beyond the scope of the thesis to exhaustively cover all of the topics. This work essentially focuses on the chemistry of specific lignans (selected lignanolactones). In particular the work for the development of an asymmetric synthesis to obtain 7'-hydroxylignanolactones is presented. In tandem, the rearrangement reactions affecting hydroxylignanolactones are discussed. Thanks to the results of this work, certain controversial data found in the literature could be corrected. Investigation into the stereochemistry of certain lignanolactones is also a core part of this work. Finally, a project for the synthesis of possible lignan metabolites will be also presented. To fully understand the meaning of this research, this thesis begins with an introduction to lignans, where the classification, the main significance, and the stereochemistry of lignanolactones are presented. This part will be followed by a review of the methods for the synthesis of lignanolactones developed as from year 2000. Both racemic and asymmetric syntheses will be discussed, showing how lignans continue to be in the spotlight, in the chemistry scene, many years after their discovery. This literature review will conclude with a section devoted to the rearrangement reactions of hydroxylignanolactones.

11 1 Introduction to lignans

Lignans represent a large family of natural products that are widely distributed in the plant kingdom. This name (originally spelled lignane) was chosen by Haworth1 in 1936 to describe compounds possessing a skeleton structure derived from two phenylpropanoid

(n-propylbenzene, C6C3) units connected at the central carbons of the side chains by a so- called β-β' linkage (Fig. 1).

β β β' 2X

Fig. 1. The basic structure of lignans according to Haworth’s definition.1

Several naturally occurring compounds containing phenylpropanoid groups exist that are related to lignans, as their names also lead us to understand. Compounds in which the two 2 C6C3 units are connected via a C-C bond different from β-β' are called neolignans. Alternatively, when the two units are linked by an ether oxygen atom, they are identified with the name oxyneolignan. Structures with more than two C6C3 monomers are oligomeric lignans like sesquilignans/sesquineolignans (three units), and dilignans/dineolignans (four units). When a lignan is attached to other types of compounds (e.g. terpenes, flavonoids, stilbenes) it is referred to as a hybrid lignan.3 The term norlignans refers to products that structurally resemble lignans, but carry one less carbon atom in the aliphatic chain. Finally, recent studies have shown that in some species, certain lignans (like secoisolariciresinol glucoside) are part of a larger and complex structure, referred to as a lignan macromolecule. This macromolecule also includes hydroxymethyl glutaric acid, p-coumaric acid glucoside 4 and ferulic acid glucoside. Examples of different types of C6C3 derivatives are shown in Fig. 2.

12 OH HO OH O OCH3 OCH OMe 3 OMe O O a neolignan O OH HO OH OH OH HO Me a sesquilignan/sesquineolignan MeO O MeO OMe OH MeO O an oxyneolignan OH O O OMe O MeO OH O OH H OH OH HO a dilignan/dineolignan O O OH

Me OMe

MeO OH Me H Me H O a norlignan O

O O O O aditerpeno-lignan hybrid

Fig. 2. Examples of C6C3 unit-containing structures related to lignans.

Lignans are secondary metabolites and have been found in all parts of the plants: wood, root, rhizome, stem, leaf, fruit and seed.5 A recently published study on the phylogenetic distribution of lignans in plants suggested that wide distribution seems to be linked to plant evolution.6 The number of naturally occurring lignans is large, due to the assorted moieties and functionalities present on the aliphatic7 and aromatic8 parts of the molecule. Lignan-O- glycosides (mono- and oligosaccharides) have been isolated with D-glucose as the principal, but not the exclusive, carbohydrate. In this context the extraction technique plays a crucial role, since glycosides are cleaved under hydrolytic conditions (acid catalysis or enzymes), therefore only the aglycone is finally isolated. The sugar moiety is more often found in the

13 aromatic portion of the lignan, although structures where sugars are appended to the aliphatic part have also been identified.9,10 Most lignans are optically active compounds, with both enantiomers being isolated at times, depending on the producing plant.11

1.1 Classification and nomenclature of lignans

1.1.1 Classification

The first effort to classify lignans dates back to Haworth12 who used the difference in reaction products between nitric acid and the ether functionality present in the molecule as a discrimination criterion. He divided lignans into: a) diarylbutanes, b) tetrahydronaphthalenes, and c) tetrahydrofurans. This initial arrangement was expanded by Hearon and MacGregor13 who, according to the oxidation level of the aliphatic portion, classified the lignans as: a) 1,4- diarylbutanes, b) 2,3-dibenzylbutyrolactones, c) tetrahydrofurans, d) tetrahydrofurofurans, and e) 4-aryltetrahydronaphthalenes. This classification was enlarged and modified in 1978 by Weinges et al.14 who separate lignans and cyclolignans (a term which replaces the previously used misleading name isolignan to indicate those compounds possessing an extra C-C bond).15 In Weinges’ classification, still on the basis of the oxidative state of the side chains, lignans were described as derivatives of: a) butane, b) butanolide, c) tetrahydrofuran, and d) 3,7-dioxabicyclo[3.3.0]octane; while cyclolignans as derivatives of: a) 1,2,3,4- tetrahydronaphthalene, b) hydroxymethyl-1,2,3,4-tetrahydronaphthalene carboxylic acid lactone, c) 3-hydroxymethylnaphthalene-2-carboxylic acid lactone, and d) 2- hydroxymethylnaphthalene-3-carboxylic acid lactone. A third class, named isolignans (derivatives of benzofuran), was introduced, but this subgroup should be included among neolignans because it does not possess a β-β' bond. Surprisingly, this detailed classification omitted the large subgroup of dibenzocyclooctadienes, identified already in the early 1960s,16 and however included later in Whiting’s review.17 Essentially, the classification commonly used today is based on Hearon and MacGregor’s proposal, integrated with Weinges’ and Whiting’s suggestions, as represented in Fig. 3.

14 O O O

O O

a) Dibenzylbutanes b) Dibenzylbutyrolactones c) Tetrahydrofurans

O O O O

d) Tetrahydrofurofurans e) Aryltetralins f) Arylnaphthalenes g) Dibenzocyclooctadienes

Fig. 3. Generally accepted and used classification of lignans.

Concerning the aromatic part, lignans possessing variously substituted rings have been isolated from natural sources. Due to their biosynthesis (dimerization of coniferyl alcohol),18,19 all plant-originated lignans are para-substituted. Aromatic substituents are oxygenated moieties such as hydroxy and methoxy.

1.1.2 Numbering

Freudenberg and Weinges15 were the first to attempt to rationalize the numbering systems, assigning the numbers 1-9 to the left half of the molecule (drawn following particular directives), according to the notation of the benzene carbon atoms (C-1 being the ipso carbon). For the right half of the molecule, the numbering is repeated using primed numbers (1'-9'). The β-β' linkage is therefore indicated as 8-8'. This system was unclear, however, as it depended on the way the lignan was drawn. To reduce ambiguities, Ayres and Loike20 proposed to use unprimed numbers for that half of the molecule which contained functionalities with priority according to IUPAC (carbonyl has priority over alcohol) (Fig. 4, structure A). In case of equivalent aliphatic chains, then the aromatic part must be taken into account (Fig. 4, structure B).

15 O O O O 9' 9 OMe O 9 9' OMe 3' 2' 2 3 3 2' 3' 8' 8 2 8 8' O OMe O OMe 4' 1' 7' 7 1 4 4 1 7 7' 1' 4' 5' 6' 6 5 5 6 6' 5' A B

Fig. 4. Lignan numbering according to Ayres and Loike.20

Regrettably, different kinds of customized numbering systems (Fig. 5) have been used over the years, generating misunderstanding and errors (especially when applied to the naming of the molecules).

2' 5 6 2' 7' 4 1' 4 4 1' 4 3' 3 3 3' 3 3 O O O O 4' 6' 4' 6' 6 2 1 5 2 1 7'' 2 1 2 5' 5' 1 1'' O O O O 6'' 2''

5'' 3'' 4'' A B C D

Fig. 5. Examples of different numbering systems that can be found in the literature.21,22,23,24

In the year 2000, IUPAC published the recommendations for the nomenclature of lignans and neolignans,25 also including the rules for a standardized numbering system. In the case of dibenzylbutyrolactones the numbering should be as described in Fig. 4 (structure A) (“lower locants required to identify the principal functional group expressed as a suffix”).

1.1.3 Nomenclature

A variety of trivial names and nomenclature systems have been adopted over the years causing mistakes and uncertainty. For example, hinokinin and cubebinolide were thought to be enantiomers,26 but finally they were found to be the same molecule.27 Additionally, more ambiguity is caused by the utilization of very similar names for different compounds (e.g. cubebinolide and cubebininolide),28 or by the use of the same trivial name for the glycoside and the aglycone. The existence of different numbering systems and classifications has also caused complications in the nomenclature. Again, the IUPAC recommendations25 assist by defining how to name a lignan. In the case of lactones the rule specifies that “lactones are named by changing the ending -ic acid of the corresponding acid to -lactone preceded by the

16 locant of the acid group and then the locant of the hydroxy group which forms the lactone in that order”. For dibenzylbutyrolactones, where the only functionality is the lactone, the IUPAC name is therefore lignano-9,9'-lactones. Nevertheless, since trivial names are still commonly used in the literature, in this work lignans will also be referred to by their trivial name, when this does not lead to misunderstanding.

1.1.4 Presentation and description of stereochemistry

Most lignans are optically active compounds, with stereogenic centers present even in the simplest structure (for instance at position 8 and 8'). This was known already in the early stage of lignan research and, in fact, the trivial names were used with the descriptors l and d.12 Also the (+) and (–) signs were used. These signs may be used to indicate the relative enantiomers, when the absolute configuration is known. Otherwise they can be misleading, as the sign of rotation of polarized light can change in different solvents. A more general and systematic description of the stereochemistry was proposed by Freudenberg and Weinges15 who introduced the designations α and β to indicate respectively the pointing up and down of hydrogen atoms from the plane of the paper. Since this notation required the molecule to be drawn according to specific rules, the use of the unequivocal Cahn-Ingold-Prelog R/S system was later suggested to avoid errors.20 Surprisingly, the IUPAC recommendations recovered the use of the descriptors α and β, providing the directions on how to orientate a lignan molecule as follows: “a) unprimed numbers on the left; b) unprimed numbers 1–6 are placed in the lower left corner; c) primed numbers 1'–6' are placed in the lower right corner. Stereochemistry associated with a ring system which includes position 8 is shown by means of α (indicating below the plane) or β (indicating above the plane) for a bridgehead hydrogen (or substituent) or for a substituent attached to the ring system. If there are two substituents at one position, preference is given to: a) the substituent that includes a skeletal atom; b) the group preferred by the sequence rules. If α or β are not applicable, then R or S should be used.”25 For the lignan shown in Fig. 6, drawn according to IUPAC recommendations, the complete name should therefore be (8α,8'β)-lignano-9,9'-lactone, which corresponds to the 8R,8'R configuration. In this thesis (and in the articles), the unequivocal and easily recognisable R/S system was preferred.

17 O O H H α β Ar Ar

Fig. 6. The structure of a lignano-9,9'-lactone, drawn according to the IUPAC recommendations, showing the stereochemistry descriptors.25

1.2 Ecological significance and biological activity

The role of lignans in plants is not yet completely clear. At the ecological level they have, together with other phenylpropanoid compounds,29 a role in the plant defense.19,30 Among others, antifungal,31 insecticidal,32 and feeding deterrent33 properties have also been reported for several lignans. In addition, plant lignans exhibit a large spectrum of biological activities both in vitro and in vivo. A comprehensive review about the bioactivity of lignans was published in 200234 and integrated a few years later by an update.35 As such, anticancer,36 antiproliferative and apoptotic effects,37 antioxidant,38 antiviral,39 and anti-inflammatory activities40 are all noted examples of the positive bioeffects of lignano-9,9'-lactones.41 The use of certain lignan-rich plants or derivatives as folk medicine for the treatment of a range of disorders has been known for centuries. It is worth mentioning that the lignan podophyllotoxin 1 was the lead compound for the semisynthetic drugs Etoposide® 2, Teniposide® 3, and Etopophos® 4, widely used against certain types of tumors (neoplasms). These molecules are shown in Fig. 7.

OH HO O R1 R2 R2, R3=H, R1= O O O O O Etoposide, 2 O O OH S HO O R2=H, R3=OH, R1= MeO OMe O OR3 O O Teniposide, 3 R1=H, R2=OH, R3=Me Podophyllotoxin, 1 OH HO O 2 3 1 R =H, R =PO3H2, R = O O O Etopophos, 4

Fig. 7. Podophyllotoxin 1 and derived drugs.

18 1.2.1 Mammalian lignans (enterolignans)

Certain plant lignans possess a very important feature: when ingested they can be metabolized by the intestinal microflora and transformed into the so-called mammalian lignans (or enterolignans), which from a structural point of view differ from plant lignans in that they are not necessarily para-substituted. The mammalian lignans that were first recognized, and are thus most studied, are enterolactone 5 and enterodiol 6 (Fig. 8). Enterolactone was the first enterolignan to be detected in mammalian fluids (like plasma and urine) at the beginning of the 1980s.42,43,44,45 In recent years, new mammalian lignans have been found and identified or tentatively identified (for example 7'-OH-enterolactone 7 and enterofuran 8, Fig. 8).46,47,48,49 They may include other plant lignan metabolites, as well as metabolic derivatives of enterolactone and enterodiol.50 Investigations in this respect have often been delayed due to the lack of reference materials and of sensitive analytical methodologies. Moreover, a major problem is that enterolignans occur mainly as glucuronides,51 which cannot be analysed directly by most available techniques. As the samples need to be pretreated (acid or enzymatic hydrolysis, extraction, purification, and derivatisation) this may cause product loss or artefacts. While enterolactone and enterodiol have been the objects of several studies, the biological effects of the minor metabolites have not yet been studied.

HO HO OH O OH

O

OH OH

Enterolactone, 5 Enterodiol, 6

OH HO HO O O

O

OH OH

7'-OH-Enterolactone, 7 Enterofuran, 8

Fig. 8. Examples of mammalian lignans (enterolignans).

19 The biological effects of mammalian lignans 5 and 6 have been under examination for years.52 Increasing evidence suggests they may have protective action against diverse health disorders such as prostate, breast, bowel, and other cancers, cardiovascular disease, brain function disorders, menopausal symptoms, and osteoporosis.53 The biological activity of enterolignans is estrogen receptor (ER) mediated, as their structures resemble that of the endogenous hormone 17-β-estradiol. For this reason, mammalian lignans are also called phytoestrogens, i.e. non steroidal naturally occurring compounds that can bind to the ER and produce estrogenic or/and antiestrogenic effects. They can also be regarded as endocrine disruptors, as they interfere with mammalian hormone function. For quite some time, only the plant lignans matairesinol 9 and secoisolariciresinol 10 (Fig. 9), or their precursor glycosides, were assumed to be metabolized into mammalian lignans, but this hypothesis was confuted when several other plants lignans (e.g. pinoresinol 11, and lariciresinol 12, Scheme 1) were found to be (at least in vitro) even more efficient mammalian lignan precursors.48,49,54 Good sources of plant lignans are cereals, beverages (like coffee, tea, and beer), fruit, berries, vegetables, and flax seed.

MeO MeO OH O OH HO HO O

OMe OMe OH OH Matairesinol, 9 Secoisolariciresinol, 10

Fig. 9. First recognised plant precursors of mammalian lignans (only aglycones are shown).

The metabolism of plant lignans has been extensively studied and is under continuous investigation. For example, the path from secoisolariciresinol 10 to enterolactone 5 has been well established and all the enzymes involved in this process have been identified.55,56 The steroisomerism of the mammalian lignans, a very important consideration, and its role in the metabolism and in the biological activity of the enterolignans, has not been thoroughly studied until now. The stereochemical properties of plant lignans have been already anticipated in chapter 1.1.4, but what about mammalian lignans? Both enterolactone 5 and enterodiol 6 exist in two forms: the mirror images (8R,8'R)-(–) and the (8S,8'S)-(+) enantiomers. Studies show that the stereochemistry of the lignan metabolites depends on the configuration of the precursors.57 It appears that the metabolic reactions (e.g. hydrolysis, demethylation, and dehydroxylation), which are involved in the conversion of plant lignans

20 into mammalian lignans, do not affect the configuration of the C8-C8' junction. Thus, for example (7S,7'S,8S,8'S)-(+)-pinoresinol diglucoside 13§ is metabolized into (8R,8'R)-(–)- enterolactone (–)-5,57 while the (8S,8'S) enantiomer (+)-5 is formed from (8S,8'S)- secoisolariciresinol diglucoside 14 (Scheme 1).58,59 The bacteria that are responsible for these transformations are capable, for instance, of stereospecifically transforming only (–)-enterodiol into (–)-enterolactone, or (+)-enterodiol into (+)-enterolactone.60,61,62

OMe OMe H OR RO OH OH O HO OH R'' H H H H H

O O RO HO OR' MeO MeO R''' (7S,7'S,8S,8'S)-Pinoresinol (7S,8R,8'R)-Lariciresinol, 12 R=Me diglucoside, 13 R=H R=Glu, 13 R'=Me R=H, 11 R'=H R''=OH R''=H R'''=OH R'''=H, (-)-6

H H H MeO HO OGlu HO OGlu O O HO H H O H O

OMe OH OH OH (8S,8'S)-Secoisolariciresinol (8S,8'S)-(+)-Enterolactone, (8R,8'R)-(-)-Enterolactone, diglucoside, 14 (+)-5 (-)-5

Scheme 1. Examples of stereocontrolled metabolism of plant lignans into enterolignans.57,58,59

The biological effects of individual mammalian lignan enantiomers have not yet been investigated. It is recognised that the chirality of the metabolites depends on the stereochemistry of their plant lignan precursors.54 Studies have shown that 7'-hydroxymatairesinol 15, precursor of (–)-5, has anticancer effects54,63 and that the action is

§ Note that the absolute configuration 8S,8'S of pinoresinol corresponds to the configuration 8R,8'R of enterolactone.

21 likely to be (–)-enterolactone-mediated. On the other hand, flax seed,64 that contains mainly (8S,8'S)-secoisolariciresinol diglucoside 14, precursor of (+)-5,65 is also recognised to possess several positive biological effects.66 Therefore if both enantiomers of enterolactone are biologically active, as the recent research seems to confirm, the means by which the enterolignans accomplish their biological activity may need to be reconsidered, and a mechanism other than receptor binding could eventually be implied.54 More studies on this topic are required.

22 2 Stereochemistry assignment

As already mentioned in chapter 1.1.4, most lignans are chiral compounds and the determination of their absolute configuration has always attracted the interest of researchers since their discovery.12 In the past, the configuration assignment was done through interconversion, chemical transformation, and derivatization of the molecules in order to find a common stereochemical pattern among various lignans, and to connect them with stereochemically known compounds.13,67,68 The first lignan whose absolute configuration was determined is (–)-guaiaretic acid, a lignan with a single chiral center. Two independent research groups established that (–)-guaiaretic acid possesses the R configuration by correlating the lignan guaiaretic acid dimethylether 16 with (S)-methylbutanol 1767 or alternatively with (S)-3,4-dihydroxyphenylalanine 1868 (Scheme 2). Since then (R)-guaiaretic acid has become the model compound for assigning the configuration of chiral lignans.

H H H MeO MeO MeO R

MeO MeO O MeO O Cl

OMe H OMe MeO (R)-dimethyl guaiaretic acid, 16 R NHTs MeO

H H HO CO2H HO S S NH HO 2 (S)-methylbutanol, 17 (S)-3,4-dihydroxyphenylalanine, 18 Scheme 2. Conversion of (R)-dimethyl guaiaretic acid 16 into stereochemically known molecules.68,67

In naturally occurring lignano-9,9'-lactones (with no substituents at the aliphatic chains) the benzyl groups almost exclusively show an 8,8'-trans relationship. Such an arrangement was already deduced in the late 1930s,69 and, was definitely proven in 1950 when the optically active (–)-matairesinol dimethyl ether (–)-19 was converted into the optically active diol (–)-20 70 by LiAlH4 (Scheme 3). A cis relationship between the two benzylic groups of the lactone would have, in fact, furnished the optically inactive meso-form of diol.

23 H H MeO MeO OH 8' O 8 LiAlH4 OH MeO MeO H O H

OMe OMe OMe OMe (-)-19 (-)-20

Scheme 3. Reduction of the lactone (–)-19 to an optically active diol (–)-20 as evidence that the benzylic groups of (–)-19 cannot possess a cis relationship.70

Correlation between 1H and 13C NMR can be used to determine the relative stereochemistry of lignano-9,9'-lactones, as this type of molecules present a spectrum with common pattern. Once the relative stereochemistry at 8 and 8' has been assigned, for example with the help of NOESY correlation, then the absolute configuration may be assigned, knowing the sign of the optical rotation, as for the 8,8'-trans lignano lactone a constancy in the sign of the optical rotation can be observed in most commonly used solvents. Thus, negative values have been consistently recorded for the 8R,8'R configuration, while the 8S,8'S isomers give positive values.20,57,71,72,73,74,75,76 This assumption can also be applied to the less common cis configuration, whereas (8R,8'S)-lignanolactones have a positive sign, while the stereoisomer has a negative.77,78 The absolute configuration of lignanolactones has been proven in some cases by X-ray diffraction.79,80,81 Furthermore, a combined conformational study of 8,8'-trans- lignano-9,9'-lactones has been recently reported.82 When additional stereogenic centers are present, the issues increase in complexity. The configuration of the 7-OH-lignano-9,9'-lactone podorhizol 21 (Scheme 4) was already determined in 1967.83 The acid hydrolysis of the naturally occurring podorhizol-O-glucoside 22 furnished, among other products, podorhizol 21 and desoxypodophyllotoxin (7'S)-23 and isodesoxypodophyllotoxin (7'R)-23 (Scheme 4). The configurations of (7'S)-23 and (7'R)-23 were already known as 7'S,8R,8'R and 7'R,8R,8'R** respectively.70 Anhydropodorhizol 24 (Scheme 4), the possible intermediate in the case of α-carbonyl epimerization, was not found. Therefore it was deduced that 22, and consequently podorhizol 21 itself, must

** Note that according to IUPAC recommendations for aryltetralignano-9,9'-lactones the numbering is different from that of lignano-9,9'-lactones.

24 possess at 8 and 8' the same trans stereochemistry as (7'S)-23 and (7'R)-23, this is to say S and R respectively.††

H H O O 8' 8 O O + O 7 H3O O GluO H O O

MeO OMe MeO OMe OMe OMe 22 24

+ H3O

H H H O O O 8 8 8' O 8' O 8 O 8' + 7' + O 7' O O 7 H H H O O HO O

MeO OMe MeO OMe MeO OMe OMe OMe OMe (7'S)-23 (7'R)-23 (-)-podorhizol, 21 (7S,8S,8'R)

Scheme 4. The transformation of podorhizol-O-glucoside 22 after acid treatment: the way how the configuration of podorhizol 21 was deduced.83

The C-7 configuration of podorhizol was determined by IR spectroscopy as only the 7R epimer possesses the right geometry to form a strong intramolecular hydrogen bond between the 7-OH and the lactone carbonyl. In this way, naturally occurring podorhizol with IR peaks at 3625 cm-1, 3490 cm-1 (disappearing with strong sample dilution), and 1762 cm-1, was established to be the 7S,8S,8'R isomer. On the other hand, the 7R epimer (named epi- podorhizol) with peaks at 3610 cm-1, 3500 cm-1 (which did not disappear upon sample dilution), and 1750 cm-1, showed the presence of an intramolecular hydrogen bond. NMR data also supported these conclusions. For podorhizol, the signal of H-7 at 5.28 ppm is in a deshielded area, while for epi-podorhizol the H-bond forces the H-7 in the shielding zone of

†† Note that the S configuration at 8' in the 7-OH-lignano-9,9'-lactones corresponds to the R configuration in aryltetralignano-9,9'-lactones and in 7'-OH-lignano-9,9'-lactones. Therefore in case of podorhizol the 8S/8'R is the trans configuration.

25 the lactone carbonyl, giving a signal for the H-7 at 4.83 ppm. These outcomes were later supported by studies on the aldol condensation, a relatively easily synthetic pathway to obtain these types of lignans, and by further NMR investigations.84,85,86 In fact, taking into account that both epimers could form an intramolecular hydrogen bond, the protons H-7 and H-8 in the erythro compound (thus the same steric arrangement as in podorhizol) are in an equatorial-axial relationship (Fig. 10), with coupling constants expected to be between 2 and 4 Hz. In the threo configuration (that of epi-podorhizol), with a trans-diaxial relationship between H-7 and H-8 (Fig. 10), the coupling constants are expected to be between 6 and 9 3 3 Hz. Indeed, the H-7 signal of podorhizol has a J7,8 = 2.0 Hz, while the epimer a J7,8 = 7.5 Hz.

Ar H H Ar 8 8 7 7 H H OO OO H H erythro threo

Fig. 10. The erythro and threo aldol products and their intramolecular hydrogen bonding.

The 7'-OH-lignano-9,9'-lactones have raised more problems. The 8R,8'R arrangement of 7'-hydroxymatairesinol 15 was determined converting this lignan to the stereochemically known (–)-α-conidendrin (the corresponding aryltetralin lignan) but the stereochemistry at 7' was not determined.87 The 8R,8'R configuration of the natural compound (–)-parabenzlactone 25 (Fig. 11) was determined through the conversion of 25 into the corresponding dehydroxylated (–)-hinokinin.88 The stereochemistry at 7' was determined by correlating the 1H NMR signal of the secondary benzylic proton of 25 with that of podorhizol 21 (hence H-7' vs. H-7). As shown in our recent revision (see paper I), the configuration assigned to (–)-parabenzlactone turned out to be wrong. Based on this erroneous assignment all the following studies referring to that assumption were flawed. However, for several years the issue was forgotten and the absolute configuration of the two 7' epimers of 7'-hydroxymatairesinol 15 was only investigated again in the late 1990s. Despite the in-depth spectroscopic studies,89,90 no conclusive statement could be made, even if apparently all data were supporting the original (incorrect) claim derived from the comparison between 25 and 21. The question was finally clarified by Eklund and co-workers,91 who thoroughly studied the natural 7'-hydroxymatairesinol 15 (mainly as mixture of the two epimers), obtained in large amounts from pine knots. By treating the mixture of the 7'-hydroxymatairesinol epimers with

LiAlH4, thus reducing the lactone moiety without affecting the stereocenters, they obtained a

26 mixture of 7-hydroxysecoisolariciresinol epimers 26 (Scheme 5). The major epimer was obtained in a crystalline form suitable for X-ray analyses and the X-ray study revealed that it possesses the S configuration at the benzylic position. As a consequence, the major 7'- hydroxymatairesinol possesses the 7'S configuration, thus opposite as previously assigned.89,92,93

OH OH OH H H H MeO MeO MeO [H-] OH OH 7' O 7 recrystal. 7 OH OH HO HO HO H O H H

OMe OMe OMe OH OH OH 15 26 (7S)-26 (major isomer)

Scheme 5. Reduction of 7'-hydroxymatairesinol 15 into a crystalline triol. The absolute configuration of isomer (7S)-26 was confirmed by X-ray analysis.91

These findings were supported by the first total stereoselective synthesis of 7'-OH-lignano- 9,9'-lactones, which appeared only in 2004.94 Fischer and co-workers prepared both 7' isomers of 7'-hydroxymatairesinol and also of the naturally occurring 7'-hydroxyarctigenin 27, demonstrating that also the latter was reported in literature with the wrong configuration, the true one being 7'R. As we will discuss later, the work of this thesis confirms the findings of Eklund and Fischer, providing unambiguous evidence with the X-ray analysis of 7'-OH- lignano-9,9'-lactones (the first X-ray for this type of lignans). In addition, the configuration of the other known naturally occurring 7'-OH-lignano-9,9'-lactone (–)-parabenzlactone 25 was also rectified as 7'S.

27 HO HO H H MeO MeO 7' 7' O O HO HO H O H O

OMe OMe OH OH (7'S)-7'-hydroxymatairesinol, (7'S)-15 (7'R)-7'-hydroxymatairesinol, (7'R)-15 HO HO H H MeO O 7' 7' O O O MeO H H O O

O OMe O OH (7'S)-parabenzlactone, (7'S)-25 (7'R)-7'-hydroxyarctigenin, (7'R)-27

Fig. 11. Naturally occurring 7'-OH-lignano-9,9'-lactones (correct absolute configurations).

As previously observed for non hydroxylated lignano-9,9'-lactones, NMR data correlations can be used to determine the relative configuration, also in the cases of 7 and 7'-OH derivatives. For example for 7'-OH-lignanolactones, once the relative stereochemistry at 8 and 8' has been assigned by a NOESY experiment, it is then possible to compare the signal for the proton at 7'. It has been observed that in CDCl3, the 7'-H for the S isomer (7S relative to a 8R,8'R configuration) occurs at about δ 4.6 (see Annex 2, Table 7) and for the R isomer (7R relative to a 8R,8'R configuration) at about δ 4.4 (see Annex 2, Table 8).94 With the addition of other spectroscopic data (i.e. knowing the sign of the optical rotation) it is possible to correlate homochiral lignans and thus to determine the absolute configuration of 7'-OH- lignano-9,9'-lactones.

28 3 Syntheses of lignano-9,9'-lactones: The state of the art as from the year 2000

Speaking in very general terms, as this is true also for lignans, the science of synthesis is continuously engaged towards new methodologies for accomplishing the desired targets in the most efficient and effective way. As natural compounds, lignans have represented an interesting target for synthetic chemists. The variety of lignan structures, from simple to more complex arrangements, and the importance as biologically active compounds have engaged many scientists in developing new and improved strategies to obtain such molecules. But why to synthesize a lignan, when such molecule can be found in nature? First of all, with the exception of a few examples like 7'-hydroxymatairesinol 15,95 the amount of material that can be obtained from a plant source may be extremely small, sufficient to determine the structure of the molecule, but not enough for studying the biological activities, or to be employed in other studies. Therefore the availability of an adequate amount of material is crucial in many fields. Furthermore, even if the nature provides us with an amazing number of different structural combinations, one can always wish to have a different pattern, which for instance is not naturally available (or at least not yet found). So when the target is a new molecule, this can be synthesised with well established synthetic methods. On the other hand, the synthesis of lignans can be the means for the development of new synthetic tools and procedures, applicable later to a broader range of molecules. The synthetic importance of lignano-9,9'-lactones immediately stands out by observing Scheme 6: these structures represent a springboard for direct or indirect access to many other lignan types. In addition, lignano-9,9'-lactones are needed for a number of purposes, such as analytical and metabolic studies, biological use, quantification in various matrices, studies on the biosynthesis and on their role in nature.

29 Ar' Ar' O O O O O Y O Ar

Ar' X X

Ar' Ar' OH O O Y Y OH O Y Ar O Ar

X X X Ar' O O Y O O Ar Ar Ar O Scheme 6. Lignano-9,9'-lactones give direct (normal arrows) or indirect (dashed arrows) access to many different lignan structures.

Several synthetic approaches towards lignano-9,9'-lactones have appeared over the years and, as a result, a number of reviews (usually including the synthesis of various classes of lignans) have been published.96,97,98,99,100,101,102,103,104,105,106 The research in this field has been very active and many new methodologies for the synthesis of the lignano-9,9'-lactone framework (or possible precursors) have been developed very recently and will be discussed in this section. The review will cover the published literature, starting from the year 2000 (with few exceptions), and will also include the syntheses of β-benzyl-γ-butyrolactones,107 ideal precursors for lignano-9,9'-lactones, via alkylation or, when a 7-OH is desired, via aldol condensation (Scheme 7). The synthesis of β-benzyl-γ-butyrolactones can therefore be considered to be a formal synthesis of 8,8'-trans-lignano-9,9'-lactones, with the second benzyl moiety (at C-8) found on the opposite face of the lactone with respect to the benzyl at C-8' (1,2 asymmetric induction).108 For that reason, those routes with a β-benzyl-γ- butyrolactone as an intermediate (but not specifically developed to obtain the lignan framework) have also been examined.

30 X X X H H Ar 7' Base Ar Base Ar 7' O O O HO 7 Ar'CH2CHO Ar'CH2Br 7 H H Ar O O Ar O

Scheme 7. Procedure for the synthesis of dibenzylbutyrolactones starting from β-benzyl-γ- butyrolactones. X=H, OR.

The developed protocols for the synthesis of lignano-9,9'-lactones have been collected according to the final product(s). Syntheses that produce racemates are discussed in chapter 3.1; syntheses that allow the formation of optically active compounds are discussed in chapter 3.2. The last chapter of this section (chapter 3.3), includes the routes that allow the formation of an oxygenated function at C-7' in the lignano-9,9'-lactones or at the benzylic position in case of β-benzyl-γ-butyrolactones. In each chapter the synthetic methods are grouped according to the mechanism of the key step. The aromatic substituents are indicated as described in Annex 1. Differently oxygenated (e.g. at C-7, C-8, or C-8'), unsaturated (e.g. C-7=C-8), or substituted (e.g. C-9') lignano-9,9'-lactones or β-benzyl-γ- butyrolactones have not been reviewed.

3.1 Racemic syntheses

3.1.1 Catalytic mechanisms

A palladium-catalyzed three-component coupling reaction (Scheme 8) was the starting point of Balme’s approach109,110 to β-benzyl-γ-butyrolactones. The mechanism of this step is a sequential hetero-Michael conjugated addition-carbopalladation-reductive elimination process, mediated by a palladium(0) catalyst obtained in situ by the reduction of

PdCl2(AsPPH3)2 with NaH. The obtained five membered heterocycle 28 was first hydrogenated to 29, and then converted into the lactone using Yb(OTf)3 as Lewis acid. The lactonization step involves a Lewis activation of the acetal with ring cleavage and formation of a stabilized malonic enolate and an oxonium species captured by methanol or water. The resulting alcohol can attack one of the ester groups to give the lactone. Finally, a one pot alkylation/decarbalkoxidation made the lignano-9,9'-lactones available as an 8,8'-trans/cis diastereomeric mixture (85:15).

31 Ar Pd I MeO C a Ar 2 O ArI + + 68-80% O MeO2C OMe MeO2C MeO2C MeO2C OMe OH MeO C OMe 2 28

b 99% H H H Ar Ar d Ar c Ar O + O O O 87-94% 89-94% MeO C MeO2C H H 2 H MeO C Ar' O Ar' O O 2 OMe 85 : 15 29 (±) (±)

Scheme 8. Reagents and conditions: (a) NaH, PdCl2(AsPPH3)2 (2 mol%) THF, DMSO, 0 → 30 °C, 3 hr; (b) H2, Pd/C, AcOEt, 10 bars, overnight; (c) Yb(OTf)3 (10 mol%), MeOH, reflux; (d) K2CO3, 5 16 14 Ar'CH2Br, DMF, LiCl, H2O, 130 °C, 4 hr. Ar=Ar , Ar ; Ar'= Ar (for Ar see Appendix 1).

The Heck reaction was used to construct the lignan skeleton from a silanol and a disiloxane, which were obtained in quantitative yield from silane 30 (Scheme 9).111,112 Diester 31 was obtained in a good yield and with a very high regio- and stereocontrol. Intermediate 31 was then hydrolyzed to acid 32, which was converted into a lignanolactone via two alternative routes. In one route (step d, Scheme 9), 32 was esterified with phenylselenomethyl chloride and the product underwent a radical cyclization providing a mixture of trans/cis lactones (the trans being the major isomer, as expected). In the other route (step f, Scheme 9) 32 was first reduced to an alcohol and then converted to a seleno ester which was in turn submitted to the previously seen radical cyclization. The products obtained from the two routes have reversed aromatic substituents. The stereomeric trans- and cis-lactones could be separated by fractional crystallization. By mean of this procedure, the authors accomplished the synthesis of (±)-matairesinol 9 from the trans isomer.112

32 Ar' Ar Ar Ar CO2Et a EtO C b EtO2C EtO2C + 2 100% CO2Et EtO C SiPh EtO C SiOH EtO2C Si O 75-84% 2 2 Ar 30 2 31

c 88-95%

H H PhSe Ar' d Ar' Ar' Ar' e O O + O 74-92% 78-87% CO2H O H O H O Ar Ar Ar Ar (±) 78 : 22 (±) 32

f 96%

H H e g Ar Ar Ar O Ar OH O + O 90% O 91% H H PhSe Ar' O Ar' O Ar' Ar' (±)78 : 22 (±)

Scheme 9. Reagents and conditions: (a) TFA, CH2Cl2, 0 °C, 1.5 hr, then aq. NH3; (b) TBAF, TBACl, DMF, Ar'I, Et3N, [Pd(allyl)Cl]2 (5 mmol%) 80 °C, 40 hr; (c) KOH, EtOH, rt, 4 hr; (d) i-Pr2NEt, PhSeCH2Cl, NaI, Me2O, 80 °C, 18 hr; (e) Bu3SnH, AIBN, PhH, 80 °C, 4 hr; (f) LiAlH4, THF, reflux, 4 hr; 10 16 (g) 1. Triphosgene, py, CH2Cl2, -15 → 10 °C, 2 min; 2. Et3N, PhSeH, 0 °C → rt, 30 min. Ar= Ar , Ar ; Ar'=Ar1, Ar10 (for Ar see Appendix 1).

Racemic dibenzylbutyrolactones were prepared by a short protocol where all the steps were metal catalyzed (Scheme 10).113 After a syn addition,114 the resulting E-vinylstannane 33 was oxidized using NMO and TPAP and two regioisomers were obtained. The bulky oxidizing reagent allowed a certain regioselectivity and, in fact, the desired 4-substituted lactone 34 was obtained as the major product. The inseparable mixture of lactones was submitted to the Stille coupling reaction. Only 34 reacted with the benzyl bromide to furnish the monobenzyl butenolide 35. Hydrogenation and consecutive alkylation afforded (±)-trans-lignanolactones. A stereoselective version of this synthesis was also developed and is discussed later in chapter 3.2.1.

33 OH Bu Sn Bu3Sn a 3 OH b

O + O 92% OH 47% Bu3Sn O OH 33 34 O 25 : 1 c 45-70% H Ar e Ar d Ar O O O 18-35% 70-97% H O O O Ar 35 (±)

Scheme 10. Reagents and conditions: (a) Bu3SnH, Pd(PPh3)4; (b) NMO, TPAP (7.5 mol%), -78 → -23 °C, 62 hr; (c) ArCH2Br, Pd2(dba)3·CHCl3, AsPh3, THF, rt → reflux, 5 hr; (d) H2, Raney–Ni (T4); (e) 1 3 9 LHMDS, DMPU, ArCH2Br, THF, -78 °C → rt. Ar=Ar , Ar , Ar (for Ar see Appendix 1).

3.1.2 Radical mechanisms

Srikrishna and Danieldoss115 based their synthetic approach on the radical cyclization of a bromoacetal 36, obtained in 3 steps from the benzaldehyde via the Horner-Wadsworth- Emmons reaction, followed by a reduction and finally by a bromoacetalization reaction

(Scheme 11). The cyclization was performed using n-Bu3SnH generated in situ. The 3:2 mixture of ethoxy epimers was successively oxidised to give the desired lactones.

a R c Ar O ArCHO Ar 65-84% Br R=CO2Et OEt b 36 R=CH2OH d 64-80%

Ar e Ar O O 65-81% O OEt (±)

Scheme 11. Reagents and conditions: (a) (EtO)2P(O)CH2CO2Et, NaH, THF; (b) LiAH4, Et2O, <0 °C, 1hr; (c) EtOCH=CH2, NBS, CH2Cl2, -70 °C, 2h; (d) n-Bu3SnCl, NaCNBH3, AIBN, t-BuOH, reflux, 90 min; (e) Jones reagent, acetone, ultrasonication, rt, 5 min. Ar=Ar5, Ar9, Ar14, Ar18 (for Ar see Appendix 1).

Organotin reactions often give low yields, mainly due to the purification step, which can be a difficult task. For this reason a solid-phase synthesis was employed as a solution for this problem.116 Thus, a bromoacetal appended to a solid support (Merrifield resin) was used as a

34 substrate for a radical cyclization to give a product easy to isolate by simple filtration (Scheme 12). The last step of the synthesis, a Jones oxidation, served both to liberate the acetal from the solid support and to oxidize it to a lactone, without further purification. The yields, however, were not astonishing and the reaction required several days to be completed.

HO O a HO O OH + Cl O 82% b 84% c O O O Br O O 92% O O Ar d

Ar O O e Ar O O O 67% (2 steps) (±) O

Scheme 12. Reagents and conditions: (a) NaH, THF, reflux, 20 hr; (b) NaH, Merrifield resin, THF, reflux, 3 days; (c) NBS, ArCH=CHCH2OH, 0 °C, 6 days; (d) Bu3SnH, AIBN, PhH, reflux, 18 hr; (e) Jones reagent, acetone, rt, 3 hr. Ar=Ar1 (for Ar see Appendix 1).

A feasible tin-free radical protocol for the synthesis of β-benzyl-γ-butyrolactones was recently published (Scheme 13).117 The key intermediate is the sulfide 37, obtained in 3 steps from 5- chloro-1-pentyne. The crucial step involves the addition of the benzenesulfanyl radical, generated in situ from PhSH and ACCN, to the triple bond of 37 to form a vinyl radical, which, in turn, generates the desired alkyl radical by homolytic substitution at the sulfur atom. The alkyl radical instead of being trapped by hydrogen abstraction, undergoes a 5-exo radical cyclization with ring enclosure onto the electron-rich C-C double bond affording the monosubstituted lactone.

COR Ar O a cd S Ar Cl O O 99% 80% 75% R=OH S b (2 steps) R=Cl O 37 (±)

Scheme 13. Reagents and conditions: (a) KOH, SHCH2CO2H, MeOH, 0 → 40-50 °C, 2 hr; (b) SOCl2, DMF, CH2Cl2, rt, overnight; (c) PhCH=CHCH2OH, Et3N, CH2Cl2, rt, 2 hr; (d) PhSH, ACCN, PhMe, reflux, 1 hr. Ar=Ar1 (for Ar see Appendix 1).

35 The oxidative coupling is a well known approach for the synthesis of lignans. A drawback of this method is the formation of a number of side products due to the random coupling of phenoxyl radical mesomers. The major side product is usually the β-5 coupling product, which involves the aromatic ring (C-5). To avoid this inconvenience, Zhu et al.118 blocked the C-5 with a t-Bu group detached at the end of the synthetic sequence (Scheme 14). The key coupling step, obtained by oxidation of the 5-substituted ferulic acid 38, furnished the cis fused dilactone 39, whereof relative cis-stereochemistry was determined by NMR (NOESY experiments) and also by the stereochemistry of the final product (±)-trans-matairesinol 9.

R' CO2H O O Ar c d H H 81% 91% MeO R Ar Ar O O OH 38 39 a R=H 83% R=t-Bu e 93% b R'=CH3 85% R'=CHO H H CO2H Ar f Ar O 85% CO2H H H Ar O Ar g Ar=Ar21 83% Ar=Ar7, (±)-9

Scheme 14. Reagents and conditions: (a) t-BuOH, 85% H3PO4, 75 °C; (b) Br2, t-BuOH, rt; (c) malonate, py, piperidine, 100 °C; (d) O2, FeCl3, MeOH-H2O, rt; (e) 10% Pd/C, H2, rt; (f) DCC, THF, 21→7 then NaBH4, THF, rt; (g) AlCl3, PhH, 50 °C. Ar=Ar (for Ar see Appendix 1).

The novelty of a recently published multistep sequence is the decyanation by lithium naphthalenide (LN) reduction of the 8-CN lignan 40 (Scheme 15).119 The substrate for this reaction was obtained by submitting the cyano ester 41 to a one-pot intramolecular Knoevenagel condensation-hydrogenation. This reaction furnished a pair of inseparable diastereomers with a ratio depending on Ar. The diastereomeric mixture was alkylated to give only cyano dibenzylbutyrolactones with the benzylic groups in a trans configuration (confirmed by X-ray crystallography). Compound 41 was synthesised in 5 steps starting from an aldehyde. First, a Corey-Fuchs reaction, quenched with paraformaldehyde (PFA), then a butanethiol addition to the vinyl alcohol, followed by acidic hydrolysis and esterification with cyanoacetic acid. LN reduction served to remove the CN group in the last step of the synthesis, for on overall yield of 19-25%. The authors reported that the asymmetric version of the synthesis is under investigation.

36 a Br b c ArCHO Ar OH Ar OH 80% 92% 82% Br Ar SBu d 72% H g f e Ar Ar Ar O Ar OH O O 80% 82% O 92% O NC O R O Ar' O CN h R=CN, 40 41 95% R=H, (±)

Scheme 15. Reagent and conditions: (a) CBr4, PPh3, CH2Cl2, 0 °C, 2 hr; (b) n-BuLi, THF, -78 °C, PFA, rt, 3 hr; (c) BuSH, NaOH, MeCN, 75 °C, 2 hr; (d) EtOH/1N H2SO4 (4:1), 50 °C, 24 hr; (e) Me2NPOCl2, py, CNCH2COOH, CH2Cl2, 0 °C → rt, 24 hr; (f) L-Proline, Hantzsch ester, EtOH, rt, 24 hr; (g) Ar'CH2Br, K2CO3, THF, rt, 24 hr; (h) LN (3.5 eq.), THF, -45 °C, 30 min; then NH4Cl (aq.), -45 °C → rt. Ar=Ar1, Ar15, Ar18; Ar'=Ar1, Ar14, Ar18 (for Ar see Appendix 1).

The last strategy presented in this section is illustrated in Scheme 16.120 The regioselectively prepared phenylselanylfuran-2(5H)-one 42 was submitted to a tandem Michael addition- alkylation reaction, affording a mixture of diastereomers, where the major isomer is the one with the benzylic groups in a trans configuration, due to steric reasons. The Michael addition required the use of DMPU as co-solvent. The reductive elimination of the phenylselanyl group was performed on the diastereomeric mixture and, through a radical mechanism, produced mainly the more stable trans-lignanolactones.

OMe H H H a b O Ar Ar c Ar O O + O O 90% 45-70% Ar Ar 94% PhSe PhSe PhSe PhSe H OMe O O O Ar O 42 d.r. up to 10 : 1 (±)

Scheme 16. Reagents and conditions: (a) 1M HCl, MeCN, 0 °C, 10 min; (b) Ar2CuLi or ArMgBr/CuI, 1 5 Et2O, -78 °C, then DMPU, Ar'CH2Br, rt, 5 hr; (c) NiCl2, NaBH4, THF, 0 °C, 5 min. Ar=Ar , Ar (for Ar see Appendix 1).

3.1.3 Miscellaneous mechanisms

A Friedel-Crafts alkylation is the starting point of Taneja’s approach.121 After the ester formation and the reduction of ketone 43, the alcohol was formylated using the Vilsmeier reagent to give 44. An additional reduction furnished the primary alcohol converted into the final butyrolactone either via acidic hydrolysis followed by double bond reduction or, alternatively, by reduction followed by acidic hydrolysis. The asymmetric version of this reaction will be presented in section 3.2.4.

37 O OH a b ArH Ar Ar 76-80% 93-99% 2 steps CO CH CO2CH3 43 2 3 c 51-60%

CH OH CHO Ar e Ar 2 d Ar O 77-86% 85% CO2CH3 CO2CH3 O 44 b 86% b 88%

CH OH Ar e Ar 2 O 70% CO2CH3 (±) O

Scheme 17. Reagents and conditions: (a) 1. AlCl3, PhNO3, succinic anhydride, 0 → 60 °C, 3 hr; 2. CH2N2, Et2O; (b) H2, Pd/C (5 mmol%), EtOH, 50 psi; (c) DMF, POCl3, 0 → 40 °C, 36 hr; (d) NaBH4, MeOH, 0 °C; (e) 10% HCl, EtOH, 70 °C. Ar=Ar9, Ar16 (for Ar see Appendix 1).

The Baeyer-Villiger (BV) oxidation represents an awesome reaction to access β-benzyl-γ- butyrolactones. Relatively easily available 3-monosubstituted cyclobutanones 45 can be oxidized to lactones using peroxyacids or hydrogen peroxide (as oxygen donor) and different catalysts. The solvent is crucial in this kind of reaction, as it must solubilize oxygen, and organic substrates, but must not be oxidable. Thus chlorinated solvents are usually selected. Bolm and co-workers122 developed a new technique where liquid or supercritical carbon dioxide was used as medium. Moreover, if CO2 is used in the presence of a combination of molecular oxygen and an aldehyde (pivaldehyde in this specific case) as co-reductant, the BV reaction does not require a catalyst to proceed (Scheme 18).

Ar Ar a O 75% (conv.) O O 45 (±)

1 Scheme 18. Reagent and conditions: (a) O2, compressed CO2, pivaldehyde, rt, 18 hr. Ar=Ar (for Ar see Appendix 1).

3.2 Stereoselective syntheses

Regardless the fact that racemic syntheses are still under continuous development (as seen in section 3.1), stereoselective syntheses nowadays represent the main objective in this field.

38 The need of optically active pure compounds, for example in the investigation of the chemical structure and behaviour of lignans, is evident, but even more obvious is the need for biological screening. It is well known that the biological activity of organic molecules is structure-dependent and, especially when looking for a possible new drug or lead compound, it is essential to use a single isomer. Thus, novel and (more) efficient methods to achieve pure enantiomers have been explored and developed in the last years.

3.2.1 Catalytic mechanisms

The asymmetric Baeyer-Villiger (BV) reaction is the stereoselective version of the reaction described in section 3.1.3 and is relatively recent, as the first chiral metal-catalyzed BV oxidation was reported only in 1994. The asymmetric BV reaction in lignan synthesis permits the access to optically active β-benzyl-γ-butyrolactones by means of desymmetrization of prochiral cyclobutanones. Bolm’s group investigated several catalytic systems with modifications affecting the metal (Al, Mg),123,124 the chiral chelating ligand (axially chiral binaphthol and derivatives),125,126 and also the oxygen donor (Scheme 19).127 The outcome (R or S) of the reaction depends on the chiral chelating ligand. Their best results (ee 73%) were obtained with the system BINOL-Al (with BINOL 46) as catalyst, and cumene hydroperoxide (CHP) as source of oxygen. For the first time in this type or reaction, a catalyst could be used in a substoichiometric amount (50 123 mol%), and furnished modest to good ee. The replacement of Al with Mg (from MgI2) had a slightly negative effect on the enantioselectivity.124 In a later study, it was found that using the vaulted biaryl VANOL 47 the catalyst load could be reduced down to 20 mol%.125 The total conversion of the substrate was still observed, with higher yield, but unfortunately the enantioselectivity dropped to lower values. Using the acetylene BINOL-derivative 48 as ligand the catalyst load was decreased to 10 mol%, but the yield and absolute configurations of the products were not reported.126

39 H H Ar a Ar Ar R O or S O 87-100*% O O O ee 34-73% ee 55-64%

OH OH OH OH

(S)-BINOL, (S)-46 (R)-BINOL, (R)-46

Me3Si

OH Ph OH OH Ph OH

Me3Si 48

(R)-VANOL, (R)-47

Scheme 19. Reagent and conditions: (a) Me2AlCl + (S)-46 (1:1, 50 mol%), CHP, PhMe, -25 °C → rt, 12 hr; or MgI2 + (R)-46 (1:1, 50 mol%), CHP, CH2Cl2, -25 °C → rt, 8 hr; or Me2AlCl + (R)-47 (1:1, 20 mol%), CHP, PhMe, -30 °C → rt, 12 hr; or Me2AlCl + 48 (1:1, 10 mol%), CHP, PhMe, -30 °C. Ar=Ar1, Ar5, Ar9, Ar16 (for Ar see Appendix 1). *Conversion of the starting material determined by GC.

Ding's group also explored the BV reaction to obtain chiral lactones.128 Their best results were obtained using a chiral phosphoric acid derived from the H8-binol-derived 49 and aqueous H2O2 as oxidant (Scheme 20). In this system, the solvent also played an important role in catalytic efficiency and asymmetric induction, and CHCl3 was proven to be the solvent of choice. The originality of this approach is due to the chiral Brønsted acid, used for the first time as an asymmetric BV oxidation catalyst.

40 X H

Ar a Ar S O O O P 99% O OH O ee 55-58% O X X=Pyren-1-yl, 49

1 9 16 Scheme 20. Reagent and conditions: (a) H2O2, 49 (10 mol%), CHCl3, -40 °C. Ar=Ar , Ar , Ar (for Ar see Appendix 1).

A variation on the motif was recently provided by Malkov et al.129 (Scheme 21) who used urea-H2O2 as stoichiometric oxidant, and the complex obtained in situ by reacting

(PhCN)2PdCl2 , ligand 50 (derived from α-pinene), and AgSbF6 as catalyst. The reaction furnished the R lactone with excellent yield but moderate ee.

H Ar a Ar R O N 91% Ph2P O ee 58% O 50

Scheme 21. Reagent and conditions: (a) (PhCN)2PdCl2 (5 mol%), 50 (5 mol%), AgSbF6 (10 mol%), 16 urea-H2O2, THF, -40 °C, overnight. Ar=Ar (for Ar see Appendix 1).

A different approach was used by Buchwald.130 The key step of the sequence is the reduction of β-substituted butenolide 51 (Scheme 22), which was synthesized in two different ways: either via ring-closing metathesis of acrylate 52 using a second-generation Grubbs catalyst 53, or via a condensation/reduction of an aldehyde.131 51 was then enantioselectively reduced to β-benzyl-γ-butyrolactones using Cu(II)/(S)-p-tol-BINAP 54 as catalyst, and polymethylhydrosiloxane (PHMS) as reducing agent. i-PrOH was used as an additive to increase the yield and the speed of the reaction.

41 Ar OR

a R=H R=COCH=CH2, 52

b 79% OH H c d Ar e CHO Ar O Ar R Ar O 79% 80-92% O (2 steps) O ee 92% O 51 O

PCy3 Cl Ph Ru P(PhCH3)2 Cl P(PhCH3)2 Mes-N N-Mes

53 (S)-p-tol-BINAP, (S)-54

Scheme 22. Reagents and conditions: (a) CH2=CHCOCl, 2,6-lutidine, THF, EtOH, rt; (b) 53 (1.5 mol%), CH2Cl2, reflux, 10 hr; (c) HCOCO2H⋅H2O, piperidine⋅HCl, dioxane-H2O, reflux, 18 hr; (d) NaBH4, MeOH, rt, 2 hr, then HCl; (e) CuCl2⋅2H2O (5 mol%), (S)-54 (5 mol%), t-BuONa (20 mol%), PMHS, pentane, rt, 2 hr; then i-PrOH, PhMe -20 °C. Ar=Ar1, Ar9 (for Ar see Appendix 1).

The method developed by Yan and Spilling132 underpinned on the enantioselective addition, promoted by a Pd(0) catalyst, of a malonate to a chiral (1R)-phosphono allylic carbonate which furnished a vinyl phosphonate further transformed into a (R)-γ-lactone (Scheme 23). The chiral (R)-phosphonate 55 was obtained by intermolecular cross-metathesis between an aryl alkene and the optically active phosphonate 56, prepared in turn by a titanium alkoxide- catalyzed phosphonylation.133 The cross-metathesis, the weak step of the route, had to be repeated four times on the same substrate to reach only a moderate yield. The enantioselective addition, instead, proceeded efficiently to give mainly the E product. The lactone ring was obtained by ozonolysis of 57, followed by in situ treatment with NaBH4. Cleavage of the t-butyl ester and decarboxylation furnished the desired (R)-lactone with complete chirality transfer.

42 MeO MeO OMe OMe Ph a Ph P b P CHO O Ar O R 4 times 71% OH OR ee 79% 36% 56 c R=H, 55 95% R=CO2Me d 85% ee 92% MeO H H H OMe f P Ar Ar e Ar R O O O 65% 57% t-BuO C ee 92% 2 t-BuO2C CO2Me O O 57

Scheme 23. Reagents and conditions: (a) Ti(i-PrO)4, dimethyl-L-tartrate, dimethyl phosphite, Et2O, -15 °C, 2 hr; (b) ArCH2CH=CH2, 53 (see Scheme 22) (5 mol%), CH2Cl2, 40 °C, 12 hr; (c) Py, ClCO2Me, DMAP, CH2Cl2, 0 °C → rt, 24 hr; (d) NaH, Pd(PPh3)4 (3 mol%), t-BuCO2CH2CO2Me, THF, 3 70 °C, 1 hr; (e) O3, MeOH, CH2Cl2, -78 °C then NaBH4, rt; (f) LiCl, DMSO, H2O, 140 °C, 17 hr. Ar=Ar (for Ar see Appendix 1).

Wennemers’ group recently developed a concise synthesis based on the asymmetric conjugated Michael-type addition of dihydrocinnamaldehyde to nitroethene (Scheme 24).134 The reation, catalyzed by peptide 58, used in a merely 1 mol%, proceeds with high yield and enantioselectivity. Transformation of the aldehyde and the nitro moieties resulted in the formation of highly optically pure β-benzyl-γ-butyrolactone, with no homoaldol reaction of the aldehyde. The authors assumed that the addition reaction proceed via enamine catalysis, because a methylation or acetylation of the N-terminal proline residue of 58 deactivates the catalyst itself. The catalyst was synthesised on a solid-phase following the general protocol for peptide synthesis.

H H a R c CHO Ar Ar Ar + S NO2 ee 98% O NO2 89% ee 97% b R=CHO O H 82% R=CH2OH N CONH2 (2 steps) N . O TFA O

NH CO2H

H-D-Pro-Pro-Glu-NH2, 58

Scheme 24. Reagents and conditions: (a) 58 (1 mol%), NMM (1 mol%), CHCl3, rt, 15 hr; (b) BH3⋅THF, 1 -15 °C, 1 hr; (c) NaNO2, AcOH, DMSO, rt, 12 hr, then MgSO4, 40 °C, 12 hr. Ar=Ar (for Ar see Appendix 1).

43 The same starting material as in Scheme 10 (chapter 3.1.1)113 was used for the asymmetric version of the synthesis. The authors’ attempt to stereoselectively reduce the benzyl butenolide 34 (Scheme 10) failed,135 and this problem was addressed by directly performing the Stille coupling on the vinylstannane 33,136 and then stereoselectively reducing the vinyldiol in the presence of an iridium catalyst (Scheme 25). For the Stille coupling optimization, several catalysts were tested, demonstrating the superiority of catalyst 59 (ee 82%), and to a lesser extent of 60 (Scheme 25). The absolute configuration of the diol was established to be R by comparing the positive value of the optical rotation with the literature data.137 The chiral diol was then submitted to TPAP-NMO oxidation which produced two regioisomer lactones in a 2:1 ratio. Chromatografically isolated (R)-β-benzyl-γ-butyrolactone was alkylated to give a trans-(8R,8'R) lignanolactone. The authors described this lactone as (+), but no value was reported. Given that (8R,8'R)-dibenzylbutyrolactones generally give negative values, while positives values are found for (R)-β-benzyl-γ-butyrolactones, there are doubts as to whether the optical rotation of the final product was actually measured, or simply extrapolated from that of the precursor.

H H R b c OH Ar OH Ar O R R O + OH 100%* OH 67-82% ee 67-82% H O a R=SnBu3, 33 O Ar 66% R=CH2Ar 2 : 1 d 1 30% (Ar=Ar )

H R CF3 Ar O P O R N B Ir H O COD Ar CF t-Bu 3 4 (8R,8'R) R=2-MePh, 59 R=Ph, 60

Scheme 25. Reagents and conditions: (a) ArCH2Br, Pd2(dba)3·CHCl3, AsPh3, THF, rt → reflux, 5 hr; (b) 59 or 60 (3 mmol%), H2, CH2Cl2, 2 hr; (c) TPAP (7.5 mol%), NMO, -78 → -23 °C, 62 hr; (d) 1 3 5 LHMDS, DMPU, ArCH2Br, THF, -78 °C → rt. Ar=Ar , Ar , Ar (for Ar see Appendix 1). *(conv.)

Isemori and Kobayashi138 established a method for the synthesis of β-benzyl-γ- butyrolactones, where the pivotal step was the stereo- and regioselective alkylation of 4-cyclopentene-1,3-diol monoacetate 61, by means of a Grignard reagent and a CuCN catalyst (10-30 mol%) (Scheme 26). The regioselectivity (1,4- vs. 1,2-regioisomer up to 94:6)

44 depends on the halogen atom in the Grignard reagent (Cl), the ratio between the Grignard reagent and the catalyst, as well as the solvent used (THF). The stereoselective reaction furnished only the isomer 62 described in Scheme 26.139 The synthesis continues with the oxidation of the alkylated pentene, the loss of a carbon atom and the formation of a lactol, further oxidized to β-benzyl-γ-butyrolactone. This approach showed general applicability, with several Ar groups tested, but in the case of Ar=Ar17, step c had to be replaced by an alternative route (steps h-j) due to the instability of the oxygenated benzyl groups to ozonolysis.

H H H AcO a Ar c Ar OH d,e Ar O 91-97% 81-84% 81% OH (2 steps) OH OR OTBS OH 61 b R=H, 62 f 93-95% ee >99% 99% R=TBS H Ar h 91% i 47% O H OH Ar OH O g 67% OTBS H Ar O

H Ar' O (8R,8'R)

Scheme 26. Reagents and conditions: (a) ArMgCl, CuCN (30 mol%), THF, -18 °C, 5 hr; (b) TBSCl; (c) O3, Py, i-PrOH, then NaBH4; (d) TBAF; (e) KIO4, EtOH-H2O; (f) PCC, NaOAc; (g) LDA, DMPU, Ar'CH2Br, THF; (h) OsO4 (cat.), NMO, MeCN-THF-H2O; (i) 1. KIO4, MeCN-THF-H2O, 35 °C, 9 hr; 2. 1 3 5 17 3 NaBH4. Ar=Ar , Ar , Ar , Ar ; Ar'=Ar (for Ar see Appendix 1).

3.2.2 Use of chiral auxiliaries

Evans and Evans-type oxazolidinone, and imidazolidinone, auxiliaries have been recently largely applied for the stereocontrolled syntheses of lignano-9,9'-lactones. The utilized chiral auxiliaries are listed in Fig. 12.

45 O O O O

Xc1= N O Xc2= NO Xc3= NO Xc4= NO

Ph Bn Bn Ph

O O O O

6 Xc5= N O Xc = NO Xc7= NNMeXc8= N NMe

i-Pr Ph Ph Bn Me

Fig. 12. Chiral auxiliaries used for the synthesis of lignans.

Sibi's group140 obtained both enantiomers of enterolactone 5 by modifying the order in which the aromatic moiety and the ester were attached to the chiral auxiliary Xc (Scheme 27). If the aromatic part was introduced first (step a), the final product is the (S,S)-lignanolactone. If the ester function was formed first (step g), then the final outcome is the (R,R) isomer. The syntheses are straightforward with an overall yield of 27% for the (S,S) enantiomer and 19% for the (R,R). The auxiliary could be recovered in more than 95% yield.

O O H H H a b d e, f Ar Ar XcH Ar Xc Ar R S O 93% 85% 74% O 54-61% (2 steps) ee 99% CO2t-Bu (2 steps) H O Ar O c R=Xc 85% R=OH (8S,8'S)

O O H H H g Xc h d Ar e,f Ar XcH Ar R R O O 85% 60% 76% 54-61% CO2Et ee 99% (2 steps) CO2Et (2 steps) H O Ar O c R=Xc 80% R=OH (8R,8'R)

Scheme 27. Reagents and conditions: (a) n-BuLi, ClCO(CH2)2Ar, THF, -78 °C, 2.5 hr; (b) NaHMDS, ICH2CO2t-Bu, THF, -78 °C, 6 hr; (c) LiOH, H2O2, THF-H2O, 0 °C → rt, 5 hr; (d) 1. BH3⋅THF, THF, 0 °C → rt, 2.5 hr; 2. p-TSA, PhH, reflux, 2 hr; (e) NaHMDS, THF, ArCH2I, -78 → -54 °C, 6-22 hr; (f) BBr3, CH2Cl2, 0 → -18 °C; (g) n-BuLi, ClCO(CH2)2CO2Et, THF, -78 °C, 2.5 hr; (h) NaHMDS, ArCH2I, THF, -78 → -54 °C, 6 hr. Xc=Xc1 (see Fig. 12); Ar=Ar3→2 (for Ar see Appendix 1).

The approach presented in Scheme 28141 is the improved version of the method developed by Charlton in 1997.142 The novelty, here, is the replacement of the t-butyl ester with allyl iodide. The use of an allyl as a masked carbonyl (see later chapter 3.3.3) prevents the

46 undesired saponification during the removal of the auxyliary. Both enantiomers (8R,8'R) and (8S,8'S) were synthesised by simply modifying the chiral auxiliary. With this protocol, (−)-5 was obtained in 20% overall yield.

O O 2 H CO H a c (Xc=Xc ) Ar 2 Ar R Ar Xc 77% 86% (3 steps) b R=OH d.r. >97:3 R=Xc 72-88% c (Xc= Xc3 or Xc4) 77% d.r. >97:3 H O H H Ar d Ar OH Ar Xc O 67 % H ee 98% Ar O e 77% (8S,8'S) (2 steps) H H Ar f Ar R O O 66% H O Ar O (8R,8'R)

Scheme 28. Reagents and conditions: (a) 1. H2, Pd/C (10%), EtOAc, rt; 2. K2CO3, BnBr, acetone- MeCN, reflux; 3. KOH, MeOH-H2O, 70 °C; (b) 1. Et3N, t-BuCOCl, THF, -70 °C → rt, 1hr; 2. n-BuLi, XcH, -70 → 0 °C, 1.5 hr; (c) KHMDS, CH2=CHCH2I, THF, -72 °C, 2 hr; (d) LiBH4, MeOH, Et2O, rt; (e) 1. OsO4, 2,6-lutidine, NaIO4, 1,4-dioxane-H2O, rt; 2. Ag2CO3, Celite, PhH, reflux, 4 hr; (f) NaHMDS, 2 3 4 3 10 ArCH2I, THF, -78 °C, 2 hr. Xc= Xc , Xc , Xc (see Fig. 12); Ar=Ar , Ar (for Ar see Appendix 1).

Evans methodology was also exploited by Pohmakotr et al.143 (Scheme 29). The chiral succinic acid derivative 63 was diastereoselectively alkylated using a conventional alkylation protocol. After removal of the chiral auxiliary Xc, the dicarboxylic acid was converted in two steps to the corresponding (R)-β-benzyl-γ-butyrolactone.

O O O H H H H O a c Xc Ar R R Ar d Ar O R O + Ar Xc 53-67% R 100% O de 99% O O O O 44% 63 b R=Xc 87-90% R=OH

Scheme 29. Reagents and conditions: (a) LDA, ArCH2Br, THF, -78 °C; (b) LiOH, THF-H2O, reflux, 7 42 hr, then HCl; (c) Ac2O, 0 °C, 10 min; (d) NaBH4, THF, 0-5 °C, 1 hr, then HCl. Xc=Xc (see Fig. 12); Ar=Ar4, Ar5, Ar9 (for Ar see Appendix 1).

47 Bennett et al.144 have developed a short, yet not optimized, synthesis of optically active (8S,8'S)-lignanolactones where the significant step was the desymmetrization of succinimide 64 by means of a chiral lithium amide base 65, capable to discriminate between a pair of enantiotopic protons. The enantioselective deprotonation generated an enolate which could be selectively alkylated (ee up to 95%) on the face anti to the bulky t-butyl substituent. The reaction continued with a second α-alkylation to obtain a trans-dibenzylderivative. Removal of the chiral auxyliary and lactonization furnished the desired lignan.

t-Bu O H O t-Bu H O t-Bu a Ar b Ar N N N 40% 44% ee 95% H O O Ar O 64 c 95%

H H Ph Ph d Ar OH t-Bu Me Me Ar O NH NLi HN 32% Ph Ph H H 65 Ar O Ar O (8S,8'S)

Scheme 30. Reagents and conditions: (a) 65, ArCH2Br, DMPU, THF, -78 °C, 3 hr; (b) LDA-LiCl, 1 5 ArCH2Br, DMPU, THF; (c) NaBH4, i-PrOH, H2O, PhMe; (d) 2N H2SO4, 80 °C. Ar=Ar (racemic), Ar (for Ar see Appendix 1).

3.2.3 Radical mechanisms

The following three approaches use chiral auxiliaries, but are included in this section because their key step involves a radical mechanism. In the method developed by Sibi et al.145 a benzylic radical was added to a fumarate, desymmetrized by a chiral oxazolidinone 66 (Scheme 31). The reaction is mediated by the Lewis acid samarium triflate which coordinates with the imide group and makes the reaction highly regio- (conjugate addition β to the imide) and stereoselective (addition on the opposite face of the bulky group of the oxazolidinone). Bu3SnH is the radical chain carrier, while 146 Et3B/O2 works as radical initiator. The Lewis acid activation is decisive for the radical addition, otherwise the reaction would be too slow and the competing hydrogenation reaction would be preferred. The lignan framework is completed with the second benzyl substituent introduced via a standard alkylation procedure. After removal of the chiral auxiliary (recovered in >95% yield) both enantiomer lactones could be easily achieved (see also Scheme 27).140

48 Xc Xc R H H a b d Ar O O Ar' O O 71-80% 50-60% 74-76% ee 99% CO2Et CO2Et CO2Et H O H H Ar' 66 Ar Ar c R=Xc (8S,8'S) 74-88% R=OH e 78-90%

H Ar' O

H Ar O (8R,8'R)

Scheme 31. Reagents and conditions: (a) Sm(OTf)3, CH2Cl2, THF, ArCH2Br, Bu3SnH, Et3B, O2, -78 °C, 3 hr; (b) NaHMDS, THF, ArCH2I, -78 → -54 °C, 26 hr; (c) LiOH, H2O2, THF-H2O, 0 °C, 3 hr; (d) LiBH4, Et2O, MeOH, rt, 24 h, then HCl; (e) 1. BH3⋅THF, THF, -15 °C, 18 hr; 2. p-TSA, PhH, reflux, 4 hr. Xc=Xc1 (see Fig. 12); Ar=Ar3, Ar9, Ar10; Ar'=Ar3, Ar10 (for Ar see Appendix 1).

When symmetrically substituted lignans are the synthetic targets, Kise’s approach147 can be used. The method is based on the oxidative homocoupling of lithium enolates of chiral oxazolidinones or imidazolidinones (Scheme 32). This sequence improves an established methodology that was already presented a few years earlier148 and pioneered by Belletire 149 and Fry. Several oxidants, including TiCl4, PhI(OAc)2, CuCl2, and I2, were tested and the best yield (81%) was obtained for both oxazolidinones and imidazolidinones with LDA-

PhI(OAc)2. Regarding their stereoselectivity, LDA-TiCl4 was observed to be the best choice for the oxazolidinones (trans:cis = 87:13), while the best results for imidazolidinones were obtained with LDA-TiCl4 and LDA-CuCl2 (trans:cis = 98:2). I2 furnished the opposite stereoselectivity ((R,R):(R,S) = 11:89). Hydrolysis of the dimers was obtained with LiOOH (for oxazolidinones) or with LiOH (for imidazolidinones). The succinic acid derivatives 67 were finally converted to anhydrides and reduced to (8R,8'R)-lignanolactones.

O O H H a Ar R c Ar Ar Xc O 38-82% R 64-76% de up to >96% H (2 steps) H O Ar O Ar (8R,8'R) b R=Xc 70-90% R=OH, 67

Scheme 32. Reagents and conditions: (a) LDA, THF, TiCl4 (or PhI(OAc)2 or CuCl2 and DMPU or I2 and DMPU), -78 °C → rt, 12-24 hr; (b) LiOH·H2O, H2O2, THF-H2O, 0 °C, 12-24 hr, then Na2SO3 4M or LiOH·H2O, THF-H2O, reflux, 24-48 hr; (c) 1. Ac2O, MeOH, -70 → 0 °C, 15 min; 2. NaBH4, THF, 0 °C, 1 hr, then HCl, rt, 1 hr. Xc=Xc3, Xc8 (see Fig. 12); Ar=Ar1, Ar3, Ar5, Ar9 (for Ar see Appendix 1).

49 Baran and DeMartino150 opened a new synthetic perspective by introducing the hetero dimerization, this is to say the coupling between the sp3-hybridized carbon atoms of two different molecules (Scheme 33). This approach, based on the intermolecular oxidative heterocoupling of two enolates, allowed the isolation of optically pure (8R,8'R)-lignanolactone in 41% overall yield. With this method unsymmetrical dibenzyllignanolactones (Ar ≠ Ar') are easily accessible. The oxidative union between LDA-enolized oxazolidinone and dihydrocinnamic ester was carried out with Cu(II)-2-ethylhexanoate (Scheme 33). The poor diastereoselectivity of the coupling (1.6:1) did not influence the outcome of the reaction because the basic conditions of the last step (DBU is a strong non-nucleophilic base) promoted the α-epimerization to give the trans-(8R,8'R) configuration.151 The brevity of the sequence is also due to the simultaneous reductive cleavage of the auxiliary and the direct formation of the secondary alcohol.

a Xc Xc Xc H OLi Ar O Ar c Ar O + d.r. = 1.6:1 O b OLi CO2t-Bu Ar' Ar' Ot-Bu Ar' Ot-Bu d

H H Ar e O Ar OH 41% (overall yield) CO t-Bu H O 2 Ar' Ar' (8R,8'R)

Scheme 33. Reagents and conditions: (a) LDA, LiCl, PhMe, -78 → 0 → -78 °C, 30 min; (b) LDA, PhMe, 30 min; (c) Cu(II) 2-ethylhexanoate, -78 °C → rt, 20 min; (d) LiBH4, MeOH, THF, -78 → -10 °C, 1.5 hr; (e) DBU, PhMe, 110 °C, 24 hr. Xc=Xc5 (see Fig. 12); Ar=Ar5, Ar'= Ar9 (for Ar see Appendix 1).

The multistep sequence by Takekawa and Shishido152,153 was centred on the fragmentation reaction of enantiopure trisubstituted cyclopropylcarbinyl radicals (Scheme 34). Both (S)- and (R)-lactone enantiomers were prepared. Intermediate 68 was obtained via a Cu-catalyzed cyclopropanation of a diazo alkene. The lactone was reduced to a diol which was desymmetrized by a lipase-mediated reaction (step d, Scheme 34), responsible for the stereoselectivity of the whole route. The regioselectivity of the C-C bond cleavage (step f, Scheme 34), is controlled by the position of the radical leaving group: the bromine 69 will lead to the S alkene, the thiocarbonyl 70 to the R isomer. The alkenes were submitted to further chemical modifications to give enantiopure (ee >99%) β-benzyl-γ-butyrolactones. The

50 chirality of the cyclopropane was entirely maintained in the radical fragmentation reaction (step f).

OR OAc H H b H Ar O O c h Ar Ar 60% 86% 100% Ar R O 68 O OH OTBDMS a R=CH2Ac d R=H 75% R=Ac 96% R=CH=N2 97% k, l ee>99% (2 steps) e 85% imid. OAc H H O i f H H S Ar OTBDMS Ar S OR 98% 92% Ar Ar CH2OH ee>99% g R=Ac 69 Br 70 OTBDMS j 60% 97% (3 steps) R=H 55% h f R=TBDMS ee>99% H 95% H H H Ar S j O Ar R Ar i, m O O Ar R OTBDMS 30% O (3 steps) O OH

Scheme 34. Reagents and conditions: (a) p-TsN3, aq. NaOH, Bu4NBr, CH2Cl2, rt; (b) CuSO4, Cu(acac)2, PhH, reflux; (c) LiAlH4, THF, rt; (d) Lipase AK, vinyl acetate, PhH, rt; (e) CBr4, Ph3P, CH2Cl2, rt; (f) Bu3SnH, AIBN, PhH, reflux; (g) KOH, aq. EtOH, rt; (h) TBDMSCl, imidazole, DMAP, DMF, rt; (i) 9-BBN, THF, 0 °C, then NaOH, H2O2, rt; (j) 1. Dess-Martin periodinane, CH2Cl2, 0 °C → rt; 2. NaClO2, NH2SO3H, t-BuOH, H2O, rt; 3. p-TSA, CH2Cl2, rt; (k) K2CO3, aq. MeOH, rt; (l) S=C(imid)2, PhMe, 50 °C; (m) 1. Dess-Martin periodinane, CH2Cl2, 0 °C → rt; 2. TBAF, THF, rt; (n) 1. PCC, 5 CH2Cl2, rt; 2. florisil, Et2O, rt, 30 min. Ar=Ar (for Ar see Appendix 1).

Ryu’s group154 implemented an atom transfer carbonylation (ATC) reaction for the formation of butyrolactones (Scheme 35). The novelty of this ATC reaction is the use of Pd(PPh3)4 as catalyst during the photoirradiation. This addition permits to increase the yield and remarkably reduce the reaction time, very lengthy with primary alkyl iodides. The authors suggested Pd(0) to operate as radical initiator by reacting with RI at the early stage, and forming a radical R• and Pd(I)I. R• and CO generate an acyl radical which forms an acylpalladium by coupling with Pd(I)I and eventually gives the lactone. The chirality of the target (R)-β-benzyl-γ-butyrolactone is attributable to the enantiomerically pure alcohol iodide 71, which was prepared in three steps following a published procedure.155

51 H H H Ar OAc a Ar OH b Ar R O 70% 74% OH 3 steps I 71 O

Scheme 35. Reagents and conditions: (a) 1. MsCl, CH2Cl2, Et3N, 0 °C; 2. NaI, acetone, 50 °C; 3. NaOMe, MeOH, rt, 1.5 hr; (b) hν (Xenon, Pyrex), Pd(PPh3)4 (5 mol%), 4-DMPA, Et3N, CO, PhH, 45 atm, 16 hr. Ar=Ar5 (for Ar see Appendix 1).

3.2.4 Enzymatic transformations

The use of chemoenzymatic methods in the stereoselective synthesis (or semi-synthesis) of lignans is relatively new (1990s). A model of this application is represented by the Baeyer- Villiger oxidation (Scheme 36), where biocatalysts have been largely used and are under continuous development. The enzymes which are capable of transforming a prochiral cyclobutanone into an optically active γ-butyrolactone are flavin-dependant monooxygenases (BVMOs). The most common enzymes are cyclohexanone monooxygenases (CHMOs), cyclopentanone monooxygenases (CPMOs) and hydroxyacetophenone monooxygenases (HAPMOs) and structural genes of these enzymes have been isolated from several microorganisms (bacteria, fungi). The BVMOs insert one oxygen atom from molecular oxygen into the substrate, while the second one is reduced to H2O by a donor. Because these enzymes are NADH or NADPH dependent, either an auxiliary enzyme that regenerates the cofactor (coenzyme regenerating enzyme) and an electron donor substrate are used together with the purified BVMOs, or a whole-cell system is employed for the reaction catalysis. Mihovilovic’s group studied several BVMOs and the outcomes are summarized in Table 1.156,157,158,159 For example, enantiomerically enriched (R)-butyrolactones were obtained using a HAPMO originating from Pseudomonas fluorescens ACB and expressed by E. coli TOP10 (Table 1, entry 1). The enantioselectivity of this enzyme may depend, according to the authors, on the distance between the bulky groups and the reactive carbonyl center. In fact a good enantioselectivity was observed with larger substituents and a spacer between the prochiral center and the phenyl group resulted in loss of chirality.156 Interesting feasibility studies have been conducted on the application of bio-engineering for the synthesis of a self-sufficient BVMO (self-sufficient because it does not require an additional catalyst for the coenzyme recycling).159 Two enzymes, the BVMO itself and a coenzyme capable of regenerating the cofactor (phosphite dehydrogenase–PTDH), were attached through a short linker peptide and expressed by E.coli TOP10. The electron donor substrate is a cheap phosphite that is converted into a phosphate. As entries 11 and 12 in Table 1 show, the results are promising because if compared to the same BVMOs, but

52 without the appended coenzyme (entries 2 and 7 in Table 1), the reaction proceeds with 100% conversion yield. In general, the BVMO technique gives good enantioselectivity with ee values up to 98-99% in several cases. Another advantage of the microbial BV compared to the metal catalyst version, is the "green chemistry" aspect of the procedure. However there are also several drawbacks, like high costs and laborious enzyme purifications steps, to mention only a few of them.

H H Ar Ar Ar BVMOs SRO or O

O O O Scheme 36. Biocatalysed Baeyer-Villiger oxidation. For the details, see Table 1.

Table 1. Baeyer-Villiger monooxygenases (BVMOs) in the desymmetrization of prochiral cyclobutanones (for Ar see Appendix 1). Entry BVMOs Ar1 Ar5 Ar3 Ar16 Ar14 Ref. Yield % 26 63 42 156 1 HAPMO (a, b) Pseudomonas fluorescens ACB ee % 44 (R) 66 (R) 56 (R) Yield % 32 35 50 55 90 157 2 CHMO (a, b) Acinetobacter ee % 88 (S) 97 (S) 99 (S) 97 (S) 90 (S) Yield % 56 35 43 89 72 157 3 CHMO (a, b) Arthrobacter ee % 93 (S) 98 (S) 93 (S) 97 (S) 94 (S) Yield % 38 60 45 64 58 157 4 CHMO (a, b) Brachymonas ee % 84 (S) 90 (S) 93 (S) 90 (S) 98 (S) Yield % 30 61 74 73 72 157 5 CHMO (a, b) Brevibacterium 1 ee % 93 (S) 75 (R) 35 (S) 26 (S) 79 (R) (a, b) Yield % 27 53 50 64 157 6 CHMO Brevibacterium 2 ee % 59 (S) 37 (S) 45 (S) 24 (R) Yield % 37 56 70 56 157 7 CPMO (a, b) Comamonas ee % 31 (S) 40 (S) 45 (S) 24 (R) Yield % 33 45 60 53 46 157 8 CHMO (a, b) Rhodococcus 1 ee % 87 (S) 98 (S) 98 (S) 80 (S) 95 (S) Yield % 31 52 72 56 55 157 9 CHMO (a, b) Rhodococcus 2 ee % 87 (S) 98 (S) 98 (S) 95 (S) 92 (S) Yield % >90c >90c >90c 158 10 CHMO (a, b) Xanthobacter sp. ZL5 ee % 88 (S) 99 (S) 88 (S) Yield % 100c 100c 159 11 PTDH-CHMO a Acinetobacter ee % 72 (S) 96 (S) Yield % 100c 100c 159 12 PTDH-CPMO a Comamonas ee % 28 (S) 41 (S) (a) Whole-cell system; (b) expressed by E.Coli; (c) conversion.

53 Enzymes have been successfully used for the desymmetrisation of racemic mixtures of alcohols, further converted into lignans. In the approach developed by Raviña et al.160 racemic hydroxyl ester 72 was desymmetrised via lipase-catalysed esterification (Scheme 37). Numerous lipases were tested and the best results were obtained using lipase from Pseudomonas cepacia (PS), which acetylated only the (R) alcohol. However, because of the unsatisfactory ee, the (R)-acetate was submitted to enantioselective lipase-catalyzed acetate hydrolysis and the optically pure alcohol was obtained. This was converted to the corresponding lactone under acidic conditions. The (S) isomer alcohol, that was not acetylated but obtained as such in 94% ee, was transformed directly into the (S)-lactone by acid treatment.

H H Ar a b O Ar OH Ar RSOAc Ar OH 81% + CO2Pr CO2Pr CO2Pr O 72 55% 40% ee 69% ee 94%

c 39% d 94% ee 96%

H H H d Ar Ar OH Ar S R O R O 94% CO2Pr O O

Scheme 37. Reagents and condition: (a) 1. KOH, MeOH, rt, 8 hr; 2. BrPr, DMF, 40 hr; (b) Lipase PS, vinyl acetate, rt, 24 hr; (c) Lipase PS, t-BuOH-phosphate buffer (1:6), rt, 18 hr; (d) p-TSA, PhH, reflux, 3 hr. Ar=Ar1 (for Ar see Appendix 1).

The work conducted by Achiwa et al.161 on desymmetrisation is shown in Scheme 38. Several lipases for the enantioselective esterification of dibenzylbutanediols were investigated. When a racemic diol was subjected to a lipase-mediated biotransformation, it furnished a separable mixture of diacetate 73, monoacetate 74 and diol 75 (Scheme 38). Diacetate 73 and monoacetate 74 were obtained either as (R,R) or (S,S) enantiomers, depending on the enzyme, in very modest ee. Diol 75, obtained in high optically purity (ee up to 99% with lipase AH), but disappointingly with a very poor yield (lower than 20%), was then oxidized with Fétizon’s reagent to the (S,S)-lignanolactone. The racemic starting material was prepared following established procedures (Stobbe condensation).

54 H Ar OH a Ar OAc Ar OAc Ar OH + + OH OAc OH OH H Ar Ar Ar Ar 73 74 75 (±)-trans 15-57% 40-65% 3-18% ee 11-64% ee 6-49% ee 12-99% (R,R or S,S) (S,S or R,R) (S,S)

b 96%

H Ar O

H Ar O (8S,8'S)

Scheme 38. Reagents and condition: (a) Lipase, vinyl acetate, rt, up to 96 hr; (b) Fétizon’s reagent 5 (AgNO3, Na2CO3, Celite), PhMe, reflux, 3 hr. Ar=Ar (for Ar see Appendix 1).

Similarly, Felluga et al.162 used a selective ester hydrolysis for the desymmetrization of nitro ester 76 (Scheme 39). Thus, racemic 76 was treated with α-chymotrypsin and the (R)-acid (ee 94%, but low yield) was then transformed into lactone.

H H a b Ar Ar NO2 Ar R NO2 + Ar NO2 78% CO2Et CO2Et CO2H CO2Et 76 21% 36% ee 94% ee 42% c

H H Ar R f Ar R O

CO2Et O d R=NO2 R=NHBoc e + R=NH3

Scheme 39. Reagents and condition: (a) MeNO2, DBU, rt, overnight; (b) α-chymotrypsin, phosphate buffer pH 7.4, H2O, rt, 15 hr; (c) TMSCl, EtOH; (d) H2, Ra-Ni, 1 atm, Boc2O, EtOH, 12 hr; (e) 2M HCl, 1 rt, 30 min; (f) 1M aq. NaNO2, H2O, 0 °C → rt, overnight. Ar=Ar (for Ar see Appendix 1).

55 The asymmetric version163 of the synthesis presented in section 3.1.3 (Scheme 17)121 was also developed (Scheme 40). Baker’s yeast was employed to simultaneously reduce the double bond (stereoselectively), and the aldehyde of 42. A treatment with diluted acid completed the sequence and the (R)-lactone was obtained in 94% ee.

H H CHO CH OH Ar a Ar 2 b Ar R O 57-60% CO2CH3 CO2CH3 (2 steps) O 42 ee 94%

Scheme 40. Reagents and condition: (a) Baker’s yeast, D-glucose, H2O, 30 °C, 50 hr; (b) 10% HCl, 70 °C, 20 min. Ar=Ar9, Ar16 (for Ar see Appendix 1).

3.2.5 Miscellaneous mechanisms

The long linear protocol developed by Ghosh164 (Scheme 41) commenced from the protected D-mannitol which was converted, through some manipulations (tandem periodate cleavage/Wittig-Horner reaction), into unsaturated alcohol 77 (E:Z = 4:1, not separated). The mixture as such was submitted to the Johnson-Claisen rearrangement165 which afforded both diastereomers 78 and 79 in a 1:1 ratio. The products were separated by flash chromatography. Compound 78, subsequent to alkylation, furnished 80 (diastereomerically pure after flash chromatography) which in turn, after a reduction, a series of oxidations, and a final decarboxylation, was converted to (S)-β-benzyl-γ-butyrolactone. Ester 79, which theoretically could lead to the (R)-isomer, was left at the stage of triol.

56 R O O OH a O H O O 67% O O O OH O

R=CO Et b 2 62% R=CH2OH, 77 c 68% H R CO2Et CO2Et H Ar (78) CH OH Ar 2 f d O O + O CH2OH 84% H H H O 70% O O H de 83% OH 78 1 : 1 79 g 87% e R=CO2Et, 80 95% R=CH2OH

H H H Ar h Ar Ar R O S O O 75% H R OH H O O

i R=CH=CH2 51% R=CO H j 2 R=H 84%

Scheme 41. Reagents and conditions: (a) NaIO4, MeCN-H2O, K2CO3, PO(OEt)2CH2CO2Et, 0 °C → rt, 16 hr; (b) LiAlH4, Et2O, -60 °C, 2 hr; (c) MeC(OEt)3, EtCO2H, 140 °C, 6 hr; (d) LDA, THF, HMPA, ArCH2Br, -78 °C, 4 hr; (e) LiAlH4, Et2O, rt, 4 hr; (f) AcOH, H2O, rt, 18 hr; (g) NaIO4, MeCN-H2O, rt, 1.5 hr; (h) Jones reagent, acetone, 0 °C; (i) NaIO4, RuCl3⋅H2O, MeCN-CCl4-H2O, 3 hr, then Jones reagent, acetone; (j) PhMe, reflux, 16 h. Ar=Ar3 (for Ar see Appendix 1).

The classical Stobbe condensation was revised in a stereoselective fashion by Xia et al.166 (Scheme 42). The key point of the protocol is the resolution of the dicarboxylic acid 81 with (–)-quinine, possible because the quinine salt of the (R)-isomer is the first to crystallise. (R)-81 was converted, after esterification and a double reduction, to a mixture of trans and cis diols. The diols were chromatographically separated and each product was oxidized to the respective lactone, thus (8R,8'R)- and (±)-cis-lignanolactones were obtained. The (S)-81 isomer could be recovered from the mother liquor.

57 CO2Et a CO2R ArCHO b Ar 92% Ar 87-90% Ar CO2Et CO2R c R=Et 95% R=H, 81

d 45%

H H H CH2OH CH2OH CO2R Ar Ar f Ar R + 82% Ar CH OH CH OH CO R H 2 H 2 2 Ar 1 : 1 Ar e R=H, (R)-81 91% R=Et g 93% g 93%

H H Ar Ar O O

H H Ar O Ar O (8R,8'R) (±)

Scheme 42. Reagent and conditions: (a) (CH2CO2Et)2, NaOEt, EtOH, reflux, 4 hr; (b) LDA, Ar'CH2Br, -78 °C, 2 hr; (c) 20% aq. NaOH, reflux, 3 hr; (d) 1. (-)-Quinine, EtOH, reflux 1 hr; 2. 2N HCl, 1 hr ; (e) EtOH:PhH:H2SO4 (100:50:1), reflux, 36 hr, (f) 1. H2, Pd/C (10%), EtOAc, 12 hr; 2. LiAlH4, THF, 10 hr; 5 then separation by chromatography; (g) Ag2CO3/celite, PhH, reflux. Ar=Ar (for Ar see Appendix 1).

3.3 Syntheses of 7'-OH- or 7'-OR-lignano-9,9'-lactones

Compared to the large number of protocols developed for the synthesis of non-oxygenated lignano-9,9'-lactones, discussed in the previous chapters and also appearing in the earlier literature, the procedures for the synthesis of 7'-OH- or 7'-OR-lignano-9,9'-lactone‡‡ have been until the recent years, surprisingly, very limited. To our knowledge, the first total racemic synthesis of a 7'-OH lignanolactone was performed in 1976,167 while the first total asymmetric synthesis appeared only in 2004.94 In this section, the syntheses have again been classified according to the key step mechanism, but the stereoselectivity of the reaction, this time, was not taken as discriminator.

3.3.1 Catalytic mechanisms

A recent approach168,169 involves a high diastereo- and enantioselective cross-aldol reaction which allowed the synthesis of (7'R)-7'-hydroxyenterolactone (7'R)-7 in few steps (Scheme

‡‡ OR = O-C, O-Si, etc.

58 43). The simultaneous formation of the new C-C bond and the desired hydroxyl group is catalysed by L-proline, which resulted to be the best catalyst, among the several tested. It is interesting that the alkylation was performed directly on the unprotected alcohol 82, as all attempts of protecting the alcohol failed, thus explaining the modest yield of the reaction. The only disappointing aspect to the work is that the authors reported an extra proton in the 1H NMR characterisation of the final product, which hopefully was only due to an oversight.§§ The synthesis of (8R,8'R)-enterolactone (−)-5 was also achieved, by reduction of the alcohol 82 followed by alkylation.

HO HO H H CHO a b Ar Ar ArCHO + O O 55% 45% CO2Me d.r. > 24:1 82 H ee 97% O Ar O d 71% c Ar=Ar3 82% Ar=Ar2, (7'R)-7 H H b Ar Ar O O 68% H Ar O O c Ar=Ar3 87% Ar=Ar2, (-)-5

Scheme 43. Reagents and conditions: (a) 1. L-Proline (20 mol%), DMF, 4 °C, 24 hr; 2. NaBH4, 35- 40 °C, 10 hr; (b) LiHMDS, THF, ArCH2Br, -78 → -54 °C, 20 hr; (c) BBr3, CH2Cl2, -18 °C, 10hr, (d). 10% 3→2 Pd/C (20 mol%), ClCH2CH2Cl, 50-60 °C, 20 hr. Ar=Ar (for Ar see Appendix 1).

In the strategy recently presented by Hodgson (Scheme 44),170 a methylene lactone carrying a secondary alcohol 83 was obtained through a Barbier-type reaction with good diastereoselectivity. The following allylation, mediated by catalyst 84 ([RhCl(cod)]2), furnished (7'S*,8R*,8'R*)-(±)-7'-hydroxymatairesinol (±)-15. The protocol has a limited number of steps and does not require the use of protecting groups.

§§ The authors claimed that the 1H NMR of (7'R)-7'-hydroxyenterolactone (7'R)-7 did not match with the previously reported data.93 This is not true if the extra signal is not accounted for.

59 HO HO H H c a, b O Ar Ar O O + O 55% 71-83% (2 steps) O O Br O 83 O d.r. = 90:10 d 85%

HO Ar Cl H Rh Rh B Ar O O Cl O 84 H O Ar 85 (±)-15

+ - Scheme 44. Reagents and conditions: (a) Me3PhN Br3 , 1,4-dioxane, rt, 15 h; (b) LiBr/Li2CO3, DMF, 60 °C, 10 h; (c) ArCHO, Zn, sat. aq. NH4Cl (trace), DMF, rt, 15h; (d) 84 (3 mol%), 85, 1,4-dioxane, 7 H2O, 100 °C, microwave, 1h. Ar=Ar (for Ar see Appendix 1).

The protocol recently presented by Coelho et al.171, is the improved version of a previously described synthesis.172,173 The starting step (Scheme 45) is a Baylis-Hillman (or Morita- Baylis-Hillman) reaction with DABCO used as catalyst. Compared to the earlier procedure, the reaction time was slightly reduced and the yield increased by using an ionic liquid

(1-butyl-3-methylimidazolium hexafluorophosphate, [bmim]PF6) and ultrasound. The addition product 86 was used as Michael acceptor for the nucleophilic cyanide anion. When the Michael addition was performed directly on the alcohol (step b, Scheme 45), the result was a mixture of diastereomers with a poor stereoselectivity (syn:anti = 1:1.5). The problem was overcome by oxidizing directly the mixture and reducing diastereoselectively the obtained lactone. The resulted anti cyanide 87 (diastereomeric ratio 10:1), purified by flash chromatography, was transformed in the final (±)-anti butyrolactone. When the Michael addition was performed on 88, the outcome was a 4:1 mixture in favour of the syn isomer 89. The mixture of diastereomers could not be entirely separated and the final product was the enantiomerically enriched (±)-syn alcohol, obtained in a similar way as the corresponding anti isomer.

60 OH HO HO a b H H ArCHO CO2Me CO2Me CO2Me 89% Ar 87% Ar + Ar 86 CN CN syn:anti = 1:1.5 e 95% c 94%

OTBDMS O H CO Me CO2Me Ar 2 Ar

88 CN b 83% d 75% d.r. = 10:1

TBDMSO TBDMSO RO H H H CO2Me CO2Me CO2Me Ar + Ar Ar CN CN 87 CN 89 syn:anti = 4:1 e R=H 78% R=TBDMS

f 91% HO H Ar HO TBDMSO O H H Ar g, h Ar O O (±) O 87% O (2 steps) OH (±)

Scheme 45. Reagents and conditions: (a) CH2=CHCO2Me, DABCO (10 mol%), [bmim]PF6, MeOH, ultrasound, rt, 80 hr; (b) NaCN, NH4Cl, DMF-H2O, rt, 12 hr; (c) Jones reagent, acetone, 0 °C ; (d) NaBH4, MeOH, 0 °C → rt, 90 min; (e) TBDMSCl, imidazol, DMF, rt, 18 hr; (f) DIBAL, CH2Cl2, -78 °C, 3 5 hr; (g) NMO, TPAP, CH2Cl2, rt, 1 hr; (h) TBAF, MeOH, 0 °C 2 hr. Ar=Ar (for Ar see Appendix 1).

3.3.2 Radical mechanisms

The key step in the synthesis developed by Quayle174 is an atom transfer radical cyclization reaction (Kharasch-type reaction, Scheme 46). Compound 90, obtained via two different methods, was treated with a copper catalyst (formed in situ) to afford a 2:3 mixture of lactone epimers. The authors proved that this ratio was due to a post-cyclization equilibration. The stereochemistry of the main threo product was confirmed by X-ray analysis. The diastereomeric mixture afforded only the single threo benzyl alcohol (d.r. > 95:5) upon solvolysis. The stereoselectivity is probably due to the formation of a planar benzylic cation, created by the scission of the C-Cl bond, which is captured by the benzyl alcohol

61 approaching from the less hindered face of the plane. A final dechlorination gave the desired lactone.

Cl Cl H H a Ar O d Ar Ar Ar CHO O + O 90% 66% Cl C O Cl Cl 3 Cl 90 Cl O O threo:erythro = 3:2

b c 91% 85% (2 steps) e d.r. >90% Ar OH BnO H H3C(H2C)6 (CH2)6CH3 Ar O R R N N O dHbipy f R=Cl 83% R=H (±)

Scheme 46. Reagents and conditions: (a) 1. Ph3P=CHCO2Et, THF, 20 °C, 2. DIBAL, THF, -78 °C, 3. ClCOCCl3, Et3N, Et2O, 0 °C, 3 hr; (b) CH2=CHMgBr, THF, -78 °C → rt; (c) ClCOCCl3, Et3N, Et2O, 0 °C, 1 hr; (d) CuCl, dHbipy (5 mol%), ClCH2CH2Cl, reflux, 3.5 hr; (e) BnOH, CH2Cl2, 50 °C; (f) Bu3SnH, AIBN (20 mol%), PhH, 80 °C. Ar=Ar5 (for Ar see Appendix 1).

Fischer and co-workers94 have the merit of having pioneered the total stereoselective synthesis of several enantiopure 7'-hydroxylignano-9,9'-lactones (Scheme 47). The first part of their strategy, where the chirality of the molecule is created effectively, is based on the Evans’ oxazolidinone protocol (Xc=Xc4, Fig. 12) with the formation of a syn-aldol adduct 91. This methodology, as will be highlighted in the next paragraphs, has been utilized by other research groups and the novelty of Sherburn’s sequence can be found in the steps that follow. After the removal of the chiral auxiliary, the synthesis proceeded with the formation of a thionocarbonate 92, which in turn underwent a radical domino reaction that in the end furnished the (7'S)-7'-TBDMSO-dibenzylbutyrolactones in acceptable yield and high diastereoselectivity. The proposed reaction mechanism involves the addition of the radical • species (MR3) to the sulfur, followed by a sequence of two 5-exo cyclizations to create two new C-C bonds, and two eliminations steps. After the removal of TBDMS protecting groups (step f), (7'S,8R,8'R)-7'-hydroxymatairesinol (7'S)-15, and (7'S,8R,8'R)-7'-hydroxyarctigenin (7'S)-27 were obtained. By means of this synthesis, the stereochemistry of naturally occurring 7'-hydroxyartigenin was proved to be incorrectly assigned (should have been 7'R instead of 7'S). The synthetic strategy was smartly concluded with the inversion of the

62 stereochemistry of the hydroxyl group at C-7' via a modified Mitsunobu reaction (step g, Scheme 47). In the case of (7'S)-15 (Ar=Ar'=Ar7), the phenols had to be re-protected as the TBDMS derivatives prior to the Mitsunobu reaction. This was not necessary in case of (7'S)- 27 (Ar=Ar9, Ar'=Ar7).

O RO O TBDMSO H H a c Xc Ar Xc Ar OR 77-85%

b R=H, 91 d R=H 83-92% R=TBDMS 81-88% R=CSOAr', 92 (2 steps) e 44%

HO RO H H Ar R [(7'S)-27] S O Ar R g, 47% O H O H Ar' Ar' O (7'R)-15 R=TBDMS (7'R)-27 f R=H, [(7'S)-15] and 86-90% [(7'S)-27] [(7'S)-15]

h, 77%; g, 83%; f, 72%

Scheme 47. Reagents and conditions: (a) 1. n-Bu2BOTf, Et3N, ArCHO, CH2Cl2, -78 → 0 °C, 1 hr; 2. H2O2, pH 7.2 buffer, Et2O, rt, 48 hr; (b) TBDMSOTf, 2,6-lutidine, CH2Cl2, rt, 30 min; (c) NaBH4, THF-H2O, rt, 16 hr; (d) Py, Ar'OCSCl, CH2Cl2, 25 °C, 2 hr; (e) TMSSiH, AIBN, PhH, 80 °C; (f) TBAF, AcOH, THF, rt, 96 hr; (g) 1. PPh3, p-NO2C6H4CO2H, DEAD, PhH, rt, 48 hr; 2. K2CO3, 4 9 KHCO3, MeOH-H2O, rt, 1 hr; (h)TBDMSCl, imidazole, DMF, 25 °C ,2 hr. Xc=Xc (see Fig. 12); Ar= Ar , Ar11→7; Ar'= Ar11→7 (for Ar see Appendix 1).

3.3.3 Use of chiral auxiliaries

The first part of the strategy discussed above (chapter 3.3.2) was also adopted by Yamauchi et al.175,176 However, as the antipode of the Evans’ oxazolidinone R-imide was used as starting material (Xc=Xc2, Fig. 12), S-butyrolactones were obtained as final product (Scheme 48).

63 O RO TIPSO TIPSO H H H a R' d, e f Xc R Ar Ar Ar O S 93% 73% O (3 steps) b R=H OH O 100% R=TIPS c R'=COXc 62% R'=CH2OH

Scheme 48. Reagents and conditions: (a) n-Bu2BOTf, Et3N, CH2Cl2, ArCHO, -78 → 0 °C, 1 hr; (b) TIPSOTf, 2,6-lutidine, CH2Cl2, rt, 1 hr; (c) LiBH4, MeOH, THF, rt, 16 hr; (d) OsO4, NMO, aq. acetone, 2 5 t-BuOH, rt, 48 hr; (e) NaIO4, MeOH, rt, 16 hr; (f) PCC, CH2Cl2, rt, 16 hr. Xc=Xc (see Fig. 12); Ar=Ar (for Ar see Appendix 1).

In a similar way, Xie et al.177,178 (Scheme 49) employed the Evans' oxazolidinone strategy using a homochiral auxiliary as that in Scheme 48, but obtaining an opposite outcome. The reason for this is that MgCl2 in the presence of TMSCl and Et3N promotes the formation the non-Evans anti-aldol adduct. Thus (7'R,8R,8'R)-hydroxymatairesinol 15 was the final product of the procedure.

O RO TBDMSO H H R' d a Ar R Ar Xc R O 70% 86%

b R=H OH 86% R=TBDMS c R'=COXc e 82% 70% R'=CH2OH

HO TBDMSO H H Ar R f, g, h Ar R O O 66% H (3 steps) Ar O O (7'R)-15

Scheme 49. Reagents and conditions: (a) MgCl2, TMSCl, Et3N, EtOAc, 0.5 hr; (b) TBDMSCl, imidazole, DMF; (c) LiBH4, Et2O-H2O; (d) OsO4, NaIO4, 2,6-lutidine, dioxane/H2O; (e) PCC, molecular sieves, CH2Cl2, 2 hr; (f) LHMDS, HMPA, ArCH2Br, THF, -78 °C, 3 hr; (g) TBAF, THF; (h) H2, Pd/C, EtOH, 4 hr. Xc=Xc6 (see Fig. 12); Ar=Ar10→7 (for Ar see Appendix 1).

The routes developed by Sherburn,94 Yamauchi175 and Xie177 represent an interesting way to control the output of a synthesis, simply by altering the catalyst, or the chiral auxiliary, showing the versatility of Evans’ chemistry in the lignan synthesis.

64 The general methodology explored by Enders and co-workers already in 1996,179 was recently revived to fashion 7'-hydroxylignanolactones.180,181 The protocol is centred on chiral α-aminonitriles 93, obtained via the Strecker reaction, which act as Michael donors, to afford diastereoselectively the (S) adducts 94. Alkylation and hydrolysis of the chiral auxiliary furnished the 8,8'-trans-7'-oxolignanolactone in high optical purity. The ketone was then reduced to give an epimeric mixture of alcohol, further oxidized to (8S,8'S)-lignanolactones.

X CN X CN H H X a b c ArCHO Ar Ar S O 59-81% Ar CN 79% O 78-100% de >98% de >98% 93 94 H O Ar' O

d 80% de >98% ee >96% R R O H H N H Ar Ar e Ar Ph O + O O X= 80-100% OO H H O H Ar' O Ar' Ar' O 83 : 17 XH, 95 f R=OH 88% R=H

Scheme 50. Reagents and conditions: (a) 93, MeOH (or H2O), AcOH (or HCl), KCN, pH = 4–5; (b) 1. LDA, THF, -78 °C, 90 min; 2. butenolide, -78 °C; (c) t-BuLi, THF, -78 °C, 90 min; 2. Ar'CH2Br, -90 → 0° C; (d). AgNO3, H2O, THF, 25 °C; (e) NaBH4, MeOH/CH2Cl2; (f) H2, Pd/C, HClO4, MeOH, 4 atm., rt. Ar=Ar1, Ar5; Ar'=Ar1, Ar5, Ar9, Ar14 (for Ar see Appendix 1).

Yamauchi182 developed a different strategy based on the aldol addition of the enantiopure (S)-trityloxypentanolide 96, prepared from L-glutamic acid, to an aromatic aldehyde (Scheme 51). In this case, the stereoselectivity of the new C-C bond is controlled by the ether moiety. The erythro and threo aldol reaction products (ratio 1:1) were chromatographically separated and individually used to obtain chiral β-benzyl-γ-butyrolactones with an opposite benzylic configuration: the S from the erythro isomer and the R from the threo. The synthesis was straightforward and according to the authors the most challenging step was the selective removal of the trityl ether without affecting the methoxymethyl ether. The orthogonal deprotection was accomplished using formic acid and prior protection of the primary alcohol as pivaloyl ester. In successive publications, the authors presented variations on some reagents and on the protecting group management steps (MOM replaced by TIPS to avoid the t-BuCO protection), but the overall yield diminished from 28% to 15-18% as a result.38,183

65

O RO HO H O H O MOMO (threo) H a Ar Ar O O + O Ar R 86% R O

TrO TrO TrO O 96 erythro:threo = 1:1 (erythro) R=H b 93% R=MOM c 88% (2 steps)

MOMO MOMO H H Ar OCOt-Bu e, f, g Ar S R O OH 81% (3 steps) O RO d R=Tr 91% R=H

Scheme 51. Reagents and conditions: (a) LDA, ArCHO, THF, -75 °C, 1 hr, then separation by chromatography; (b) MOMCl, i-Pr2NEt, CH2Cl2, rt, 16 hr; (c) 1. LiAlH4, THF, 0 °C, 30 min; 2. t-BuCOCl, py, rt, 17 hr; (d) HCO2H, Et2O, -5 °C, 1 hr; (e) 1N aq. NaOH, EtOH, rt, 20 hr; (f) NaIO4, aq. t-BuOH (or 5 10 20 MeOH), rt, 2 hr; (g) Ag2CO3-Celite, PhMe, reflux, 1 hr. Ar=Ar , Ar , Ar (for Ar see Appendix 1).

A similar strategy was adopted to synthesise a cis-8,8'-lignanolactone (Scheme 52).184,185 Compared to the procedure described in Scheme 51, the stereoselectivity of the aldol reaction was improved by using, as starting material, (3R,4S)-R-hydroxy-S-trityloxy-4- pentanolide 97, prepared from L-arabinose. The hydroxyl group of the chiral auxiliary is the crucial moiety that directs the stereoselectivity of the new C-C bond and also increases the yield of threo alcohol at the expense of the erythro product (9:1 ratio). The epimers were chromatographically separated and the major isomer 98 was used to carry on the synthesis similarly as described in Scheme 51. A key point of the protocol is the formation of the hemiacetal 99 as a single isomer, with epimerization at C-3. After conversion of 99 into a lactone, removal of the hydroxyl group via a thionoformate, and alkylation, the last step involved a rearrangement of the trisubstituted 7',8'-cis-8,8'-trans-lignano-9,7'-lactone, which will be discussed in more detail in chapter 4. Thus the cis-8,8'-parabenzlactone was synthesised in a 5% overall yield from pentanolide 97.

66 RO HO O H O H O a Ar Ar O + O 62% O HO RO RO TrO TrO TrO 97 threo:erythro = 9:1 (threo) b R=H, 98 84% R=MOM 88% c (2 steps)

Ar H H Ar PivO OMOM e, f PivO O R OH 70% S MOMO (2 steps) MOMO OH OR 99 R=Tr d g 87% 76% R=H HO H H Ar H Ar Ar R l PivO j PivO O O O 41% 56% R H H Ar O Ar O O h R=OMOM 56% k R=OH 89% R=OCSOPh i 87% R=H

Scheme 52. Reagents and conditions: (a) LDA, ArCHO, THF, -75 °C, 1 hr, then separation by chromatography; (b) MOMCl, i-Pr2NEt, CH2Cl2, rt, 3 hr; (c) 1. LiAlH4, THF, rt, 30 min; 2. PivCl, py, rt, 1 hr; (d) PPTS, MeOH, reflux, 3 hr; (e) NaIO4, MeOH, rt, 3 hr; (f) PPTS, t-BuOH, reflux, 2 hr; (g) Ag2CO3- Celite, PhMe, reflux, 2 hr; (h) TMSBr, CH2Cl2, -15 °C, 3 hr; (k) ClCSOPh, py, DMAP, MeCN, rt, 1h; (i) n-Bu3SnH, AIBN, PhMe, reflux, 1 hr; (j) LHMDS, ArCH2Br, THF, -75 °C, 1 hr; (l) 1M aq. NaOH, EtOH, rt, 16 hr; then 6M aq. HCl (twice). Ar=Ar5 (for Ar see Appendix 1).

3.3.4 Miscellaneous mechanisms

Mäkelä et al.93 were the first to accomplish the total synthesis of (7'R*,8R*,8'R*)-7'- hydroxymatairesinol 15 and (7'S*, 8R*,8'R*)-7'-hydroxyenterolactone 7. The 7' epimers were separated by semi-preparative HPLC and fully characterised, but because at that time the absolute configuration of this type of lignan had not yet been definitely assigned, the C-7' stereochemistry was erroneously allocated based on the available literature.89,90 The dithiophenol 100, prepared by a modification of the tandem Michael addition-alkylation protocol, fashioned by Ziegler and Schwartz186 and developed by Pelter et al.,187 was

67 converted to a ketone, in turn reduced to a 3:2 mixture of isomers separated chromatographically.

SPh PhS O HO HO H H H H Ar a Ar b Ar Ar O O O + O 61-75% 64-73% H H H H Ar O Ar O Ar O Ar O 100 (±) (±) 3:2

Scheme 53. Reagents and conditions: (a) HgO, BF3·Et2O, THF-H2O, rt, 4 hr; (b) H2, Raney nickel, EtOH, rt, 1h; separation by chromatography. Ar=Ar3→2, Ar10→7 (for Ar see Appendix 1).

The strategy presented by Yoda and co-workers188,189 started from hydroxylactone 101190 readily converted to a keto-amide 102, which was arylated using a Grignard reaction (de up to 99%).191 The monobenzyl compound furnished the target hydroxylactone after a protection-deprotection-cyclization series. Despite the fact that the authors had developed a diastereomeric separation to obtain enantiomerically pure Bn-protected 101,192 the synthesis was performed on the racemate.

O H H HO O a RO OH b H OR 73-96% NMe 2 NMe2 O O 101 O R=Bn or TBDPS 102 37-76% c (2 steps) de up to 99% R'O R'O H H Ar d Ar OR O 61-78% NMe2 3 steps (±) O O R'=H R'=TIPS or Bn

Scheme 54. Reagents and conditions: for R=Bn: (a) 1. BnBr, Ag2O, EtOAc; 2. Me2NH, MeOH; (b) (COCl)2, DMSO, THF, then Et3N, -78 → -45 °C; (c) ArMgBr, THF, 0 °C; (d) 1. TIPSOTf, 2,6-lutidine, CH2Cl2; 2. H2, 5% Pd/C, MeOH; 3. PPTS, PhMe, 45 °C; for R=TBDPS: (a) 1. TBDPSCl, CH2Cl2, imidazole; 2. Me2NH, MeOH; (b) as before; (c) as before; (d) 1. BnBr, Ag2O, EtOAc; 2. TBAF, THF; 3. p-TsOH, PhMe. Ar=Ar5, Ar14 (for Ar see Appendix 1).

68 Uchiyama and co-workers193 developed this synthesis aiming to reproduce the natural compound (±)-arthrinone, a biologically active compound containing a semiquinone structure, and thus not to a lignan (Scheme 55). 2,3-Dihidrofuran 103 was converted to acetal 104 via a selenium intermediate. A stereoselective [2,3]-Wittig rearrangement of 104 furnished the primary alcohol which was transformed in 105, a suitable substrate for a regioselective Diels–Alder reaction with 106. The associated retro-Diels–Alder furnished the desired aromatic moiety of the benzyl-butyrolactone.

R'O H a b c O O O O 94% 86% 95% R 103 PhSe O O d R=TMS 98% R=H OMe e R'=H 85% R'=TBDPS R=H TMS TMS f 96% R=CO2Et 104 OTMS 88% 106 g, h (2 steps)

TBDPSO TBDPSO TBDPSO H MeO H H Ar i O O O 81% EtO2C CO Et TMSO 2 (±) O O O 105

Scheme 55. Reagents and conditions: (a) PhSeCl, CH2Cl2, TMS-2-propyn-1-ol, -78 → 50 °C; (b) 30% H2O2, NaHCO3, AcOEt, THF, 0 °C → rt; (c) n-BuLi, THF, -78 °C; (d) TBAF, THF, 0 °C → rt; (e) TBDPSCl, imidazole, DMF, rt; (f) n-BuLi, THF, -78 °C, then ClCO2Et, -78 °C → rt; (g) AcOH, THF, 20 H2O, rt; (h) PCC, MS 4 Å, CH2Cl2, 0 °C → rt, (i) 106, 160 °C; then THF, 5% HCl. Ar=Ar (for Ar see Appendix 1).

69 4 Rearrangement reactions of hydroxylated lignanolactones under basic conditions

In 1976, Nishiba et al.88 reported the formation of a mixture of two hydroxylactones during the deacetylation of naturally occurring acetylparabenzlactone 107 with KOH/MeOH (Scheme 56). Spectroscopic data indicated that the two products were the expected deacetylated (–)- parabenzlactone 25 and its structural isomer 108a. As discussed in Chapter 2, 25 was considered to possess the stereochemistry described in Scheme 56, this is to say (7'R,8R,8'R), and thus the same absolute configuration was attributed to 107. The regioisomer of 25, which was named isoparabenzlactone, was established to be a lariciresinol-type lignan, a structure reported only once before.87 Unfortunately, in the same way that the absolute configuration of (–)-parabenzlactone was erroneously assigned, an incorrect configuration (7',8'-cis-8,8'-cis, as that described in Scheme 56) was also attributed to the new lignan 108a. Investigations into the stereochemistry of these compounds and the assignment of the correct configuration will be discussed in chapter 5.2. In addition, the same group reported that treating what was then believed to be (7'S,8R,8'R)-7'- acetylhydroxyarctigenin monoacetate (7'S)-109 with similar basic conditions, no rearrangement occurred (Scheme 56), but no explanation was given for this behaviour.21

AcO HO H H H Ar Ar 7' 2% KOH Ar HO 8' O + 8' 7' 8 MeOH O 8 O H H H Ar O Ar O Ar O (7'R)-107 (7'R)-25 108a 65% (all-cis) 21%

AcO HO H H H Ar' Ar' 4% KOH Ar' HO O O O MeOH + H O H O H O Ar'' Ar''' Ar''' (7'S)-109 (7'S)-27 0% 65%

Scheme 56. Saponification of (7'R)-acetylparabenzlactone 10788 and (7'S)-7'-acetylhydroxyarctigenin monoacetate 10921 as reported by the authors. As it will be discussed in section 5.2, the configurations of the molecules (starting materials and products) are incorrect. Ar=Ar5, Ar'=Ar9, Ar''=Ar24, Ar'''=Ar7 (for Ar see Appendix 1).

70 During their investigation on the reactivity of 7'-hydroxymatairesinol 15, Eklund et al.194 found that a mixture of such diastereomers in alkaline/THF solution (pH 8-9) furnished, along with the other products, two rearranged lactones (Scheme 57) identified as the two isomers 7',8'-trans-8,8'-cis 110 and 7',8'-cis-8,8'-cis 108b. The outcome of the reaction is strictly connected to the starting material 7'-hydroxymatairesinol 15 and to the particular mechanism involved. On the basis of some evidence, the authors proposed the formation of a para- quinone methide as intermediate in the rearrangement reaction,195 which is made possible by the free para-hydroxyl group of the starting material 15. In this way, depending on which face the carboxylic acid anion approached the 7' position (route a or b, Scheme 57), both rearranged lactones could be formed (with a preference for the 7',8'-trans isomer 110, due to geometric reasons). The outcome of the rearranged lactones in this case does not depend on the stereochemistry at 7' of 15, as with the quinone methide mechanism, 7' loses its chirality prior to when the rearrangement occurs. It was observed, however, that (7'S)-15 reacted more easily than the R isomer. In addition, Eklund reported that under Nishibe’s conditions (2% KOH/MeOH) 7'-hydroxymatairesinol 15 furnished a small amount of 7'- methoxymatairesinol 111, as a result of the nucleophilic attack of methanol to the quinone methide intermediate.194 Therefore the reactivity of 15 is strictly dependent on the structure of the molecule and, in particular, on the presence of the para-OH.

b OH OH H H H MeO MeO MeO 7' aq. NaOH 8' O O a O- HO H2O/THF O O H pH 8-9 H H Ar O Ar O Ar O 15 a b (7'S:7'R = 96:4)

2% KOH/MeOH H Ar H Ar

OMe HO 8' 7' HO 8' 7' H O + O Ar H H O Ar O Ar O H 110 108b Ar O (7',8'-trans-8,8'-cis) (all-cis) 111 5% 1%

Scheme 57. Rearrangement of 7'-hydroxymatairesinol 15 in aqueous alkaline conditions.194 Ar=Ar7 (for Ar see Appendix 1).

Iwasaki’s group in order used the rearrangement reaction to obtain a suitable stereoisomer capable of being stereoselectively alkylated and to provide, after a reduction followed by a

71 lactone ring reinstallation, 8-alkyl- or 8-OH-lignano-9,9'-lactones.196,197 Thus the racemic mixture of (±)-112 was treated with NaH/DMF, giving a rearranged lactone as the only product in more than 75% yield (Scheme 58). During the rearrangement reaction, the epimerization at C-8 occurred to yield the thermodynamically more favourable 7',8'-trans-8,8'- trans product (±)-113 (all-trans).

HO H H Ar Ar NaH/DMF HO O O

H H Ar' O Ar' O (±)-112 (±)-113 (all-trans) >75%

Scheme 58. Rearrangement of (±)-112 in NaH/DMF to yield (±)-113 as only product.196,197 Ar=Ar5, Ar9; Ar'=Ar5, Ar9, Ar10 (for Ar see Appendix 1).

As briefly discussed in the previous chapter (Scheme 52), Yamauchi31,185 used the translactonization reaction to access a cis-8,8' lignan, which is difficult to obtain with other strategies. Compared to the rearrangements observed above, this one starts from the optically pure trisubstituted 7',8'-cis-8,8'-trans-lignano-9,7'-lactone 114, which upon treatment with aq. NaOH in MeOH, gives the deprotected 7',8'-cis-8,8'-trans-lactone 115 and the rearranged 8,8'-cis (7'R)-7'-hydroxylignano-9,9'-lactone (7'R)-116 in 49 and 25% yield respectively (Scheme 58). Upon an additional treatment of 115 with aq. NaOH/EtOH, (7'R)- 116 was obtained in a further 31%, while the starting material 116 was recovered in 69% yield. No side reactions (e.g. epimerization) were reported.

OH H Ar H Ar H PivO 1M aq.NaOH HO Ar O O + O EtOH H H H Ar O Ar O Ar O 115 (7'R)-116 114 49% 25% 1M aq. NaOH EtOH 69% 115, 31% 116

Scheme 59. Rearrangement of lactone 114 in aq. NaOH/EtOH. After two basic treatments 115 was obtained in 41% yield and (7'R)-116 in 34%. Ar=Ar5, Ar10 (for Ar see Appendix 1).

72 After this excursus, it is acknowledged that 7'-OH-lignano-9,9'-lactones and 9'-OH-lignano- 9,7'-lactones can rearrange into each other under various basic hydrolytic conditions. The outcome of the reaction depends on the starting material (structure, stereochemistry and Ar substituents, like in the peculiar case of 7'-OH-matairesinol 15) and on the reaction conditions. As will be discussed in detail in the next part of the thesis, our investigations on this subject revealed some errors in the published literature which could be rectified. At the same time, some mechanistic aspects concerning the intramolecular translactonization of these kinds of compounds will also be discussed and clarified.

73 5 Aim of the study

Our interest in lignans, and in 7'-OH-lignano-9,9'-lactones in particular, originated from the possible biological effects of these compounds and by the fact that, despite the number of various methods discussed in the sections above (chapter 4), only few syntheses of 7'-OH-lignano-9,9'-lactones were available in the literature when this project started. These methods were not stereoselective, or needed certain requirements (particular aromatic substituents, certain starting materials’ configuration) that would not allow a general synthetic approach. The lack of synthetic protocols was one of the reasons why the knowledge about 7'-OH-lignano-9,9'-lactones was still limited (for example the absolute configuration had not been definitely determined) and the situation was worsened by the conflicting and wrong data reported in the literature. Thus the main task of this project was to develop a general procedure for the stereoselective synthesis of 7'-OH-lignano-9,9'-lactones that allows the eventual production of both optically pure trans-(8R,8'R) and trans-(8S,8'S) isomers, and could be flexible in regard to the aromatic substitution patterns (i.e. not only naturally occurring substituents). During the optimization of this synthetic protocol, several problems were encountered and new challenges appeared along the way. The unexpected rearrangement reaction with L-Selectride® and the re-assignment of the configuration of (–)-parabenzlactone drove our interest in the study of rearrangement reactions in basic conditions. Being aware that some published data were incorrect, we also aimed to clarify and harmonize these conflicting data, especially concerning the stereochemistry and reactivity of 7'-OH-lignano-9,9'-lactones. Thus, the mechanism of the rearrangement of hydroxylignanolactones in basic conditions was investigated. In addition, the absolute configuration of two lignan regioisomers was also definitely proven by X-ray spectroscopy. In particular, the X-ray structures of 7'-OH-lignano-9,9'-lactones were obtained for the first time. A parallel project, connected to an international consortium study on lignans, phytoestrogens, and human health, concerned the synthesis of new possible mammalian lignan metabolites, for a possible use as reference material in metabolic and biology studies (for example as deuterated derivatives). In this chapter, the aromatic substituents will be indicated as described in Appendix 1.

74 5.1 Stereoselective synthesis of 7'-hydroxylignano-9,9'-lactones

5.1.1 The construction of the (8R,8'R)-lignano-9,9'-lactone framework

To fashion the lignan skeleton, the tandem Michael addition-alkylation reaction was chosen because our laboratory had a long experience with this type of protocol (for racemic lignano- 9,9'-lactones as well as 7'-hydroxy derivatives),93 and also because the reaction exhibited shortness, versatility, and sturdiness: the desired requirements in this synthesis. Optically pure 7'-OH-lignano-9,9'-lactones were obtained using the asymmetric tandem Michael addition-alkylation reaction, pioneered by Feringa’s and Wards' groups.198,199 The stereoselectivity of the tandem reaction was controlled by the chiral Michael acceptor (5R)-(–)-menthyloxy-2(5H)-furanone (5R)-117 (Fig. 13)200 which furnished the trans-(8R,8'R) lactone. The 5S isomer would eventually provide the less common trans-(8S,8'S) isomer.

O O

R S O O

O O (5R)-117 (5S)-117

Fig. 13. (5R)-(–)-menthyloxy-2(5H)-furanone (5R)-117 and its isomer (5S)-117.

Heartened by the reported positive results,197,201,202 initially an asymmetric tandem Michael addition-alkylation was attempted using a silyloxy-nitrile 118 as nucleophile (Scheme 60). This approach, however, was not satisfactory due to the silyloxy-nitrile hydrolysis and the reduction of the formed ketone occurring during the menthyloxy removal step, which yielded a complex mixture of products that were difficult to handle. Therefore the dithiane 119 (Scheme 60) was chosen as nucleophile, but in this case the tandem product resisted all attempts at removing the menthyloxy auxiliary.

75 R' R H R'' R a Ar + (5R)-117 + Ar'CH2Br O Ar R' 118 or 119 H Ar O R''=O-(-)-menthyl 118, R=TBDMSO, R'=CN b 119, R-R'=S(CH2)3S R''=H

Scheme 60. Reagents and conditions: (a) n-BuLi, THF, DMPU, -78 °C; (b) NaBH4, KOH, EtOH, 0 °C. Ar=Ar4 (for Ar see Appendix 1).

Thus the Michael addition product 120 (Scheme 61) was submitted to basic hydrolysis conditions followed by the in situ NaBH4 reduction of the formed aldehyde to furnish the chiral lactones 121 on ordinary acidic workup.203 No isomerization was detected by 1H NMR, using the chiral shift reagent Eu(hfc)3 (ee ≥98%). The lignan skeleton was then completed by alkylation of 121 with the suitable benzylic bromides to get only the trans-(8R,8'R)- lignanolactones 122.186 It is noteworthy that both steps a and c require the use of co-solvents DMPU and DMI respectively.

SS R SS SS a c Ar Ar O Ar 75-88% O 66-78% de 99% 119a-e H O Ar' O b R=O(-)menthyl, 120a-e 122a-f 60-82% R=H, 121a-e ee 99% a Ar=Ar'=Ar5 5 a Ar=Ar b Ar=Ar'=Ar11 11 b Ar=Ar c Ar=Ar'=Ar23 23 122 119,120,121 c Ar=Ar d Ar=Ar'=Ar22 22 d Ar=Ar e Ar=Ar9, Ar'=Ar11 e Ar=Ar9 5 9 f Ar=Ar , Ar'=Ar

Scheme 61. Reagents and conditions: (a) n-BuLi, THF, DMPU, (5R)-117, -78 °C; (b) 1. KOH, NaBH4, EtOH, 0 °C; 2. 0.1N HCl (pH 5-6); (c) LDA or LHMDS, THF, DMI, Ar'CH2Br, -78 °C → rt; (for Ar see Appendix 1).

5.1.2 The removal of the dithiane moiety and the reduction with L-Selectride

The removal of the dithiane moiety proved to be a challenging task. In the search for a general method, which would not affect the aromatic substituents (e.g. silyl groups), several protocols to reinstate the ketone were tested.204,205 Attempts were initially made with mercury

76 salts as HgCl2–HgO, HgCl2–CaCO3, Hg(ClO4)2•3H2O–DIPA. Target compounds, but in a very low yield, due to the problems in the product purification, or unreacted starting material were obtained. In addition, the high toxicity of these reagents prompted the search for a more efficient alternative. Thus standard protocols using CuCl2–CuCO or MeI–CaCO3, and more 206 207 208 recent protocols with NaNO2–AcCl, Dess-Martin periodinane or NaI–TaCl5–H2O2 were tried but all attempts were inefficient. Finally, the ketone 123 was unmasked using the Stork– 209 Zhao method but with 3 eq. of (CF3CO2)2IPh (Scheme 62). The reaction proceeded very rapidly and the purification procedure did not pose any problems. The only drawback was the moderate yield (50 to 80%), however neither starting material nor side products (e.g. desilylated compound) were recovered from the workup mixture. With the 7'-oxolignanolactone 123 available, an easy access to the final target was anticipated by using L-Selectride (step b, Scheme 62). Previously reported results with racemic lignans197 and 7-OH lignans179 as substrates were encouraging. As expected, a hydride attack from the less hindered face of the ketone lead to 7'S products 25, 124 and 112b-e, obtained in all cases with a high stereoselectivity. However, the reduction turned out to be more challenging than expected. Considerable dissimilarities in the outcome of the reduction were registered with de values ranging from 83 to 97% (Table 2). The de values were determined by 1H NMR and were based on the ratios between the H-7' signals (δ ≈ 4.6 for the S isomer and δ ≈ 4.4 for the R isomer in CDCl3, irrespective of the aromatic substituents). The reasons for this variation are not immediately obvious, as the differences in structures reside at relatively remote locations (aromatic substituents). Earlier works do not report this kind of behaviour.179,197 It is worthwhile to mention that, if the spectra are registered in solvents other than CDCl3, the chemical shifts for these diagnostic signals are shifted. For example, in acetone-d6 the H-7' signal for the R isomer occurs at about δ 4.5 (see Annex 2, Table 8), while for the S isomer at δ 4.8 (see Annex 2, Table 7).

77 O HO H H H Ar a Ar 7' b Ar S HO 7' 122a-f O O + O 50-80% H H H Ar' O Ar' O Ar' O 123a-f (7'S)-25, Ar=Ar'=Ar5 126a-f c 5 9 (126b) (7'S)-124, Ar=Ar , Ar'=Ar 110 5 a Ar=Ar'=Ar (7'S)-112b-e b Ar=Ar'=Ar11 23 c Ar=Ar'=Ar c (7'S)-15, Ar=Ar'=Ar7 112,123,126 22 d Ar=Ar'=Ar (7'S)-125, Ar=Ar'=Ar8 9 11 e Ar=Ar , Ar'=Ar (7'S)-7, Ar=Ar'=Ar2 5 9 f Ar=Ar , Ar'=Ar (7'S)-27, Ar=Ar9, Ar'=Ar7

Scheme 62. Reagents and conditions: (a) (CF3CO2)2IPh, MeCN-H2O, rt; (b) L-Selectride, THF, -78 °C; (c) TBAF/AcOH, THF, 0 °C; (for Ar see Appendix 1).

Table 2. L-Selectride reduction of 123a-f. Entry Ketone Alcohol de (%)a δ (H-7'S) δ (H-7'R) Yield (%)b 1 123a 25 97 4.57 4.37 70 2 123b 112b 93 4.62 4.38 75 3 123c 112c 83 4.60 4.38 77 4 123d 112d 88 4.53 4.35 68 5 123e 112e 96 4.61 4.38 80 6 123f 124 97 4.66 4.39 61 1 (a) Determined by H NMR of crude products (CDCl3). (b) After flash chromatography (7'S + 7'R).

The reduction with L-Selectride (followed by protecting groups removal, when necessary) provided the target (7'S)-7'-hydroxyenterolactone (7'S)-7 (a mammalian lignan, obtained for the first time in enantiopure form), (7'S)-7'-hydroxymatairesinol (7'S)-15, (7'S)- parabenzlactone (7'S)-25, (7'S)-7'-hydroxyarctigenin (7'S)-27, (7'S)-7'-hydroxybursehernin (7'S)-124 and (7'S)-7'-hydroxyprestegane B (7'S)-125 (the last two not found in natural matrices yet). Spectroscopic data of (7'S)-7 were similar to those reported earlier by Mäkelä93 for the racemic material. It must be noted, however, that our 7'S isomer corresponds to Mäkelä’s 7'R compound, as Mäkelä’s results were based on outdated and incorrect data. Spectroscopic data of (7'S)-15 and (7'S)-27 were in agreement with published data and supported the findings of Eklund91 and Fisher.94 Finally, spectroscopic data of (7'S)-25 showed that the stereochemistry of naturally occurring (–)-parabenzlactone, and consequently of 7'-acetylparabenzlactone 107 (Scheme 56), had been erroneously assigned as 7'R. The unquestionable evidence of the absolute configuration of (7'S)-(–)- parabenzlactone is given by the X-ray structures (Fig. 14). Also (7'S)-124 afforded crystals suitable for X-ray study. These are the first ever reported X-ray analyses for (7'S,8R,8R')-7'-

78 OH-lignano-9,9'-lactones. Very close similarity can be observed for the two structures. In both crystals, the γ-butyrolactone ring presents a chair conformation with the C8 atom pointing below the lactone plane. In addition, the aromatic rings are in a folded conformation, which appears to be the less stable one in a recently published computer-based conformational study of 7'-hydroxymatairesinol.210

Fig. 14. ORTEP plot of (7'S)-parabenzlactone (7'S)-25 (50% probability ellipsoids).

Fig. 15. ORTEP plot of (7'S)-7'-hydroxybursehernin (7'S)-124 (50% probability ellipsoids).

79 Treatment with L-Selectride not only furnished the desired secondary alcohol, but also a byproduct that was identified as the rearranged lactone 126 (Scheme 62). In some cases this lactone was only detected by NMR, but in the cases of 126a, 126b, and 126f the products were isolated and characterized. The 7',8'-trans-8,8'-cis stereochemistry was assigned by 2D NMR (COSY and NOESY) and on the basis of the presumable mechanism of formation: a translactonization reaction where the 7'-alkoxyborane from the selectride reaction acted as nucleophile towards the lactone ring, with the formation of a primary alcohol as a consequence. This rearranged lactone was found to be a homochiral analogue of the major isomer 110 (7',8'-trans-8,8'-cis), one of the two lariciresinol-type lactones reported by Eklund shortly before,194 and obtained by a basic treatment of 7'-hydroxymatairesinol 15 (chapter 4, Scheme 57). Correlations between the spectroscopic data of 126a, 126b, 126f and 110 can be observed in Annex 2, Table 10. In particular, they all have a diagnostic signal for H-7' 3 (δ ≈ 5.5, J7',8' = 2.0-2.6 Hz), which appears to be typical for this configuration. Thus, even if two different mechanisms of formation are involved (simple transesterification for 126 and via a quinone methide for 110), the two molecules have the same stereochemistry. The absolute configuration of the rearranged product was definitely established by X-ray analysis of 126a (Fig. 16). In addition, compound 126b was converted to 110 and the spectroscopic data were in agreement with those reported by Eklund.194 Compound 110 was also established to be a naturally occurring lignan as well and not only a synthetic product.194

Fig. 16. ORTEP plot of the rearranged lactone 126a (50% probability ellipsoids), obtained as side product in the L-Selectride reduction of 123a.

80 Additional studies on the L-Selectride reaction, using 123a as substrate, showed that increasing the amount of reagent, the rearranged product 126a increased at the expense of (7'S)-25 (Table 3). However, the reasons for this behaviour are not yet disclosed and studies are still ongoing in our laboratory.

Table 3. Outcome of the reduction of 123a with increasing amount of L-Selectride Ratio Entry eq. of L-Selectride (7'S)-25:126aa

1 1.1 82:18 2 5 69:31 3 10 53:47 (a) determined by 1H NMR of the crude material.

5.2 Rearrangement and stereochemistry studies

The reassignment of the stereochemistry of naturally occurring (–)-parabenzlactone 25 and 7'-acetylparabenzlactone 107 as 7'S (thus different as described in Scheme 56), also involved a re-examination of the lignano-9,7'-lactone 108a obtained from the treatment of (7'S)-107 with 2% KOH/MeOH and originally described with a 7',8'-cis-8,8'-cis configuration (all-cis, Scheme 56).88 Based on the corrected stereochemistry of (7'S)-107, the rearranged product obtained by a simple translactonization should be the 7',8'-trans-8,8'-cis isomer 126a, but the reported NMR data of 108a88 (in particular the 7'-H signal, doublet at δ 5.13,

3 J7',8' = 8 Hz) are not well-matched with those obtained for compound 126a (doublet at δ 5.45, 3 1 J7',8' = 2.6 Hz) in the L-selectride reduction. It was noticed, however, that the H NMR data of 3 108a (7'-H doublet at δ 5.12, J7',8' = 9.3 Hz) were in fact consistent with those of 108b, the 7',8'-cis-8,8'-cis minor rearranged lactone reported by Eklund et al.194 (chapter 4, Scheme 3 57), which has the H-7' signal at about δ 5.1, J7',8' = 9.3 Hz. Because an all-cis configuration could only be obtained from an inversion at 7', it was considered that, in competition with the acetate hydrolysis and thus the formation of (7'S)-25, the lactone may also undergo hydrolysis with the formation of a primary alcohol and a carboxyl group (Scheme 63). The COO generated in basic conditions can act as a nucleophile and attack the 7' center via a

SN2 mechanism, with simultaneous detachment of the AcO group. The final result would be a new lactone ring with inversion of configuration at the 7' position (Scheme 63). One should notice that for acetylparabenzlactone, mere AcO hydrolysis is not sufficient to trigger the lactone rearrangement, and no other rearranged lactones were mentioned in Nishibe’s paper.88 In addition, the stereochemistry at 7' should be such that the approach of COO is allowed, thus giving a possible explanation as to why (7'R)-7'-acetylhydroxyarctigenin

81 monoacetate (7'R)-109 (the stereochemistry of this compound had been erroneously assigned as in Scheme 5621 and should have been 7'R as described in Scheme 63) failed to furnish the rearranged lactone.

AcO AcO Ar H H H Ar 7' 2% KOH A OH HO O 7' O MeOH O- H O H H Ar Ar O Ar O (7'S)-107 108a

AcO H AcO 2% KOH H Ar' 7' O Ar' OH MeOH O- H O H Ar'' Ar''' O (7'R)-109

Scheme 63. Possible mechanism of formation of 108a from (7'S)-acetylparabenzalctone (7'S)-107. Ar=Ar5; Ar'=Ar9; Ar''=Ar24; Ar'''=Ar7 (for Ar see Appendix 1).

5.2.1 Translactonization of 18O-labeled substrates

To verify whether this mechanism could be plausible some studies were pursued using 18O-labeled compounds and monitoring the results by mass spectroscopy (MS).

To distinguish between the simple translactonization and the SN2 rearrangement mechanism, the labeled (7'S)-7'-18O-acetylhydroxylignanolactone (7'S)-18O-107 was prepared and assessed by MS for the conservation vs. loss of the label. In the simple translactonization route, the 18O should be retained in the product, but should be missing in the rearranged 18 lactone derived by the SN2 mechanism. The O label was introduced during the deprotection 18 of the dithiane moiety of 122 using H2 O as co-solvent (Scheme 64). The treatment of 18O-123a,b with L-Selectride furnished, as expected, the (7'S)-18O-25 and (7'S)-18O-112b along with the rearranged 9,7'-lactones 18O-126a,b in a ca. 85:15 ratio (Scheme 64). The MS spectra showed that the 18O label was totally conserved in the rearranged lactone18O-126, as anticipated for a simple translactonization (L-Selectride reduction). The acetylated (7'S)-18O- 107 was submitted to hydrolytic conditions and a mixture of (7'S)-18O-25 and rearranged lactone was obtained in a ca. 60:40 ratio. However the MS analysis of the latter compound showed that the molecule still contained the 18O label. This result was in conflict with a possible SN2 mechanism, which was therefore ruled out. Moreover, a simple

82 translactonization would have furnished the product obtained by L-Selectride 18O-126 and easily distinguished from the hydrolysis product 18O-25, thanks to the diagnostic 7'-H NMR signal. Finally, a careful examination by MS and 2D NMR revealed that this material is the 7',8'-trans-8,8'-trans isomer (all-trans) 18O-113a (Scheme 64), which arose from a translactonization accompanied by an α-epimerization. This material is the chiral version of (±)-113 obtained before by Iwasaki after treating (7'S*,8R*,8'R*)-7'-hydroxylignanolactone (±)-112 with NaH/DMF,196,197 and discussed earlier in Scheme 58. Thus, the coincidental NMR resemblance between the all-cis 108b and the all-trans 113a lactones (see Appendix 2) erroneously lead us to the wrong SN2 mechanism proposal. (±)-113a has been previously reported in unlabeled form.211

18 18O OH H H H Ar a Ar b Ar HO 122a,b O O + 18O

H H H Ar O Ar O Ar O 18 85 :15 18 O-123a,b (7'S)-18O-25 O-126a,b (7'S)-18O-112b c (Ar=Ar5)

18OAc 18OH H H H Ar Ar d Ar HO O O + 18O

H H H Ar O Ar O Ar O (7'S)-18O-107 (7'S)-18O-25 18O-113a

18 Scheme 64. Reagents and conditions: (a) (CF3CO2)2IPh, MeCN/H2 O, rt; (b) L-Selectride, THF, 5 11 -78 °C; (c) Ac2O-py, rt; (d) 2% KOH/MeOH, rt. For a Ar=Ar , for b Ar=Ar (for Ar see Appendix 1).

The absolute configuration of 113a was confirmed by X-ray analysis of the 3,5-dinitrobenzoate derivative 127 (Fig. 17).

83

Fig. 17. ORTEP plot of 127 (50% probability ellipsoids).

5.2.2 Translactonization studies with NaH/DMF and KOH/MeOH

Due to the similarity (as for the stereochemistry) in the outcome using different reaction conditions (NaH/DMF and 2% KOH/MeOH), it was interesting to directly compare the two translactonization/α-epimerization reactions. Thus (–)-parabenzlactone (7'S)-25 was submitted to both Nishibe’s88 and Iwasaki’s197 rearrangement conditions. The KOH/MeOH reaction furnished, as expected, a mixture of starting material (7'S)-25 and rearranged lactone 113a in a ca. 60:40 ratio, as observed for the hydrolysis of (7'S)-18O-107 (Scheme 64). Surprisingly, in our hands, a mixture of the starting material (7'S)-25 and the rearranged all-trans lactone 113a in a ca. 40:60 ratio was furnished also by NaH/DMF. This proved, therefore, to be very different from Iwasaki’s results, that showed a 75% yield and neither recovery nor detection of starting material reported.197 To exclude any effects due to the unlike aromatic substituents, Nishibe’s and Iwasaki’s rearrangement conditions were used on the enantiopure compound (7'S)-7'-hydroxybursehernin (7'S)-124 (whilst Iwasaki was using the racemic mixture) as substrate. With NaH/DMF (7'S)-124 gave results comparable to those obtained for (7'S)-25 for both yield and ratio between the starting material and the rearranged lactone (Table 4). As expected, the nature of phenolic protection did not play a role in these reactions.

84 OH OH H H H Ar Ar a or b Ar HO O O + O

H H H Ar' O Ar' O Ar' O (7'S)-25 (7'S)-25 113a (7'S)-124 (7'S)-124 113f

(7'S)-25, 113a, Ar=Ar'=Ar5 (7'S)-124, 113f, Ar=Ar5, Ar'=Ar9

Scheme 65. Reagents and conditions: (a) 2% KOH/MeOH, rt; (b) NaH/DMF, 0 °C; (for Ar see Appendix 1).

Table 4. Outcome of the reaction described in Scheme 65. Ratio of products Reagents (7'S)-25:113a (7'S)-124:113f 2% KOH/MeOH 64:36 57:43 NaH/DMF 40:60 36:64

The difference in the predominant product obtained under Nishibe’s and Iwasaki’s conditions may be explained by the reaction mechanisms. When NaH/DMF is employed, the lactone moiety of (7'S)-25 undergoes a nucleophilic attack by the formed alkoxide ion and the α- epimerization occurs more or less simultaneously to give the thermodynamically more stable all-trans product 113a (Scheme 66).

HO O- H H H Ar Ar NaH/DMF Ar HO O O O

H H H Ar O Ar O Ar O (7'S)-25 113a

Scheme 66. Mechanism of translactonization/α-epimerization of (7'S)-25 in NaH/DMF. Ar=Ar5 (for Ar see Appendix 1).

In the case of KOH/MeOH (Scheme 67), the lactone moiety of (7'S)-25 undergoes transesterification by the action of methoxide present in the reaction medium, forming the methyl ester diol intermediate, which can be deprotonated at the α position. After acidification, relactonization can occur with either hydroxy group present in the intermediate, but the preference is determined by the selectivity of α-proton reinsertion into the molecule,

85 prior to relactonization. A reprotonation from the less hindered upper face leads back to the starting material (7'S)-25, whereas protonation from the lower face will lead to the rearranged 113. From the (7'S)-25:113a ratio values, it is clear that the structure (7'S)-25 is preferred. Evidence for this equilibrium was obtained by treating 113a with 2% KOH/MeOH. Translactonization/α-epimerization is observed and the ratio (7'S)-25:113a = 66:34 was obtained after 48 hours (compared to 1.5 hours required when (7'S)-25 was the starting material). The 9,7'-lactone 113a is probably more hindered than (7'S)-25, making transesterification less favorable. Furthermore, when (7'S)-25 and 113a were treated with 2% KOH/t-BuOH, no rearrangement was observed.

(7'S)-25 113a

2% KOH 2% KOH HCl HCl MeOH MeOH

HO OH HO OH HO OH HO OH H H H H + +H+ Ar -H Ar Ar Ar OMe +H+ OMe OMe -H+ H H Ar O Ar O Ar O Ar O

Scheme 67. Mechanism of translactonization/α-epimerization in 2% KOH/MeOH. Ar=Ar5 (for Ar see Appendix 1).

5.2.3 Translactonization studies with aq. NaOH/EtOH

The translactonization conditions used to obtain the unnatural (7'R,8S,8'R)-8,8'-cis- parabenzlactone (7'R)-116 from 7',8'-cis-8,8'-trans-lignano-9,7'-lactone 114 or 115 (aq. NaOH/EtOH, Scheme 59)185 were tested on the 7',8'-trans-8,8'-trans 113a in order to check whether it was possible to obtain the 7'S isomer of (7'S)-116 (Scheme 68). Unfortunately no translactonization was observed and only the starting material was recovered. Since the only difference between our and Yamauchi’s reaction can be found in the stereochemistry at the 7' site of the starting material, it is likely that this must play a crucial role in the formation of an all-cis 7'-hydroxylignano-9,9'-lactone. To confirm that steric hindrance is a key feature in these types of reaction, also (7'S)-25 was treated with 1M aq. NaOH/EtOH (Scheme 68). As expected, also in this case no rearrangement product (126a) was observed. Thus for substrates with stereochemistry like 113a and (7'S)-25, a rearrangement in basic hydrolytic conditions is allowed only with an α-epimerization (Scheme 67), which is not observed in aq.

86 NaOH/EtOH. In support of these considerations, 113a and (7'S)-25 were further treated with 2% aq. KOH/MeOH (50:50) and again only starting material was obtained.

OH H Ar H a HO Ar O O

H H Ar O Ar O 113a (7'S)-116 OH H H Ar a Ar HO O O

H H Ar O Ar O (7'S)-25 126a

Scheme 68. Reagents and conditions: (a) 1. 1M aq. NaOH, EtOH (or 2% aq. KOH, MeOH), rt; 2. 2N HCl. Ar=Ar5 (for Ar see Appendix 1).

5.2.4 Translactonization studies of (±)-8-methylparabenzlactone

To discriminate between the role of the stereochemistry at 7' and that of α-epimerization, translactonizations under basic conditions were further tested on (±)-8- methylparabenzlactone 128a (Scheme 69). Lactone 128a, a compound where no α protons are available, was synthesized with the Michael addition–alkylation protocol using 129 as Michael acceptor. The reduction of ketone 130 with L-Selectride furnished as expected (±)- (7'S)-8-methylparabenzlactone 128a (de 99%) and the 7',8'-trans-8,8'-cis translactonization product 131a in a ca. 84:16 ratio (Scheme 69). Lactone 128a was subjected to the various conditions discussed in chapter 5.2.2 and 5.2.3 (2% KOH/MeOH, NaH/DMF and 1M aq. NaOH/EtOH), and furnished, in all cases, the starting material and the rearranged product 131a in a ca. 65:35 ratio. Thus the presence of the methyl group at C-8 clearly affects the behaviour of the rearrangement reaction. In the case of 1M aq. NaOH/EtOH, the presence of the water in the reaction medium does not prevent (as seen for (7'S)-25, Scheme 68) the reaction from occurring, providing the same ratio as with 2% KOH/MeOH, in turn comparable with those obtained for (7'S)-25 and (7'S)-124. On the other hand, NaH/DMF furnished less rearranged lactone 131a when compared with the results obtained with (7'S)-25 and (7'S)- 124 (where the ratio starting material/rearranged product was 40:60), likely due to steric hindrance exerted by the 8-methyl in the nucleophilic attack and by the 8,8'-cis configuration of the product.

87 SS SS a Ar b Ar + O O O SS Me Me Me Ar O O Ar O 119a 129 c

HO Ar O H H H d Ar HO Ar O + O O

Me Me Me Ar O Ar O Ar O 128a 131a 130 e, or f, or g 128a:131a = 65:35

Scheme 69. Reagents and conditions: (a) n-BuLi, THF, DMPU, -78 °C; (b) LHMDS, THF, DMI,

ArCH2Br, -78 °C → rt; (c) (CF3CO2)2IPh, MeCN-H2O, rt; (d) L-Selectride, THF, -78 °C; (e) 1. 2% KOH/MeOH, rt; 2. 2N HCl; (f) 1. NaH/DMF, 0 °C; 2. 2N HCl; (g) 1. 1M aq. NaOH, EtOH, rt; 2. 2N HCl. Ar=Ar5 (for Ar see Appendix 1).

5.2.5 Translactonization studies of Ar-O-silylated (7'S)-7'-hydroxymatairesinol

In connection with another study, we have observed that phenolic silyl ethers can be cleaved selectively in the presence of aliphatic silyl ethers by means of NaH in DMF.212 Thus the aryl- O-silyl protected (7'S)-7'-hydroxymatairesinol (7'S)-112b (see Scheme 62) was subjected to the NaH/DMF translactonization conditions (Scheme 70). As anticipated the reaction proved to be a powerful desilylating tool, and the all-trans rearranged/α-epimerized product with free phenolic groups 113c was obtained. However, the major product of this reaction was the unexpected TBDMS-derivative 132 formed by a migration of the silyl group to the aliphatic hydroxyl (Scheme 70). Traces of totally and partially silyl protected starting material were also detected by NMR, but were not isolated. The combined yield of rearranged lactones 113c and 132 is higher than the yield of the rearranged lactones obtained by us from (–)- parabenzlactone (7'S)-25 and (7'S)-7'-hydroxybursehernin (7'S)-124 under the same reaction conditions, but is in line with the original results obtained by Iwasaki on (±)-112 (Scheme 58).197 The difference in the outcome of the reaction could not be explained. Maybe a retro- translactonization is also involved in the reaction with NaH/DMF and additional investigations in this respect would be necessary.

88 HO H H Ar' H Ar' Ar a HO TBDMSO O O + O

H H H Ar O Ar' O Ar' O (7'S)-112b 113c 132 26% 55%

Scheme 70. Reagents and conditions: (a) 1. NaH/DMF, 0 °C; 2. 2N HCl. Ar=Ar11; Ar'=Ar7 (for Ar see Appendix 1).

As can be observed in Table 9 (Annex 2), the 1H NMR spectra of all-trans lignano-9,7'- lactones 113a, 113b, 113c, 127, and 132, as well as of reported MOM protected lactones 133a-c,197 show close resemblance, particularly regarding the diagnostic peak of H-7'. A similar behaviour can also be observed for 7',8'-trans-8,8'-cis-lignano-9,7'-lactones 126a, 126b, 126f and 110 (Table 10, Annex 2), as well as for 7',8'-cis-8,8'-trans-lignano-9,7'- lactones 114 and 115 reported by Yamauchi185 (Table 11, Annex 2). It is interesting to notice that regardless of the presence of a free or a protected alcohol (TBDMS, MOM, Piv) there are no remarkable differences in the NMR spectra of homochiral lignans. This feature can be useful in determining the stereochemistry of stereorelated trisubstituted lignano-9,7'-lactones, as it was possible for 7'-HO-lignano-9,9'-lactones (Table 7 and Table 8, Annex 2). Also in the case of 8-alkylated lignano-9,7'-lactones, there is a correlation between the 1H NMR spectra of 128a and of the other reported 128b,c197 all possessing a 7',8'-trans-8,8'-cis configuration (see Table 13, Annex 2). However, NMR correlations may not be possible for the all-cis stereoisomers. In fact, the only two all-cis lignano-9,7'-lactones that have been reported to date, 108b194 and (±)-134211 (Fig. 18), present a different 1H NMR spectrum in both chemical shifts and coupling constants (Table 12, Annex 2). As an example, the diagnostic peak of H- 3 3 7' is found at δ 5.12 ( J7',8' = 9.3 Hz) for 108b and at δ 5.43 ( J7',8' = 5.4 Hz) for (±)-134. The sizeable variation of these values could be caused by the bulky TBDMS group which may significantly affect the conformation in case of 7',8'-cis-8,8'-cis-lignano-9,7'-lactones (not observed for the other configurations). Surprisingly, the spectroscopic data ([α], 1H and 13C NMR) of 108b and of the all-trans isomers show notable similarity (see Table 9 and Table 12, Annex 2) and one could speculate if the Eklund’s all-cis lignan 108b possesses that configuration or if it is, in fact, the all-trans stereoisomer. In his paper Eklund presented convincing evidence (NOESY study, mechanism of formation) that the minor rearranged lactone is the all-cis lignano-9,7'-lactones 108b.194 In addition, the tentatively identification of the 8-epimer of 110 (that is to say the all-trans 113c) in the extractives of Norway spruce

89 knots was reported. Thus, the resemblance of spectroscopic data between the all-trans 113 and all-cis 108b should be seen just as an accidental coincidence.

H Ar H Ar'

HO 8' 7' TBDMSO 8 O O H H Ar O Ar' O 108b (±)-134

Fig. 18. Known 7',8'-cis-8,8'-cis-lignano-9,7'-lactones.194,211 Ar=Ar7; Ar'=Ar5 (for Ar see Appendix 1).

A summary of the basic reactions of lignanolactones is presented in Appendix 3.

5.3 Synthesis of potential mammalian metabolites of lignans

After the discovery that, in addition to matairesinol 9 and secoisolariciresinol 10 and their O-glucosides (Fig. 9), several other plant lignans could be converted to mammalian lignans (or enterolignans) by the metabolic action of intestinal microflora,48,49 interest grew in the study of these new precursors, on their intake, on the possible metabolism and metabolites. Present interest is focused on developing lignan databases to estimate the lignan content of food (fruits, vegetables, cereals, whole grain products, beverages such as wine, beer, tea and coffee) and lignan intake in various populations.213 The plant lignans found to be precursors of enterolignans and thus under investigation, include pinoresinol 11, lariciresinol 12 (Scheme 1), 7'-hydroxymatairesinol 15, syringaresinol 135, medioresinol 136, and arctigenin 137 (Fig. 19).48,49

OMe MeO OH O O HO OR1 O MeO O HO OMe OMe MeO

R1=Me, Syringaresinol, 135 Arctigenin, 137 R1=H, Medioresinol, 136

Fig. 19. Additional naturally occurring precursors of mammalian lignans.

90 These studies on the new precursors have also led to the detection of previously unknown metabolic products. Unfortunately, the small amount of the products and the lack of authentic reference compounds often made the identification and the quantification of these new metabolites impossible. For example, the degradation of syringaresinol and medioresinol produced several compounds whose possible structure was postulated by Heinonen et al. on the basis of the mass spectra.48 These new metabolites should possess a dibenzylbutyrolactone- or a dibenzylbutanediol-type structure with variation concerning the presence and position of OH and OMe in the aromatic rings (Fig. 20). Authentic reference compounds are required to establish the structures of these potential new metabolites, thus we challenged the synthesis of several of the postulated metabolites presented in Fig. 20.

HO R3O HO OH OH O OH OH R6 1 R4O R O R2 R5O OH OH R7O OR8 3 4 5 R1=OH, R2=H R =R =R =H R6=R7=R8=H 3 4 5 R1=H, R2=OH R =R =R =Me R6=H, R7=R8=Me 3 4 5 R1=OMe, R2=H R =R =Me, R =H R6=R7=H, R8=Me 3 4 5 R1=H, R2=OMe R =H, R =R =Me R6=OH, R7=R8=Me R3=R5=H, R4=Me

Fig. 20. Structures of the postulated metabolites of syringaresinol 135 and medioresinol 136.48

The new (±)-trans-dibenzylbutyrolactones 138a-g were prepared by a modification of the tandem Michael addition–alkylation procedure.186,187,214 The protected dithioacetals 139a–g were obtained by reacting the acyl anion equivalents 140a-e with 2-butenolide 141 followed by in situ alkylation with appropriately substituted benzylic bromides 142a-e. In this case, the presence of co-solvent DMI is also essential for the reaction to succeed. Treatment of 139a- g with Raney nickel provided desulfurization and simultaneous removal of the benzylic protecting groups to afford the desired dibenzylbutyrolactones (±)-138a-g. The new dibenzylbutanediols (±)-143a-f were obtained by reducing lactones 138a-f with LiAlH4.

91 XX H H PhS SPh a Ar c Ar OH + O + Ar'CH Br 2 O OH Ar 142a-e H H 140a-g O Ar' O Ar' 141 X=SPh, (±)-139a-g b (±)-143a-f a Ar=Ar4 X=H, (±)-138a-g b Ar=Ar9 140 c Ar=Ar25 d Ar=Ar26 9 25 a Ar=Ar9, Ar'=Ar30 e Ar=Ar27 a Ar=Ar , Ar'=Ar b Ar=Ar25, Ar'=Ar25 b Ar=Ar30, Ar'=Ar30 26 27 6 8 a Ar'=Ar4 c Ar=Ar , Ar'=Ar c Ar=Ar , Ar'=Ar 27 28 8 28 b Ar'=Ar25 139 d Ar=Ar , Ar'=Ar 138,143 d Ar=Ar , Ar'=Ar 26 28 6 28 142 c Ar'=Ar27 e Ar=Ar , Ar'=Ar e Ar=Ar , Ar'=Ar 4 29 4 29 d Ar'=Ar28 f Ar=Ar , Ar'=Ar f Ar=Ar , Ar'=Ar 29 g Ar=Ar25, Ar'=Ar4 g Ar=Ar30, Ar'=Ar4 e Ar'=Ar

Scheme 71. Reagents and conditions: (a) n-BuLi, THF, DMI, -78 °C; (b) Raney nickel, EtOH, refl. 3-10 hr; (c) LiAlH4, THF, rt, 2-12hr; (for Ar see Appendix 1).

In the tandem reaction, ortho- and meta-substituted aromatic starting materials (in particular when benzyloxy groups are present in both meta positions) reacted less readily providing only a satisfactory yield, presumably due to steric reasons. Optimum yields were obtained using highly concentrated solutions of starting materials in THF. The crucial step in the three–component reaction is undoubtedly the alkylation step. In the majority of cases, unreacted Michael addition products were recovered in up to 30% yield while only small amounts of diphenylthioacetals 140a-e were isolated. It was observed that for a better yield the addition of the benzylbromide to the reaction mixture should be done relatively quickly after the addition of butenolide 141. Lactones (±)-138a–g and diols (±)-143a–f were fully characterized by IR, mass, and NMR spectroscopy. Two-dimensional NMR experiments (COSY, HSQC and HMBC) were used to unequivocally allocate the 13C signals. Subsequently, some of the synthesised (±)-trans-dibenzylbutyrolactones were deuterated in order to be used as internal standards in analytical studies.215

92 6 Conclusions

In this thesis a general asymmetric synthesis for obtaining optically active (7'S)-7'- hydroxylignano-9,9'-lactones was developed. The interest in these types of lignans is due to their possible biological effects. The optimized protocol allowed the preparation of several compounds: (7'S)-7'-hydroxyenterolactone (7'S)-7 (a mammalian lignan, obtained for the first time in enantiopure form), (7'S)-7'-hydroxymatairesinol (7'S)-15, (7'S)-parabenzlactone (7'S)- 25, (7'S)-7'-hydroxyarctigenin (7'S)-27, (7'S)-7'-hydroxybursehernin (7'S)-124, and (7'S)-7'- hydroxyprestegane B (7'S)-125. With the X-ray of (7'S)-25 and (7'S)-124, reported for the first time, it was possible to establish the absolute configuration of these stereoisomeric lignans and thus demonstrate that the reported configuration for the 7' position of naturally occurring (–)-parabenzlactone was incorrect. During the reduction with L-Selectride, a rearranged 7',8'- trans-8,8'-cis-lignano-9,7'-lactone was formed as a byproduct, likely by a translactonization mechanism. The absolute configuration was determined by X-ray crystallography. Similar lactones, but with different stereochemistry and different mechanism of formation, were reported in literature. As a result, the rearrangement reactions of various hydroxylignano-9,9'- and 9,7'-lactones were investigated. Studies with 18O labeled substrates showed that our presumed mechanism of formation was erroneous and the absolute stereochemistry of the rearranged lactone obtained by treating (7'S)-7'-acetylparabenzlactone with 2% KOH/EtOH was determined to be 7',8'-trans-8,8'-trans by X-ray analysis of its dinitrobenzoate derivative. This compound was derived from a translactonization accompanied by an α-epimerization. Different basic conditions were evaluated showing that the stereochemistry of the lactone and the possibility of an α-epimerization determine whether a rearrangement is observed or not. Rearrangement studies were also performed on an 8-methyl substituted lignanolactone. In this case the stereochemistry was not decisive for the formation of the rearranged lactone, which was obtained under all conditions. Finally, the treatment of an aromatic silyl protected lignan with NaH/DMF, furnished a concomitant rearrangement, α-epimerization and aryl silyl deprotection, together with an unexpected migration of the silyl to the aliphatic hydroxy group. The resemblance observed in the 1H NMR for hydroxylignano-9,9'- and -9,7'-lactones can be a useful expedient for the assignment of the stereochemistry.

In a parallel project, several new dibenzylbutyrolactones and dibenzylbutanediols were prepared, in two and three steps respectively, from aromatic starting materials. The new compounds were fully characterized by mass and NMR spectroscopy. The aromatic patterns of the new compounds were based on the postulated metabolites derived from the newly identified mammalian lignan precursors syringaresinol and medioresinol. These new

93 synthetic compounds could be used in the identification of new mammalian lignan metabolites, as well as in other biological or metabolic studies.

94 7 Experimental part

The synthesis of (7'S)-7'-hydroxybursehernin (7'S)-124, and its spectroscopic data, are included in paper II. The X-ray data (not included in the original articles I-IV) are presented in this section.

(7'S)-7'-hydroxybursehernin, (7'S)-124, (7'S,8R,8'R)-7'-hydroxy-3',4'-dimethoxy-3,4-methylenedioxylignano-9,9'-lactone

Table 5. Crystal data and structure refinement of (7'S)-124, (Fig. 16). Empirical formula C21H22O7 Formula weight 386.39 Temperature 150(2) K Wavelength 1.54184 Crystal system monoclinic Space group P21 Unit cell dimensions a=8.3530(3) b=6.4972(6) c=16.9997(7) alpha=90.018(5) beta=92.987(3) gamma=90.016(5) Volume 921.34(10) Z 2 Density (calculated) 1.393 mg/m3 Absorption coefficient, µ 0.875 mm-1 F(000) 408 Theta range for data collection 5.21-60.57 Independent reflections 2327 Completeness to theta 0.972 Extinction coefficient 0.875 Max. and min. transmission 1.000 and 0.607 Absorption correction MULTI-SCAN method Goodness-of-fit on F2 (S) 1.092 R indices (all data) R(reflections)=0.0251(2222) wR2(reflections)=0.0628(2327)

95 Table 6. Selected bond lengths (Å) and angles (°), (Fig. 16). Bond lengths (Å) Bond angles (°) O3-C3 1.3761 (0.0022) C9–O1–C9' 110.93 (0.13) O3–C10 1.4223 (0.0022) C8–C7–C1 111.26 (0.12) O4–C4 1.3826 (0.0021) C8'–C7'–C1' 111.31 (0.13) O4–C10 1.4211 (0.0026) C8'–C7'–O5 106.24 (0.12) O4'1-C4' 1.3738 (0.0021) C1'–C7'–O5 112.31 (0.14) O4'1-C4'1 1.4254 (0.0023) C8'–C8–C9 103.35 (0.15) O3'-C3' 1.3744 (0.0018) C7–C8–C9 109.03 (0.12) O3'-C3'1 1.4334 (0.0024) C7–C8–C8' 112.90 (0.13) O5–C7' 1.4274 (0.0019) C9'–C8'–C8 102.93 (0.14) O1–C9 1.3401 (0.0029) C7'–C8'–C8 112.38 (0.14) O2–C9 1.2139 (0.0023) C7'–C8'–C9' 112.36 (0.14) C7'–C8' 1.5402 (0.0022) O3–C10–O4 109.13 (0.15) C8–C9 1.5150 (0.0023) C8–C8' 1.5267 (0.0027) C8'–C9' 1.5365 (0.0025)

96 References

1 Haworth RD. Natural resins. Ann. Rep. Prog. Chem. 1936, 33, 266–279. 2 Gottlieb OR. Neolignans. Fortschs. Chem. Org. Naturst. 1978, 35, 11–72. 3 Ward RS. Lignans, neolignans and related compounds. Nat. Prod. Rep. 1997, 14, 43–74. 4 Struijs K, Vincken J-P, Verhoef R, Voragen AGJ, Gruppen H. Hydroxycinnamic acids are ester- linked directly to glucosyl moieties within the lignan macromolecule from flaxseed hulls. Phytochemistry 2008, 69, 1250–1260. 5 MacRae WD, Towers GHN. Biological activities of lignans. Phytochemistry 1984, 23, 1207– 1220 (ref. 15-30 therein). 6 Umezawa T. Phylogenetic distribution of lignan–producing plants. Wood Res. 2003, 90, 27–110. 7 Zhao C, Nagatsu A, Hatano K, Shirai N, Kato S, Ogihara Y. New lignan glycosides from Chinese medicinal plant, Sinopodophillum emodi. Chem. Pharm. Bull. 2003, 51, 255–261. 8 Koubaa I, Damak M. A new dilignan from Cynara cardunculus. Fitoterapia 2003, 74, 18–22. 9 Tsukamoto H, Hisada S, Nishibe S. Lignans from bark of the Olea plants. II. Chem. Pharm. Bull. 1985, 33, 1232–1241. 10 Wang HB, Mayer R, Rucker G, Neugebauer M. Bisepoxylignan glycosides from Stauntonia hexaphylla. Phytochemistry 1993, 34, 1621–1624. 11 Umezawa T, Okunishi T, Shimada M. Stereochemical diversity in lignan biosynthesis. Wood Res. 1997, 84, 62–75. 12 Haworth RD. Constituents of natural phenolic resins. Nature (London) 1941, 147, 255–257. 13 Hearon WM, MacGregor WS. The naturally occurring lignans. Chem. Rev. 1955, 55, 957–1068. 14 Weinges K, Nader F, Kuenstler K. Chemistry of lignans. Introduction. In Chemistry of Lignans, Rao CBS, Ed., Andhra University Press, Waltair, 1978, 1–37. 15 Freudenberg K, Weinges K. Systematik und Nomenklatur der Lignane. Tetrahedron 1961, 15, 115–128. 16 Kochetkov NK, Khorlin AY, Chizhov OS, Sheichenko VI. Schizandrin, a lignan of unusual structure. Tetrahedron Lett. 1961, 2, 730–734. 17 Whiting DA. Lignans and neolignans. Nat. Prod. Rep. 1985, 2, 191–211. 18 Suzuki S, Umezawa T. Biosynthesis of lignans and norlignans. J. Wood Sci. 2007, 53, 273–284. 19 Lewis NG, Kato MJ, Lopes N, Laurence BD. Lignans: diversity, biosynthesis, and function. ACS Symposium Series 1995, 588 (Chemistry of the Amazon), 135–167. 20 Ayers DC, Loike JD. Lignans: Chemical, biological, and clinical properties. Cambridge University Press, Cambridge, 1990. 21 Nishibe S, Chiba M, Sakushima A, Hisada S, Yamanouchi S, Takido M, Sankawa U, Sakakibara A. Introduction of an alcoholic hydroxyl group into 2,3-dibenzylbutyrolactone lignans with oxidizing agents and carbon-13 nuclear magnetic resonance spectra of the oxidation products. Chem. Pharm. Bull. 1980, 28, 850–860. 22 Botta B, Delle Monache G, Ricciardi P, Vitali A, Vinciguerra V, Misiti D, Kutney JP, Stoynov N. Studies with plant cell cultures of Cassia didymobotrya. VII. Enzyme catalyzed biotransformation of dibenzylbutanolides to podophyllotoxin analogues and related compounds. Heterocycles 1996, 43, 2443–2456.

97

23 Ward RS, Pelter A, Brizzi A, Sega A, Paoli P. Synthesis of diversely functionalized dibenzylbutyrolactones and aryltetralins from silylated cyanohydrin anions. J. Chem. Res. (S) 1998, 226–227. 24 Takemoto M, Matsuoka Y, Achiwa K, Kutney JP. Biocatalytic dediastereomerization of dibenzylbutanolides by plant cell cultures. Tetrahedron Lett. 2000, 41, 499–502. 25 Moss GP. Nomenclature of lignans and neolignans: (IUPAC recommendations 2000). Pure Appl. Chem. 2000, 72, 1493–1523. 26 Briggs LH. Hinokinin as the enantiomorph of cubebinolide. J. Am. Chem. Soc. 1935, 57, 1383– 1384. 27 Mameli E. Cubebin. VIII. Identity of cubebinolide with hinokinin. Gazz. Chim. Ital. 1935, 65, 886– 888. 28 Elfahmi, Ruslan K, Batterman S, Bos R, Kayser O, Woerdenbag HJ, Quax WJ. Lignan profile of Piper cubeba, an Indonesian medicinal plant. Biochem. Syst. Ecol. 2007, 35, 397–402. 29 Dixon RA, Achnine L, Kota P, Liu C-J, Reddy MKS, Wang L. The phenylpropanoid pathway and plant defence - a genomics perspective. Mol. Plant Path. 2002, 3, 371–390. 30 Lewis NG, Davin LB. Lignans: biosynthesis and function. Compr. Nat. Prod. Chem. 1999, 1, 639–712. 31 Akiyama K, Yamauchi S, Nakato T, Maruyama M, Sugahara T, Kishida T. Antifungal activity of tetrasubstituted tetrahydrofuran lignan, (–)-virgatusin, and its structure-activity relationship. Biosci., Biotechnol., Biochem. 2007, 71, 1028–1035. 32 Messiano GB, Vieira L, Machado MB, Lopes LMX, de Bortoli SA, Zukerman-Schpector J. Evaluation of insecticidal activity of diterpenes and lignans from Aristolochia malmeana against Anticarsia gemmatalis. J. Agric. Food Chem. 2008, 56, 2655–2659. 33 Harmatha J, Nawrot J. Insect feeding deterrent activity of lignans and related phenylpropanoids with a methylenedioxyphenyl (piperonyl) structure moiety. Entomol. Exp. Appl. 2002, 104, 51– 60. 34 Riós JL, Giner RM, Prieto JM. New findings on the bioactivity of lignans. In Studies in Natural Products Chemistry, Bioactive Natural Products (Part G), Atta-ur-Rahman ed., Elsevier Science, 2002, vol. 26, 26183–26292. 35 Saleem M, Kim HJ, Ali MS, Lee YS. An update on bioactive plant lignans. Nat. Prod. Rep. 2005, 22, 696–716. 36 Li N, Wu J, Hasegawa T, Sakai J, Wang L, Kakuta S, Furuya Y, Tomida A, Tsuruo T, Ando M. Bioactive dibenzylbutyrolactone and dibenzylbutanediol lignans from Peperomia duclouxii. J. Nat. Prod. 2006, 69, 234–239. 37 Matsumoto T, Hosono-Nishiyama K, Yamada H. Antiproliferative and apoptotic effects of butyrolactone lignans from Arctium lappa on leukemic cells. Planta Med. 2006, 72, 276–278. 38 Yamauchi S, Hayashi H, Nakashima Y, Kirikihira Y, Yamada K, Masuda T. Effect of benzylic oxygen on the antioxidant activity of phenolic lignans. J. Nat. Prod. 2005, 68, 1459–1470. 39 Charlton JL. Antiviral activity of lignans. J. Nat. Prod. 1998, 61, 1447–1451. 40 Fan C-Q, Zhu X-Z, Zhan Z-J, Ji X-Q, Hui L, Yue J-M. Lignans from Saussurea conica and their NO production suppressing activity. Planta Med. 2006, 72, 590–595. 41 Aehle E, Müller U, Eklund PC, Willför SM, Sippl W, Dräger B. Lignans as food constituents with estrogen and antiestrogen activity. Phytochemistry 2011, 72, 2396–2405. 42 Stitch SR, Toumba JK, Groen MB, Funke CW, Leemhuis J, Vink J, Woods GF. Excretion, isolation and structure of a new phenolic constituent of female urine. Nature 1980, 287, 738– 740.

98

43 Setchell KDR, Lawson AM, Mitchell FL, Adlercreutz H, Kirk DN, Axelson M. Lignans in man and animal species. Nature 1980, 287, 740–742. 44 Setchell KDR, Lawson AM, Conway E, Taylor NF, Kirk DN, Cooley G, Farrant RD, Wynn S, Axelson M. The definitive identification of the lignans trans-2,3-bis(3-hydroxybenzyl)-γ- butyrolactone and 2,3-bis(3-hydroxybenzyl)butane-1,4-diol in human and animal urine. Biochem. J. 1981, 197, 447–458. 45 Axelson M, Sjoevall J, Gustafsson BE, Setchell KDR. Origin of lignans in mammals and identification of a precursor from plants. Nature 1982, 298, 659–660. 46 Struijs K, Vincken J-P, Gruppen H. Bacterial conversion of secoisolariciresinol and anhydrosecoisolariciresinol. J. Appl. Microbiol. 2009, 107, 308–317. 47 Niemeyer HB, Honig DM, Kulling SE, Metzler M. Studies on the metabolism of the plant lignans secoisolariciresinol and matairesinol. J. Agric. Food Chem. 2003, 51, 6317–6325. 48 Heinonen S, Nurmi T, Liukkonen K, Poutanen K, Wähälä K, Deyama T, Nishibe S, Adlercreutz H. In vitro metabolism of plant lignans: new precursors of mammalian lignans enterolactone and enterodiol. J. Agric. Food Chem. 2001, 49, 3178–3186. 49 Smeds AI, Saarinen NM, Eklund PC, Sjöholm RE, Mäkelä SI. New lignan metabolites in rat urine. J. Chromatogr. B 2005, 816, 87–97. 50 Jacobs E, Kulling SE, Metzler M. Novel metabolites of the mammalian lignans enterolactone and enterodiol in human urine. J. Steroid Biochem. Mol. Biol. 1999, 68, 211–218. 51 Adlercreutz H, van der Wildt J, Kinzel J, Attalla H, Wähälä K, Mäkelä T, Hase T, Fotsis T. Lignan and isoflavonoid conjugates in human urine. J. Steroid Biochem. Mol. Biol. 1995, 52, 97– 103. 52 Adlercreutz H. Lignans and human health. Crit. Rev. Cl. Lab. Sci. 2007, 44, 483–525. 53 Landete JM. Plant and mammalian lignans: A review of source, intake, metabolism, intestinal bacteria and health. Food Res. Int. 2012, 46, 410–424. 54 Saarinen NM, Smeds A, Mäkelä, SI, Ämmälä J, Hakala K, Pihlava J-M, Ryhänen E-L, Sjöholm R, Santti R. Structural determinants of plant lignans for the formation of enterolactone in vivo. J. Chromatogr. B 2002, 777, 311–319. 55 Clavel T, Henderson G, Engst W, Doré J, Blaut M. Phylogeny of human intestinal bacteria that activate the dietary lignan secoisolariciresinol diglucoside. FEMS Microbiol Ecol. 2006, 55, 471– 478. 56 Clavel T, Lippman R, Gavini F, Doré J, Blaut M. Clostridium saccharogumia sp. nov. and Lactonifactor longoviformis gen. nov., sp. nov., two novel human faecal bacteria involved in the conversion of the dietary phytoestrogen secoisolariciresinol diglucoside. Syst. Appl. Microbiol. 2007, 30, 16–26. 57 Xie L-H, AkaoT, Hamasaki K, Deyama T, Hattori M. Biotransformation of pinoresinol diglucoside to mammalian lignans by human intestinal microflora, and isolation of Enterococcus faecalis strain PDG-1 responsible for the transformation of (+)-pinoresinol to (+)-lariciresinol. Chem. Pharm. Bull. 2003, 51, 508–515. 58 Wang L-Q, Meselhy MR, Li Y, Qin GW, Hattori M. Human intestinal bacteria capable of transforming secoisolariciresinol diglucoside to mammalian lignans, enterodiol and enterolactone. Chem. Pharm. Bull. 2000, 48, 1606–1610. 59 Jin J-S, Hattori M. Further Studies on a Human Intestinal Bacterium Ruminococcus sp. END-1 for transformation of plant lignans to mammalian lignans J. Agric. Food Chem. 2009, 57, 7537– 7542.

99

60 Jin J-S, Zhao Y-F, Nakamura N, Akao T, Kakiuchi N, Min B-S, Hattori M. Enantioselective dehydroxylation of enterodiol and enterolactone precursors by human intestinal bacteria. Biol. Pharm. Bull. 2007, 30, 2113–2119. 61 Jin J-S, Kakiuchi N, Hattori M. Enantioselective oxidation of enterodiol to enterolactone by human intestinal bacteria. Biol. Pharm. Bull. 2007, 30, 2204–2206. 62 Jin J-S, Hattori M. Human intestinal bacterium, strain END-2 is responsible for demethylation as well as lactonization during plant lignan metabolism. Biol. Pharm. Bull. 2010, 33, 1443–1447. 63 Miura D, Saarinen NM, Miura Y, Santti R, Yagasaki K. Hydroxymatairesinol and its mammalian metabolite enterolactone reduce the growth and metastasis of subcutaneous AH109A hepatomas in rats. Nut. Canc. 2007, 58, 49–59. 64 Bylund A, Lundin E, Zhang JX, Nordin A, Kaaks R, Stenman UH, Aman P, Adlercreutz H, Nilsson TK, Hallmans G, Bergh A, Stattin P. Randomised controlled short-term intervention pilot study on rye bran bread in prostate cancer. Eur. J. Cancer Prev. 2003, 12, 407–415. 65 Saarinen NM, Smeds AI, Peñalvo JL, Nurmi T, Adlercreutz H, Mäkelä S. Flaxseed ingestion alters ratio of enterolactone enantiomers in human serum. J. Nutr. Metab. 2010, vol. 2010, article ID 403076, 5 pages. 66 Touré A, Xueming X. Flaxseed lignans: source, biosynthesis, metabolism, antioxidant activity, bio-active components, and health benefits. Compr. Rev. Food Sci. Food Safety 2010, 9, 261– 269. 67 Carnmalm B. Absolute configuration in the lignan series. Ark. Kemi 1960, 15, 215–222. 68 Schrecker AW, Hartwell JL. The absolute configuration of lignans. J. Am. Chem. Soc. 1957, 79, 3827–3831. 69 Haworth RD, Kelly W. The constituents of natural phenolic resins. Part VIII. Lariciresinol, cubebin and some stereochemical relationships. J. Chem. Soc. 1937, 384–391. 70 Haworth RD, Wilson L. The constituents of natural phenolic resins. Part XXII. Reduction of some lactonic lignans with lithium aluminium hydride. J. Chem. Soc. 1950, 71–72. 71 Kuo Y-H, Chen C-H, Lin Y-L. New lignans from the heartwood of Chamaecyparis obtusa var. formosana. Chem. Pharm. Bull. 2002, 50, 978–980. 72 Li N, Wu J-L, Sakai J Ando M. Dibenzylbutyrolactone and dibenzylbutanediol lignans from Peperomia duclouxii. J. Nat. Prod. 2003, 66, 1421–1426. 73 Chen Y-C, Liao C-H, Chen I-S. Lignans, an amide and anti-platelet activities from Piper philippinum. Phytochemistry 2007, 68, 2101–2111. 74 Chyu C-F, Kuo Y-H. New Lignans from the roots of Taiwania cryptomerioides HAYATA. Helv. Chim. Acta 2007, 90, 738–747. 75 Singh M, Tiwari N, Shanker K, Verma RM Gupta AK, Gupta MM. Two new lignans from Phyllanthus amarus. J. Asian Nat. Prod. Res. 2009, 11, 562–568. 76 Min B-S, Na M-K, Oh S-R, Ahn K-S, Jeong G-S, Li G, Lee S-K, Joung H, Lee H-K. New furofuran and butyrolactone lignans with antioxidant activity from the stem bark of Styrax japonica. J. Nat. Prod. 2004, 67, 1980–1984. 77 Chenevert R, Rose YS. A chemoenzymatic synthesis of both enantiomers of a cis-lignan lactone. Tetrahedron: Asymmetry 1998, 9, 2827–2831. 78 Ward RS, Pelter A, Edwards MI, Gilmore J. Asymmetric synthesis of a lignan lactone from a meso anhydride. Tetrahedron 1996, 52, 12799–12814. 79 Gao H, Li G, Zhang J, Zeng J. Arctigenin: a lignan from Arctium lappa. Acta Crystallogr. 2008, E64, o1538.

100

80 Caira MR, Watson WH, Silva M. Thujaplicatin methyl ether, C21H24O7. Acta Crystallogr. 1984, C40, 299–301. 81 Cooley G, Farrant RD, Kirk DN, Patel S, Wynn S, Buckingham MJ, Hawkes GE, Hursthouse MB, Galas AMR, Lawson AM, Setchell KDR. Structural analysis of the urinary lignan 2,3-bis(3- hydroxybenzyl)butan-4-olide (enterolactone). A 400 MHz nuclear magnetic resonance study for the solution state and X-ray study for the crystal state. J. Chem. Soc. Perkin Trans. 2 1984, 489–497. 82 Kiralj R, Ferreira MMC, Donate PM, da Silva R, Albuquerque S. Conformational study of (8α,8'β)-bis(substituted phenyl)-lignano-9,9'-lactones by means of combined computational, database mining, NMR, and chemometric approaches. J. Phys. Chem. A. 2007, 111, 6316– 6333. 83 Kuhn M, von Wartburg A. Podophyllum-Lignane: Struktur und Absolutkonfiguration von Podorhizol-β-D-glucosid (= Lignan F). Helv. Chim. Acta 1967, 50, 1546–1565. 84 House HO, Crumrine DS, Teranishi AY, Olmstead HD. Chemistry of carbanions. XXIII. Use of metal complexes to control the aldol condensation. J. Am. Chem. Soc. 1973, 95, 3310–3324. 85 Yamauchi S, Machi M, Kinoshita Y. Stereoselective syntheses of (–)-podorhizol lignan and its derivatives: erythro and threo preferential aldol condensation of potassium enolate from γ-butyrolactone with alkoxybenzaldehyde. Biosci., Biotechnol., Biochem. 1999, 63, 1453–1462. 86 Yamauchi S, Tanaka T, Kinoshita Y. First highly stereoselective synthesis of (+)-dihydrosesamin, a trisubstituted tetrahydrofuran-type of lignan, by using highly erythro- selective aldol condensation. J. Chem. Soc. Perkin Trans. 1 2001, 2158–2160. 87 Freudenberg K, Knof L. Die Lignane des Fichtenholzes. Chem. Ber. 1957, 90, 2957–2969. 88 Niwa M, Iguchi M, Yamamura S, Nishibe S. The stereostructure of acetylparabenzlactone and its conversion to a lariciresinol-type lactone. Bull. Chem. Soc. Jpn. 1976, 49, 3395–3360. 89 Kawamura F, Ohashi H, Kawai S, Teratani F, Kai Y. Photodiscoloration of western hemlock (Tsuga heterophylla) sapwood II. Structures of constituents causing photodiscoloration. Mokuzai Gakkaishi 1996, 42, 301–307. 90 Mattinen J, Sjöholm R, Ekman R. NMR- spectroscopic study of hydroxymatairesinol, the major lignan in Norway spruce (Picea Abies) heartwood. ACH Models Chem. 1998, 135, 583–590. 91 Eklund P, Sillanpää R, Sjöholm R. Synthetic transformation of hydroxymatairesinol from Norway spruce (Picea abies) to 7-hydroxysecoisolariciresinol, (+)-lariciresinol and (+)-cyclolariciresinol. J. Chem. Soc. Perkin trans. 1 2002, 1906–1910. 92 Kawamura F, Miyachi M, Kawai S, Ohashi H. Photodiscoloration of western hemlock (Tsuga heterophylla) sapwood III. Early stage of photodiscoloration reaction with lignans. J. Wood Sci. 1998, 44, 47–55. 93 Mäkelä TH, Kaltia SA, Wähälä KT, Hase TA. α,β-Dibenzyl-γ-butyrolactone lignan alcohols: total synthesis of (±)-7'-hydroxyenterolactone, (±)-7'-hydroxymatairesinol and (±)-8- hydroxyenterolactone. Steroids 2001, 66, 777–784. 94 Fischer J, Reynolds AJ, Sharp LA, Sherburn MS, Radical carboxyarylation approach to lignans. Total synthesis of (–)-arctigenin, (–)-matairesinol, and related natural products. Org. Lett. 2004, 6, 1345–1348. 95 Willför S, Hemming J, Reunanen M, Eckerman C, Holmbom B. Lignans and lipophilic extractives in Norway spruce knots and stemwood. Holzforschung 2003, 57, 27–36. 96 Ward RS. The synthesis of lignans and neolignans. Chem. Soc. Rev. 1982, 11, 75-125. 97 Chatterjee A, Banerji A, Banerji J, Pal SC, Ghosal T. Recent advances in the chemistry of lignans. Proc. Indian Acad. Sci. 1984, 93, 1031–1057.

101

98 Pelter A. Lignans: some properties and syntheses. In Recent advances in phytochemistry, Conn EE ed., Plenum Press, New York and London, 1986, 20, 201–241. 99 Ward RS. Asymmetric synthesis of lignans. Tetrahedron 1990, 46, 5029–5041. 100 Brown E. Total syntheses of optically active lignans. Trends Org. Chem. 1991, 2, 59–72. 101 Stevenson R. Some aspects of the chemistry of lignans. In Studies in natural products chemistry, Structure and chemistry (Part D), Atta-ur-Rahman ed., Elsevier, 1995, 17, 311–356. 102 Ward RS. Recent advances in the chemistry of lignans. In Studies in natural products chemistry, Bioactive natural products (Part E), Atta-ur-Rahman ed., Elsevier, 2000, 24, 739–798. 103 Ward RS. Different strategies for the chemical synthesis of lignans. Phytochem. Rev. 2003, 2, 391–400. 104 Del Signore G, Berner OM. Recent developments in the asymmetric synthesis of lignans. In Studies in natural products chemistry, Bioactive natural products (Part M), Atta-ur -Rahman ed., Elsevier, 2006, 33, 541–600. 105 Bouyssi D, Monteiro, Balme G. Transition metal-mediated strategies towards lignans and related natural compounds. Curr. Org. Chem. 2008, 12, 1570–1587. 106 Pan J-Y, Chen S-L, Yang M-H, Wu J, Sinkkonen J, Zou K. An update on lignans: natural products and synthesis. Nat. Prod. Rep. 2009, 26, 1251–1292. 107 Gil S, Parra M, Rodriguez P, Segura J. Recent developments in γ-lactone synthesis. Mini-Rev. Org. Chem. 2009, 6, 345–358. 108 Tomioka K, Ishiguro T, Koga K. First asymmetric total synthesis of (+)-steganacin determination of absolute stereochemistry. Tetrahedron Lett. 1980, 21, 2973–2976. 109 Ferrié L, Bouyssi D, Balme G. Selective Lewis Acid catalyzed transformation (γ-butyrolactone versus cyclopropane) of 2-methoxy-4-benzyltetrahydrofuran derivatives. Efficient synthesis of lignan lactones. Org. Lett. 2005, 7, 3143–3146. 110 Balme G, Bouyssi D, Monteiro N. Palladium-mediated cascade or multicomponent reactions: a new route to carbo- and heterocyclic compounds. Pure Appl. Chem. 2006, 78, 231–239. 111 Singh R, Singh GC, Ghosh SK. A novel approach to functionalized (E)-1,4-diaryl-1-butenes by Heck reaction and their applications for the construction of dibenzylbutyrolactone lignan skeletons by radical cyclization. Tetrahedron Lett. 2005, 46, 4719–4722. 112 Singh R, Singh GC, Ghosh SK. A novel approach towards dibenzylbutyrolactone lignans involving Heck and radical reactions: application to (±)-matairesinol synthesis. Eur. J. Org. Chem. 2007, 5376–5385. 113 Kamlage S, Sefkow M, Peter MG, Pool-Zobel BL. A short synthesis of biologically active lignan analogues. Chem. Comm. 2001, 331–332. 114 Zhang HX, Guibe F, Balavoine G. Palladium- and molybdenum-catalyzed hydrostannation of alkynes. A novel access to regio- and stereodefined vinylstannanes. J. Org. Chem. 1990, 55, 1857–1867. 115 Srikrishna A, Danieldoss S. Radical cyclisation based approach to lignans. Synthesis of 4-arylmethyldihydrofuran-2-ones. Synth. Commun. 2001, 31, 2357–2364. 116 Watanabea Y, Ishikawaa S, Takaoa G, Toru T. Radical cyclization on solid support: Synthesis of γ-butyrolactones. Tetrahedron Lett.1999, 40, 3411–3414. 117 Bencivenni G, LanzaT, Leardini R, Minozzi M, Nanni D, Spagnolo P, Zanardi G. Tin-free generation of alkyl radicals from alkyl 4-pentynyl sulfides via homolytic substitution at the sulfur atom. Org. Lett. 2008, 10, 1127–1130. 118 Zhu F, Li W, Wang Q, Hou Z. Regioselective oxidative coupling approach to the synthesis of (±)-matairesinol and (±)-secoisolariciresinol. Synlett 2006, 1780–1782.

102

119 Amancha PK, Liu H-J, Ly TW, Shia K-S. General approach to 2,3-dibenzyl-γ-butyrolactone lignans: application to the total synthesis of (±)-5'-methoxyyatein, (±)-5'-methoxyclusin, and (±)-4'-hydroxycubebinone. Eur. J. Org. Chem. 2010, 3473–3480. 120 Bella M, Piancatelli G, Pigro MC. 3-Phenylselanylfuran-2(5H)-one: a versatile building block in the synthesis of lignans. A new approach towards 3,4-dibenzyl-γ-butyrolactones. Tetrahedron 1999, 55, 12387–12398. 121 Koul S, Singh B, Taneja SC, Qazi GN. New chemo and chemo-enzymatic synthesis of β-benzyl- γ-butyrolactones. Tetrahedron 2003, 59, 3487–3491.

122 Bolm C, Palazzi C, Francio G, Leitner W. Baeyer-Villiger oxidation in compressed CO2. Chem. Comm. 2002, 1588–1589. 123 Bolm C, Beckmann O, Palazzi C. Chiral aluminum complexes as catalysts in asymmetric Baeyer-Villiger reactions of cyclobutanones. Can. J. Chem. 2001, 79, 1593–1597. 124 Bolm C, Beckmann O, Cosp A, Palazzi C. Enantioselective Baeyer-Villiger oxidations catalyzed by chiral magnesium complexes. Synlett 2001, 1461–1463. 125 Bolm C, Frison J-C, Zhang Y, Wulff WD. Vaulted biaryls: efficient ligands for the aluminum- catalyzed asymmetric Baeyer-Villiger reaction. Synlett 2004, 1619–1621. 126 Frison J-C, Palazzi C, Bolm C. Ligand effects in aluminum-catalyzed asymmetric Baeyer-Villiger reactions. Tetrahedron 2006, 62, 6700–6706. 127 Bolm C, Beckmann O, Kuhn T, Palazzi C, Adam W, Rao PB, Saha-Moller CR. Influence of hydroperoxides on the enantioselectivity of metal-catalyzed asymmetric Baeyer-Villiger oxidation and epoxidation with chiral ligands. Tetrahedron: Asymmetry 2001, 12, 2441–2446. 128 Xu S, Wang Z, Zhang X, Zhang X, Ding K. Chiral Brønsted acid catalyzed asymmetric Baeyer- Villiger reaction of 3-substituted cyclobutanones by using aqueous H2O2. Angew. Chem., Int. Ed. 2008, 47, 2840–2843. 129 Malkov AV, Friscourt F, Bell M, Swarbrick ME, Kočovský Pl. Enantioselective Baeyer–Villiger oxidation catalyzed by palladium(II) complexes with chiral P,N-ligands. J. Org. Chem. 2008, 73, 3996–4003. 130 Hughes G, Kimura M, Buchwald SL. Catalytic enantioselective conjugate reduction of lactones and lactams. J. Am. Chem. Soc. 2003, 125, 11253–11258. 131 Bourguignon JJ, Wermuth CG. Lactone chemistry. Synthesis of β-substituted, γ-functionalized butanolides and butenolides and succinaldehydic acids from glyoxylic acid. J. Org. Chem. 1981, 46, 4889–4894. 132 Yan B, Spilling CD. Stereospecific Pd(0)-catalyzed malonate additions to allylic hydroxy phosphonate derivatives: a formal synthesis of (–)-enterolactone. J. Org. Chem. 2004, 69, 2859–2862. 133 Rowe BJ, Spilling CD. The synthesis of 1-hydroxy phosphonates of high enantiomeric excess using sequential asymmetric reactions: titanium alkoxide-catalyzed P-C bond formation and kinetic resolution. Tetrahedron: Asymmetry 2001, 12, 1701–1708. 134 Wiesner M, Revell JD, Tonazzi S, Wennemers H. Peptide catalyzed asymmetric conjugate addition reactions of aldehydes to nitroethylene – a convenient entry into γ2-amino acids. J. Am. Chem. Soc. 2008, 130, 5610–5611. 135 Kamlage S, Sefkow M, Zimmermann N, Peter MG. Concise synthesis of (+)-β-benzyl-γ- butyrolactones from butynediol. Synlett 2002, 77–80. 136 Kamlage S, Sefkow M, Peter MG. Stereoselective cross-coupling of benzylic bromides and vinyl stannanes. J. Org. Chem. 1999, 64, 2938–2940.

103

137 Shao L, Miyata S, Muramatsu H, Kawano H, Ishii Y, Saburi M, Uchida Y. Asymmetric synthesis of (R)- and (S)-4-(substituted benzyl)dihydrofuran-2(3H)-ones: an application of the ruthenium- binap complex–catalysed asymmetric hydrogenation of alkylidenesuccinic acids. J. Chem. Soc. Perkin Tran. 1 1990, 1441–1445. 138 Isemori Y, Kobayashi Y. An approach to β-substituted γ-butyrolactones and its application to the synthesis of lignans. Synlett 2004, 1941–1944. 139 Ito M, Murugesh MG, Kobayashi Y. Regio- and stereoselective installation of alkyl groups onto cis 4-cyclopentene-1,3-diol monoacetate by using reagents derived from alkylmagnesium halides. Tetrahedron Lett. 2001, 42, 423–427. 140 Sibi MP, Liu P, Johnson MD. A short synthesis of both enantiomers of enterolactone. Can. J. Chem. 2000, 78, 133–138. 141 Allais F, Pla TJL, Ducrot P-H. An access to chiral β-benzyl-γ-butyrolactones and its application to the synthesis of enantiopure (+)-secoisolariciresinol, (–)-secoisolariciresinol, and (–)-enterolactone. Synthesis 2011, 1456–1464. 142 Charlton JL, Chee G-L. Asymmetric synthesis of lignans using oxazolidinones as chiral auxiliaries. Can. J. Chem. 1997, 75, 1783–1794. 143 Pohmakotr M, Soorukram D, Tuchinda P, Prabpai S, Kongsaeree P, Reutrakul V. Highly diastereoselective alkylation of vicinal dianions of chiral succinic acid derivatives: a new general strategy to (R)-β-arylmethyl-γ-butyrolactones. Tetrahedron Lett. 2004, 45, 4315–4318. 144 Bennett DJ, Pickering PL, Simpkins NS. A novel asymmetric route to succinimides and derived compounds: synthesis of the lignan lactone (+)-hinokinin. Chem. Comm. 2004, 1392–1393. 145 Sibi MP, Liu P, Ji J, Hajra S, Chen J. Free-radical-mediated conjugate additions. Enantioselective synthesis of butyrolactone natural products: (–)-enterolactone, (–)-arctigenin, (–)-isoarctigenin, (–)-nephrosteranic acid, and (–)-roccellaric acid. J. Org. Chem. 2002, 67, 1738–1745. 146 Sibi MP, Ji J, Sausker JB, Jasperse CP. Free radical-mediated intermolecular conjugate additions. Effect of the Lewis acid, chiral auxiliary, and additives on diastereoselectivity. J. Am. Chem. Soc. 1999, 121, 7517–7526. 147 Kise N, Ueda T, Kumada K, Terao Y, Ueda N. Oxidative homocoupling of chiral 3-arylpropanoic acid derivatives. Application to asymmetric synthesis of lignans. J. Org. Chem. 2000, 65, 464– 468. 148 Kise N, Tokioka K, Aoyama Y, Matsumura Y. Enantioselective synthesis of 2,3-disubstituted succinic acids by oxidative homocoupling of optically active 3-acyl-2-oxazolidones. J. Org. Chem. 1995, 60, 1100–1101. 149 Belletire JL, Fry DF. Oxidative coupling of carboxylic acid dianions: the total synthesis of (±)-hinokinin and (±)-fomentaric acid. J. Org. Chem. 1987, 52, 2549–2555. 150 Baran PS, DeMartino MP. Intermolecular oxidative enolate heterocoupling. Angew. Chem., Int. Ed. 2006, 45, 7083–7086. 151 DeMartino MP, Chen K, Baran PS. Intermolecular enolate heterocoupling: scope, mechanism, and application. J. Am. Chem. Soc. 2008, 130, 11546–11560. 152 Takekawa Y, Shishido K. Selective cleavage of the trisubstituted cyclopropanes via cyclopropylcarbinyl radical fragmentation. Tetrahedron Lett. 1999, 40, 6817–6820. 153 Takekawa Y, Shishido K. Fragmentation reactions of optically active trisubstituted cyclopropylcarbinyl radicals. J. Org. Chem. 2001, 66, 8490–8503. 154 Fukuyama T, Nishitani S, Inouye T, Morimoto K, Ryu I. Effective acceleration of atom transfer carbonylation of alkyl iodides by metal complexes. Application to the synthesis of the hinokinin precursor and dihydrocapsaicin. Org. Lett. 2006, 8, 1383–1386.

104

155 Itoh T, Chika J, Takagi Y, Nishiyama S. An efficient enantioselective total synthesis of antitumor lignans: synthesis of enantiomerically pure 4-hydroxyalkanenitriles via an enzymic reaction. J. Org. Chem. 1993, 58, 5717–5723. 156 Mihovilovic MD, Kapitan P, Rydz J, Rudroff F, Ogink FH, Fraaije MW. Biooxidation of ketones with a cyclobutanone structural motif by recombinant whole-cells expressing 4-hydroxyacetophenone monooxygenase. J. Mol. Catal. B: Enzym. 2005, 32, 135–140. 157 Rudroff F, Rydz J, Ogink FH, Fink M, Mihovilovic MD. Comparing the stereoselective biooxidation of cyclobutanones by recombinant strains expressing bacterial Baeyer–Villiger monooxygenases. Adv. Synth. Catal. 2007, 349, 1436–1444. 158 Rial DV, Bianchi DA, Kapitanova P, Lengar A, van Beilen JB, Mihovilovic MD. Stereoselective desymmetrizations by recombinant whole cells expressing the Baeyer–Villiger monooxygenase from Xanthobacter sp. ZL5: a new biocatalyst accepting structurally demanding substrates. Eur. J. Org. Chem. 2008, 1203–1213. 159 Torres Pazmiño DE, Snajdrova R, Baas B-J, Ghobrial M, Mihovilovic MD, Fraaije MW. Self- sufficient Baeyer–Villiger monooxygenases: effective coenzyme regeneration for biooxygenation by fusion engineering. Angew. Chem., Int. Ed. 2008, 47, 2275–2278. 160 Caro Y, Masaguer CF, Raviña E. Synthesis of optically active β-benzyl-γ-butyrolactone through lipase-catalyzed kinetic resolution. Tetrahedron: Asymmetry 2001, 12, 1723–1726. 161 Morimoto T, Nagai H, Achiwa K. Lipase-catalyzed esterification of a (±)-2,3-di(arylmethyl)-1,4- butanediol and its application to the synthesis of (S,S)-(+)-hinokinin. Synth. Commun. 2005, 35, 857–865. 162 Felluga F, Pitacco G, Valentin E, Venneri CD. A facile chemoenzymatic approach to chiral non- racemic β-alkyl-γ-aminoacids and 2-alkylsuccinic acids. A concise synthesis of (S)-(+)- pregabalin. Tetrahedron: Asymmetry 2008, 19, 945–955. 163 Taneja SC, Kaul S, Singh B, Qazi GN. Novel chemo-enzymatic reduction process for the preparation of optically enriched β-benzyl-γ-butyrolactones. U.S. Pat. Appl. Publ. 2004 US 2004014993, A1 20040122. 164 Ghosh M. Convenient route to enantiopure substituted butyrolactones: application in a formal synthesis of both enantiomers of enterolactone. Tetrahedron 2007, 63, 11710–11715. 165 Banerjee S, Ghosh S, Sinha S, Ghosh S. Convenient route to both enantiomers of a highly functionalized trans-disubstituted cyclopentene. Synthesis of the carbocyclic core of the nucleoside BCA. J. Org. Chem. 2005, 70, 4199–4202. 166 Xia Y, You J, Zhang Y, Su Z. Synthesis, anti-virus and anti-tumour activities of dibenzylbutyrolactone lignans and their analogues. J. Chem. Res. 2009, 565–569. 167 Asano Y, Kamikawa T, Tokoroyama T. A simple method for the synthesis of lignan skeleton. Syntheses of (±)-parabenzlactone and (±)-hinokinin. Bull. Chem. Soc. Jpn. 1976, 49, 3232– 3235. 168 Hajra S, Giri AK. Organocatalytic and enantioselective synthesis of β-(hydroxyalkyl)-γ- butyrolactones. J. Org. Chem. 2008, 73, 3935–3937. 169 Hajra S, Giri AK, Hazra S. Asymmetric syntheses of (–)-enterolactone and (7'R)-7'- hydroxyenterolactone via organocatalyzed aldol reaction. J. Org. Chem. 2009, 74, 7978–7981. 170 Hodgson DM, Talbot EP, Clark BP. Stereoselective synthesis of β-(hydroxymethylaryl/alkyl)-α- methylene-γ-butyrolactones. Org. Lett. 2011, 13, 2594–2597. 171 Trazzi G, Andre MF, Coelho F. Diastereoselective synthesis of β-piperonyl-γ-butyrolactones from Morita-Baylis-Hillman adducts. Highly efficient synthesis of (±)-yatein, (±)-podorhizol and (±)-epi-podorhizol. J. Braz. Chem. Soc. 2010, 21, 2327–2339.

105

172 Masunari A, Ishida E, Trazzi G, Almeida WP, Coelho F. An alternative route to the synthesis of lignans intermediates. Synth. Commun. 2001, 31, 2127–2136. 173 Almeida WP, Coelho F. Piperonal as electrophile in the Baylis-Hillman reaction. A synthesis of hydroxy-β-piperonyl-γ-butyrolactone derivative. Tetrahedron Lett. 1998, 39, 8609–8612. 174 Edlin CD, Faulkner J, Helliwell M, Knight CK, Parker J, Quayle P, Raftery J. Atom transfer radical cyclization reactions (ATRC): synthetic applications. Tetrahedron 2006, 62, 3004–3015. 175 Yamauchi S, Okazaki M, Akiyama K, Sugahara T, Kishida T, Kashiwagi T. First enantioselective synthesis of (–)- and (+)-virgatusin, tetra-substituted tetrahydrofuran lignan. Org. Biomol. Chem. 2005, 3, 1670–1675. 176 Masao M, Yamauchi S. First diastereoselective construction of butane-type and butyrolactone- type secocyclolignane structures. Biosci., Biotechnol., Biochem. 2009, 73, 2445–2451. 177 Jiao X-Z, Jiang Y-J, Feng W-H, Xie P, Liang X-T. Total synthesis of enantiomer of (–)-α- Conidendrin. Chin. J. Synth. Chem. 2007, 15, 34–37. 178 Jiao X-Z, Jiang Y-J, Xie P, Liang X-T. Total Synthesis of (–)-7(R)-Hydroxymatairesinol. Chin. J. Org. Chem. 2007, 27, 1537–1541. 179 Enders D, Kirchhoff J, Lausberg V., Diastereo- and enantioselective synthesis of lignan building blocks by tandem Michael addition/electrophilic substitution of lithiated α-amino nitriles to furan- 2(5H)-one. Liebigs Ann. 1996, 1361–1366. 180 Enders D, Lausberg V, Del Signore G, Berner OM. A general approach to the asymmetric synthesis of lignans: (–)-methyl piperitol, (–)-sesamin, (–)-aschantin, (+)-yatein, (+)- dihydroclusin, (+)-burseran and (–)-isostegane. Synthesis 2002, 515–522. 181 Enders D, Milovanović M. Asymmetric synthesis of (+)-hinokinin, (+)-dihydrocubebin and cubebin dimethyl ether, a new lignan from Phyllanthus niruri. Z. Naturforsch. 2007, 62b, 117– 120. 182 Yamauchi S, Yamamoto N, Kinoshita Y. Improved stereoselective synthesis of optically active methylene lactone, key intermediate for the synthesis of 1,2-oxidized furofuran lignan, by direct α-methylenation to butanolide. Biosci., Biotechnol., Biochem. 2000, 64, 2209–2215. 183 Yamauchi S, Yamaguchi M. Synthesis of (+)-aptosimon, a 4-oxofurofuran lignan, by erythro selective aldol condensation and stereoconvergent cyclization as the key reactions. Biosci., Biotechnol., Biochem. 2003, 67, 838–846. 184 Yamauchi S, Kinoshita Y. Synthesis of optically active olivil type of lignan from L-arabinose using threo-selective aldol condensation as a key reaction. Biosci., Biotechnol., Biochem. 2000, 64, 2320–2327. 185 Yamauchi S, Kinoshita Y. Synthesis of cis-lactone lignan, cis-(2S,3R)-parabenzlactone, from L-arabinose. Biosci., Biotechnol., Biochem. 2001, 65, 1669–1672. 186 Ziegler FE, Schwartz JA. Synthetic studies on lignan lactones: aryl dithiane route to (±)- podorhizol and (±)-isopodophyllotoxone and approaches to the stegane skeleton. J. Org. Chem. 1978, 43, 985–991. 187 Pelter A, Ward RS, Satyanarayana P, Collins P. Synthesis of lignan lactones by conjugate addition of thioacetal carbanions to butenolide. J. Chem. Soc. Perkin Trans. 1 1983, 643–647. 188 Yoda H, Suzuki Y, Matsuura D, Takabe K. A new synthetic entry to furofuranoid lignans, methyl piperitol and fargesin. Heterocycles 2005, 65, 519–522. 189 Takahashi M, Takada K, Matsuura D, Takabe K, Yoda H. First total syntheses of new phenylpropanoid lignans, (α)-aglacin K stereoisomer and analogue of (α)-arborone. Heterocycles 2007, 71, 2113–2118. 190 Yoda H, Mizutani M, Takabe K. Stereoselective synthesis of tetrasubstituted furan from dihydroxyacetone dimer. Synlett 1998, 855–856.

106

191 Yoda H, Kimura K, Takabe K. Asymmetric syntheses of trisubstituted tetrahydrofuran lignans, sesaminone and 4-epidihydrosesamin. Synlett 2001, 400–402. 192 Yoda H, Mizutani M, Takabe K. First total synthesis of tetrasubstituted tetrahydrofuran lignan, (–)-virgatusin. Tetrahedron Lett. 1999, 40, 4701–4702. 193 Uchiyama M, Kimura Y, Ohta A. Stereoselective total syntheses of (±)-arthrinone and related natural compounds. Tetrahedron Lett. 2000, 41, 10013–10017. 194 Eklund PC, Willför SM, Smeds AI, Sundell FJ, Sjöholm RE, Holmbom BR. A new lariciresinol- type butyrolactone lignan derived from hydroxymatairesinol and its identification in spruce wood. J. Nat. Prod. 2004, 67, 927–931. 195 Eklund PC, Sundell FJ, Smeds AI, Sjöholm RE. Reactions of the natural lignan hydroxymatairesinol in basic and acidic nucleophilic media: formation and reactivity of a quinone methide intermediate. Org. Biomol. Chem. 2004, 2, 2229–2235. 196 Moritani Y, Fukushima C, Ogiku T, Ukita T, Miyagishima T, Iwasaki T. A highly stereoselective synthesis of α-substituted cis-α,β-dibenzyl-γ-butyrolactones. Tetrahedron Lett. 1993, 34, 2787– 2790. 197 Moritani Y, Fukushima C, Ukita T, Miyagishima T, Ohmizu H, Iwasaki T. Stereoselective syntheses of cis- and trans-isomers of α-hydroxy-α,β-dibenzyl-γ-butyrolactone lignans: new syntheses of (±)-trachelogenin and (±)-guayadequiol. J. Org. Chem. 1996, 61, 6922–6930. 198 van Oeveren A, Jansen JFGA, Feringa BL. Enantioselective synthesis of natural dibenzylbutyrolactone lignans (–)-enterolactone, (–)-hinokinin, (–)-pluviatolide, (–)-enterodiol, and furofuran lignan (–)-eudesmin via tandem conjugate addition to γ-alkoxybutenolides. J. Org. Chem. 1994, 59, 5999–6007. 199 Pelter A, Ward RS, Jones DM, Maddocks P. Asymmetric synthesis of lignans of the dibenzylbutanediol and tetrahydrodibenzocyclooctene series. J. Chem. Soc. Perkin Trans. 1 1993, 2631–2637. 200 Feringa BL, de Lange B, de Jong JC. Synthesis of enantiomerically pure γ-(menthyloxy)butenolides and (R)- and (S)-2-methyl-1,4-butanediol. J. Org. Chem. 1989, 54, 2471–2475. 201 Moritani Y, Ukita T, Hiramatsu H, Okamura K, Ohmizu H, Iwasaki T. A highly stereoselective synthesis of 3-hydroxy-1-aryltetralin lignans based on the stereoselective hydroxylation of α,β- dibenzyl-γ-butyrolactones: the first synthesis of (±)-cycloolivil. J. Chem. Soc. Perkin Trans. 1 1996, 2747–2753. 202 Pelter A, Ward RS, Storer NP. The asymmetric synthesis of an isomer of podophyllotoxin. Tetrahedron 1994, 50, 10829–10838. 203 Van Speybroeck R, Guo H, Van der Eycken J, Vandewalle M. Enantioselective total synthesis of (–)-epipodophyllotoxin and (–)-podophyllotoxin. Tetrahedron 1991, 47, 4675–4682. 204 Greene TW, Wuts PGM. Protective groups in organic synthesis, 3rd Edition, John Wiley & Sons, Canada, Ltd., 1999. 205 Kocieński PJ. Protecting groups. 2nd Edition, Thieme, Stuttgart, 2000. 206 Khan AT, Mondal E, Sahu PR. A highly efficient procedure for regeneration of carbonyl groups from their corresponding oxathioacetals and dithioacetals using sodium nitrite and acetyl chloride in dichloromethane. Synlett 2003, 377−381. 207 Langille NF, Dakin LA, Panek JS. A mild, chemoselective protocol for the removal of thioketals and thioacetals mediated by Dess−Martin periodinane. Org. Lett. 2003, 5, 575–578. 208 Kirihara M, Harano A, Tsukiji H, Takizawa R, Uchiyama T, Hatano A. Deprotection of dithioacetals using the tantalum(V) chloride catalyzed oxidation of iodide ion by hydrogen peroxide. Tetrahedron Lett. 2005, 46, 6377–6380.

107

209 Stork G, Zhao K. A simple method of dethioacetalization. Tetrahedron Lett. 1989, 30, 287−290. 210 Li Manni G, Barone G, Duca D, Murzin DY. Systematic conformational search analysis of the SRR and RRR epimers of 7-hydroxymatairesinol. J. Phys. Org. Chem. 2010, 23, 141–147. 211 Stevens DR, Whiting DA. Stereospecific synthesis of the 2,3-trans-3,4-cis trisubstituted tetrahydrofuran lignan (±)-dihydrosesamin. J. Chem. Soc. Perkin Trans. 1 1992, 633−637. 212 Kiuru P, Wähälä K. Unpublished results. 213 Schwartz H, Sontag G, Plumb J. Inventory of phytoestrogen databases. Food Chem. 2009, 113, 736−747. 214 Mäkelä TH, Wähälä KT, Hase TA. Synthesis of enterolactone and enterodiol precursors as potential inhibitors of human estrogen synthetase (aromatase). Steroids 2000, 65, 437−441. 215 Leppälä E, Pohjoispää M, Koskimies J, Wähälä K. Synthesis of new deuterium–labeled lignanolactones. J. Label. Compd. Radiopharm. 2008, 51, 407–412.

108 Appendix Appendix 1. Aryl substituents present in the synthetic procedures (Chapter 3, 4, and 5).

Ar1= Ar2= Ar3= Ar4= Ar5= OH OMe OBn O O

10 Ar6=Ar7= Ar8= Ar9= Ar = OH OMe OH OMe OMe OH OH OMe OMe OBn

Ar11= Ar12= Ar13= Ar14= OMe MeO OMe MeO OH MeO OMe OTBDMS OH OMe OMe

15 16 17 Ar = Ar = Ar = Ar18= MeO OMe O O OMe OBn OMe OMe O

OMe CO2Et Ar19= Ar20= Ar21= Ar22= O MeO OH t-Bu OMe OTBDMS O OH

26 27 Ar23= Ar24= Ar25= Ar = Ar = OTBDMS OMe BnO OBn OBn OBn OMe OAc OBn OMe

OMe Ar28= Ar29= Ar30= OMe MeO OMe HO OH

109 Appendix 2. 1H NMR data of lignanolactones (selected molecules). Only signals of lignan framework (2-8; 2'-9') are reported.

Table 7. 1H NMR data of (7'S,8R,8'R)-7'-hydroxylignano-9,9'-lactones. OH H Ar O H Ar' O I,94 I,94 II,197 II 93 (7'S)-27 (7'S)-15 (7'S)-124 (7'S)-125f I (7'S)-7 (7'S)-25 (7'S)-7'- (7'S)-7'- (7'S)-7'- (7'S)-7'- (7'S)-7'- (7'S)-parabenzlactone hydroxyarctigenin hydroxymatairesinol hydroxybursehernin hydroxyprestegane B hydroxyenterolactone Ar=Ar9, Ar'=Ar7 Ar=Ar'=Ar7 Ar=Ar9, Ar'=Ar5 Ar=Ar'=Ar8 Ar=Ar'=Ar5 Ar=Ar'=Ar2 24 [α] D -23.9° (c 0.56, CHCl3) -9.3° (c 4.0, THF) -20.09° (c 0.34, CHCl3) -1.5° (c 0.12, CHCl3) -1.5° (c 0.12, CHCl3) -1.32° (c 0.31, Acetone) MHz, 500 MHz, 500 MHz, 500 MHz, 500 MHz, 500 MHz, 300 MHz, 300 MHz, Solvent CDCl3 CDCl3 CDCl3 CDCl3 CDCl3 acetone-d6 acetone-d6 δ; multipl.; J δ; multipl.; J Position δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) (Hz) (Hz) 7a 2.90-2.96; m 2.89-2.95; m 2.92-2.97; m 2.89-2.95; m 2.94-3.00; m 2.76-2.90, m 2.71; dd; 13.7, 5.0 7b 3.00-3.08; m 3.01; dd; 15.3, 7.7 2.92-2.97; m 3.01; dd; 15.5, 8.0 2.94-3.00; m 2.76-2.90, m 3.02; dd; 13.7, 4.7 2.90; dd; 12.0, 8 2.90-2.96; m 2.89-2.95; m 2.89; dd; 13.0, 5.5 2.89-2.95; m 2.76-2.90, m 2.99; m 6.0 4.61; dd; 6.2, 7' 4.66; dd; 6.0, 1.5 4.63; d; 6.6 4.64; dd; 6.5, 1.5 4.57; dd; 6.7, 2.3 4.75; m 4.75; d; 5.7 2.8 8' 2.59-2.67; m 2.56-2.63; m 2.56-2.62; m 2.60-2.66; m 2.54-2.60; m 2.62; m 2.68; m 4.05; dd; 9.0, 9'a 3.91; app. t; 8.8 3.88-3.98; m 3.95; app. t; 8.8 3.85-3.88; m 3.91; m 3.88; t; 8.1 7.8 4.09; dd; 9.0, 9'b 3.99; dd; 9.5, 7.0 3.88-3.98; m 3.99; dd; 9.0, 6.5 3.89-3.92; m 3.91; m 4.02; dd; 8.1, 7.8 6.9 Aromatic 6.61; dd; 8.0, 2.0 6.60; dd; 7.9, 1.9 6.54-6.56; m, 2H 6.66; dd; 8.5, 2.0 6.58; dd; 8.0, 6.55-6.58; m 6.57; dt; 7.3, 1.2 protons 6.65; d; 2.0 6.61; s 6.65-6.69, m, 2H 6.72; d; 2.0 1.5 6.66-6.69; m, 6.67; ddd; 7.8, 1.9, 1.0 6.73; d; 2.0 6.67; d; 1.8 6.78; dd; 8.5, 1.5 6.73-6.75; m, 2H 6.61; d; 1.0 2H 6.69; t; 1.2 6.75; dd; 8.5, 2.0 6.71; dd; 8.1, 1.8 6.81; d; 8.5 6.79; d; 2.0 6.67; dd; 8.0, 6.75-6.80; m, 6.76; ddd; 8.1, 2.4, 1.0 6.79; d; 8.0 6.78; d; 7.8 6.80; d; 8.5 1.0 3H 6.82; dt; 7.6, 1.2 6.81; d; 8.0 6.86; d; 8.0 6.68; d; 8.0 6.91; dd; 2.0, 1.9 6.70; d; 1.5 7.06; t; 8.6 6.74; d; 8.0 7.17; t; 7.8

110 Table 8. 1H NMR data of (7'R,8R,8'R)-7'-hydroxylignano-9,9'-lactones. OH H Ar O H Ar' O 169 I,21,94 I,21,94 93,169 (7'R)-144 (7'R)-27 (7'R)-15 (7'R)-7 (7'R)-7'-hydroxyenterolactone (7'R)-7'-hydroxyarctigenin (7'R)-7'-hydroxymatairesinol (7'R)-7'-hydroxyenterolactone dimethyl ether Ar=Ar9, Ar'=Ar7 Ar=Ar'=Ar7 Ar=Ar'=Ar2 Ar=Ar'=Ar3 24 [α] D +8.8° (c 1.4, EtOH) -9.8° (c 4.0, THF) -13.04° (c 1.5, Acetone) –13.98° (c 3.0, CHCl3) MHz, 500 MHz, 400 MHz, 300 MHz, 200 MHz, Solvent CDCl3 CDCl3 Acetone-d6 CDCl3 Position δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) 7a 2.65; ddd; 6.5, 6.5, 6.5 2.77; dd; 13.8, 5.2 2.76; dd; 13.7, 5.5 2.92-2.78; m 7b 2.76; s 2.70; dd; 13.8, 7.9 2.85; dd; 13.7, 6.5 2.92-2.78; m 8 2.75; s 2.64-2.59; m 2.95; ddd; 8.7, 6.5, 5.5 2.92-2.78; m 7' 4.41; d; 6.8 4.38; d; 7.8 4.51; d; 4.8 4.39; d; 5.8 8' 2.53; dddd; 7.8, 6.8, 6.7, 6.5 2.54-2.47; m 2.64; qd; 8.7, 4.8 2.59-2.52; m 9'a 4.13; dd; 9.3, 7.8 4.16; dd; 9.5, 7.6 3.90; t; 8.7 4.0; dd; 9.2, 8.0 9'b 4.39; dd; 9.3, 6.7 4.42; dd; 9.5, 5.6 4.21; t; 8.7 4.32; dd; 9.4, 7.4 Aromatic 6.78; d; 8.3, 2H 6.43; d; 1.9 6.67; dm; 7.8 7.27-7.14; m, 2H protons 6.70; d; 8.3 6.48; dd; 8.0, 1.9 6.70; dm; 7.8 6.84-6.63; m, 6H 6.67; dd; 8.3, 1.5 6.55; d; 1.9 6.73; dm; 7.8 6.63; m 6.62; dd; 8.1, 1.9 6.76; dm; 7.8 6.52-6.47; m, 2H 6.76; d; 7.9 6.77; m 6.82; d; 8.0 6.86; t; 2.0 7.11; t; 7.8 7.15; t; 7.8

111 Table 9. 1H NMR data of 7',8'-trans-8,8'-trans-lignano-9,7'-lactones, (7'S,8S,8'R). Ar H H RO O

H Ar' O 113a88,197,II,III 113b197,II 113cII 132I 127II 133a197 133b197 133c197 Ar=Ar'=Ar5 Ar=Ar'=Ar5 Ar=Ar9, Ar'=Ar5 Ar=Ar'=Ar7 Ar=Ar'=Ar7 O Ar=Ar9, Ar'=Ar5 Ar=Ar5, Ar'=Ar9 Ar=Ar9, Ar'=Ar10

R=H R=H R=H R=TBDMS R= R=MOM R=MOM R=MOM O2N NO2 +47.45° (c 1.08 +52.68° (c 0.2, +36.83° (c 0.57, +66.81° (c 0.31, [α]24 +90.14° (c 0.4 CHCl ) (±) (±) (±) D 3 THF) THF) THF) THF) MHz, 500 MHz, 500 MHz, 500 MHz, 500 MHz, 500 MHz, n.g., CDCl3 n.g., CDCl3 n.g., CDCl3 solvent CDCl3 CDCl3 CDCl3 CDCl3 CDCl3 Position δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) 7a 3.15; dd; 14.0, 5.0 2.91-3.21; m 3.14-3.05; m 3.05; d; 5.5 3.22; dd; 14.0, 4.0 2.96-3.21; m 2.94-3.20; m 3.00-3.22; m 7b 2.95; dd; 14.0, 7.0 2.91-3.21; m 3.14-3.05; m 3.05; d; 5.5 3.00; dd; 14.0, 7.5 2.96-3.21; m 2.94-3.20; m 3.00-3.22; m 3.06; ddd; 12.0, 7.0, 8 2.91-3.21; m 3.14-3.05; m 3.15-3.10; m 2.91-2.87; m 2.96-3.21; m 2.94-3.20; m 3.00-3.22; m 5.0 7' 5.13; d; 8.5 5.16; d; 9.2 5.11; d; 9.5 5.11; d; 9.0 5.04; d; 9.0 5.15; d; 9.2 5.12; d; 9.1 5.13; d; 10.7 8' 2.28-2.22, m 2.18-2.39; m 2.32-2.27; m 2.26- 2.21; m 2.82-2.76; m 2.12-2.40; m 2.19-2.37; m 2.20-2.38; m 3.44; dd; 10.3, 3.42; dd; 10.6, 3.44; dd; 10.3, 9'a 3.55; dd; 3.5, 11.0 3.40-3.65; m 3.58; dd; 3.5, 11.0 3.49; dd; 3.5, 10.5 4.23; d; 6.0 3.7 2.8 3.6 3.30; dd; 10.3, 3.27; dd; 10.3, 3.28; dd; 10.3, 9'b 3.48; dd; 4.0, 11.0 3.40-3.65; m 3.50; dd; 4.0, 11.0 3.41; dd; 3.5, 10.5 4.23; d; 6.0 3.8 3.9 3.8 Aromatic 6.60; d; 1.5 6.56-6.88; m, 6H 6.49; d; 2.0 6.46; d; 2.0 6.59; d; 1.5 6.58-6.86; m, 6H 6.46-6.84; m, 6H 6.50-6.87; m, 6H protons 6.66; dd; 8.0, 1.5 6.69; dd; 8.0, 1.5 6.66; dd; 8.0, 2.0 6.63; d; 7.5 6.69; dd; 7.5, 1.5 6.70; dd; 8.0, 2.0 6.67; dd; 8.0, 2.0 6.68; dd; 7.5, 1.5 6.72; d; 1.5 6.81; d; 1.5 6.79; d; 2.0 6.71; dd; 7.5, 1.5 6.73; d; 7.5 6.83; d; 8.0 6.81; d; 8.0 6.74; d; 1.5 6.75; d; 8.0 6.84; d; 8.0 6.84; d; 8.0 6.75; d; 7.5

112 Table 10. 1H NMR data of 7',8'-trans-8,8'-cis-lignano-9,7'-lactones, (7'S,8R,8'R). Ar H H HO O

H Ar' O 126aII 126bII 126fI 110194 Ar=Ar'=Ar5 Ar=Ar9, Ar'=Ar5 Ar=Ar'=Ar11 Ar=Ar'=Ar7 24 [α] D +49.3° (c 0.3, THF) +44.47° (c 0.54, THF) +37.5° (c 0.3, CHCl3) +51.0° (c 0.005, THF)

MHz, solvent 500 MHz, CDCl3 500 MHz, CDCl3 500 MHz, CDCl3 500 MHz, CDCl3 Position δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) 7a 3.22; dd; 15.0, 5.0 3.20; dd; 15.0,5.0 3.23; dd; 15.1, 4.5 3.16; dd; 15.0, 5.0 7b 2.73; dd; 15.0, 10.5 2.74; dd; 15.0, 11.0 2.75; dd; 15.1, 10.6 2.76; dd; 15.0, 10.2 8 3.04; ddd; 10.5, 8.0, 4.5 3.04; ddd; 11.0, 8.5, 5.0 3.09; ddd; 10.6, 8.2, 4.5 3.05; ddd; 10.2, 8.1, 5.0 7' 5.49; d; 2.0 5.53; d; 2.0 5.45; d; 2.6 5.46; d; 2.6 8' 2.56; ddd; 7.0, 4.5, 2.0 2.62-2.58; m 2.59-2.63; m 2.6; dddd; 8.1, 6.9, 4.4, 2.6 9'a 3.95; dt; 10.5, 4.0 3.96; dd; 11.0, 4.5 3.93; dd; 10.5, 4.0 3.94; dd; 10.7, 4.4 9'b 3.77; ddd; 10.5, 7.0, 5.5 3.78; dd; 11.0, 7.5 3.73-3.78; m 3.76; dd; 10.7, 6.9 Aromatic 6.66; dd; 8.0, 1.5 6.65; dd; 8.0,1.5 6.65; dd; 8.0, 2.0 6.67; dd; 8.1, 2.1 protons 6.69; d; 1.5 6.69; d; 1.5 6.71-6.73; m, 2H 6.72; d; 2.1 6.72; d; 8.0 6.71; d; 8.0 6.74-6.77; m, 2H 6.75; ddd; 8.1, 2.0, 0.8 6.75; dd; 8.0, 1.5 6.78; d; 2.0 6.82; d; 8.1 6.77; d; 2 6.76; d; 1.5 6.82; dd; 8.0, 2.0 6.80; d; 8.1 6.80; d; 8.0 6.85; d; 8.0 6.88; d; 8.1

Table 11. 1H NMR data of 7',8'-cis-8,8'-trans-lignano-9,7'-lactones, (7'R,8S,8'R) and (7'S,8R,8'S). H Ar H Ar H H RO RO O O

H H Ar' O Ar' O (−)-115185 (−)-114185 (+)-11531 (+)-11438 Ar=Ar'=Ar5 Ar=Ar'=Ar5 Ar=Ar'=Ar10 Ar=Ar'=Ar10

R=H R=Piv R=H R=Piv 24 [α] D -42.1° (c 1.07, CHCl3) -46.8° (c 1.52, CHCl3) +20.0° (c 1.6, CHCl3) +30.0° (c 0.6, CHCl3)

MHz, solvent 400 MHz, CDCl3 400 MHz, CDCl3 400 MHz, CDCl3 400 MHz, CDCl3 Position δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) 7a 3.08; dd; 14.2, 5.3 3.07; dd; 14.2, 5.1 3.08; dd; 14.0, 5.5 3.08; dd; 14.2, 6.8 7b 2.99; dd; 14.2, 6.8 3.00; dd; 14.2, 7.3 3.02; dd; 14.0, 5.5 3.02; dd; 14.2, 4.9 8 2.90; m 2.87; m 2.94; m 2.87; m 7' 5.38; d; 7.8 5.38; d; 7.3 5.32; d; 7.7 5.32; d; 6.8 8' 2.67; m 2.79; m 2.64; m 2.79; m 9'a 3.31; m 3.60; dd; 11.7, 6.4 3.27; m 3.77; dd; 11.5, 5.9 9'b 3.21; m 3.62; dd; 11.7, 6.4 3.16; m 3.62; dd; 11.5, 6.4 Aromatic 6.69-6.80; m, 6H 6.66-6.72; m, 4H 6.76-6.87; m, 4H 6.66; d; 8.3 protons 6.75-6.78; m, 2H 6.67-6.72; m, 2H 6.72-6.75; m, 2H 6.82-6.85; m, 3H

113 Table 12. 1H NMR data of 7',8'-cis-8,8'-cis-lignano-9,7'-lactones, (7'R,8R,8'R). H H Ar RO O

H Ar' O 108b194 (±)-134211 Ar=Ar'=Ar7 Ar=Ar'=Ar5

R=H R=TBDMS 24 [α] D +59.8 (c 0.005, THF) (±)

MHz, solvent 500 MHz, CDCl3 400 MHz, CDCl3 Position δ; multipl.; J (Hz) δ; multipl.; J (Hz) 7a 3.08; d; 5.8 2.84; dd; 15.5, 10.5 7b 3.08; d; 5.8 3.35; dd; 15.5, 3.4 8 3.12; ddd; 10.2, 5.8, 5.8 3.10; m 7' 5.12; d; 9.3 5.43; d; 5.4 8' 2.30; dddd; 10.2, 9.3, 4.0, 4.1 2.52; m 9'a 3.50; ddd; 11.0, 4.1, 3.9 3.28; d; 10.8 9'b 3.57; ddd; 11.0, 4.0, 3.9 3.57; dd; 10.8, 2.6 Aromatic 6.49; d; 2.0 6.71-6.79; m, 5H protons 6.69; dd; 8.0, 2.0 6.93; s 6.70; ddd; 8.1, 2.0, 0.4 6.81; d; 2 6.82; d; 8.1 6.83; d; 8.1

Table 13. 1H NMR data of 7',8'-trans-8,8'-cis-8-alkyl-lignano-9,7'-lactones, (7'S*,8R*,8'R*).

H Ar H RO O

R' Ar' O 128aII 128b197 128c197 Ar=Ar'=Ar5 Ar=Ar'=Ar5 Ar=Ar'=Ar5 R=H R=MOM R=MOM R'=Me R'=Me R'=Et 24 [α] D (±) (±) (±)

MHz, solvent 500 MHz, CDCl3 CDCl3 CDCl3 Position δ; multipl.; J (Hz) δ; multipl.; J (Hz) δ; multipl.; J (Hz) 7a 2.99; d; 13.5 3.29; d; 13.6 3.38; d; 13.6 7b 2.85; d; 13.5 2.71; d; 13.6 2.62; d; 13.5 8 – – – 7' 4.84; d; 10.5 4.83; d; 10.2 4.81; d; 10.5 8' 2.44; ddd; 10.5, 8.0, 5.5 2.51-2.71; m 2.63-2.79; m 9'a 4.01; dd; 11.5, 8.0 3.61-3.76; m 3.74; dd; 10.9, 9.6 9'b 3.84; dd; 11.5, 5.5 3.44; dd; 9.8, 4.6 3.43; dd; 9.8, 4.6 1.16; t; 7.5 R' 1.37; s 1.41; s 1.69-2.03, m Aromatic 6.67; dd; 1.5, 8.0 6.26; d; 2.0 6.21; d, 1.9 protons 6.71; d; 1.5 6.62; dd; 8.1, 1.9 6.60; dd; 8.1, 2.0 6.75; m, 2H 6.72; s 6.72; d; 8.5 6.78; d; 8.0 6.74; s, 2H 6.75; br s, 2H 6.81; d; 1.5 6.78; s 6.81; br s

114 Appendix 3. Table 14. Summary of the outcomes of the basic treatments of hydroxylignanolactones.

Starting material Reagents Product(s) Outcome Ar H H 2% KOH/MeOH HO O OH translactonization and H Starting material H α-epimerization Ar Ar' O O NaH/DMF 113a H Ar=Ar'=Ar5 Ar' O 1M aq. NaOH/EtOH (7'S)-25 5 Ar=Ar'=Ar 2% KOH/t-BuOH Starting material no reaction

2% aq. KOH/MeOH OH Ar Ar H H H H H Ar HO TBDMSO O O O translactonization, α-epimerization, NaH/DMF H H H aromatic silyl cleavage and Ar' O Ar' O Ar' O silyl migration

(7'S)-112b 113c 132 7 7 Ar=Ar'=Ar11 Ar=Ar'=Ar Ar=Ar'=Ar OH Ar H H H Ar 2% KOH/MeOH HO O O translactonization and Starting material H H α-epimerization Ar' O Ar' O NaH/DMF (7'S)-124 113f 5 9 Ar=Ar5, Ar'=Ar9 Ar=Ar , Ar'=Ar OH H Ar Ar O H translactonization and H 2% KOH/MeOH Starting material HO H α-epimerization O Ar' O

(7'S)-25 H 5 Ar' O Ar=Ar'=Ar

113a 1M aq. NaOH/EtOH Ar=Ar'=Ar5 2% KOH/t-BuOH Starting material no reaction

2% aq. KOH/MeOH OH H Ar H 2% KOH/MeOH H Ar HO O O NaH/DMF Starting material translactonization Me Me Ar' O Ar' O

128a 1M aq. NaOH/EtOH 131a 5 Ar=Ar'=Ar5 Ar=Ar'=Ar

115 Errata

Page 17. In the last paragraph the complete name of the lignan in Fig. 6 should be (8β,8'α)- lignano-9,9'-lactone, and not (8α,8'β). Page 18. Fig. 6 should be as below.

O O H H β α Ar Ar

We apologize for these oversights.