Stereodivergent synthesis of oligo-1,2-naphthylenes and studies towards the total synthesis of naphthacemycin B1

Stereodivergent Synthesis of Oligo-1,2-naphthylenes and

Studies Towards the Total Synthesis of Naphthacemycin B

Inauguraldissertation

zur Erlangung der Würde eines Doktors der Philosophie vorgelegt der

Philosophisch-Naturwissenschaftlichen Fakultät der Universität Basel

von

Dominik Lotter aus Luzern (LU), Schweiz

Basel, 2019

Originaldokument gespeichert auf dem Dokumentenserver der Universität Basel edoc.unibas.ch

Genehmigt von der Philosophisch-Naturwissenschaftlichen Fakultät auf Antrag von

Prof. Dr. Christof Sparr Prof. Dr. Marcel Mayor

Basel, den 21. Mai 2019

Prof. Dr. Martin Spiess Dekan

Für meine Eltern

i

I. Acknowledgements

At first, I want to thank my “Doktorvater” and mentor Prof. Dr. Christof Sparr for the opportunity to work in his group. I highly appreciate not only the atmosphere he creates – which truly allows improving and growing as a scientist – but also his confidence in me and the freedom I experienced during the last four years. The very interesting and challenging projects have been a very wonderful confirmation thereof and resulted in an unforgettable time in his research group.

I would like to acknowledge Prof. Dr. Marcel Mayor for kindly accepting the co-examination of this thesis, as well as Prof. Dr. Oliver S. Wenger for chairing the defense.

Special thanks go to Reto M. Witzig for almost ten years of friendship throughout the entire chemistry studies. The countless discussions about chemistry had undoubtedly a major influence on my projects. Nevertheless, equally important for the success of this thesis were the numerous discussions not about chemistry.

I very much appreciate the efforts Felix C. Raps, Mirjam Schreier, Reto M. Witzig and Isabelle Lotter invested into proof-reading the thesis, not hesitating to say, when a paragraph had to be rewritten.

I would like to thank Fabian Bissegger and Pascale Welsch for their important and fruitful contributions to the oligo-1,2-naphthylene project during their Master internships.

I thank the whole Sparr-Group, former and present members, for many years of great discussions, very valuable group-meetings, help and advice and unforgettable Friday-beers, Christmas-dinners and group-outings.

Without Dr. Elias Kaufmann, Dr. Christophe Daeppen and Dr. Manuel Scherrer this thesis would not have been possible. I am very grateful for all the things I learned from them during my internship and in the beginning of my PhD studies.

PD Dr. Daniel Häussinger, Dr. Thomas Müntener, Daniel Joss and Raphael Vogel did a wonderful job maintaining the NMR facilities and I am especially grateful to Thomas for all the instructions and advice I obtained during the studies of the “not so easy to analyze” oligomers.

Without the help of a team of analytical chemists this work would not have been possible. For their efforts I thank Dr. Markus Neuburger, Dr. Heinz Nadig, Sylvie Mittelheisser and Dr.

Michael Pfeffer.

ii

I would furthermore like to acknowledge the people, who keep the department running: the secretary team, especially Marina Mambelli Johnson, the “Werkstatt”-team and Oliver Ilg.

Warm thanks go to the former and present members of the department, where I spent the last nine years, for friendship, apéros, discussions, shared disappointment and success.

For being a second family I thank the ”Elite“. Ritz, Päscu, Limi, Luki, Schtindli, Märi and Brünu have always been there, when necessary. Our friendship is, and has always been, an anchor in my life.

I am extremely grateful to my sisters Isa and Claudia, for their support, patience and love in hard and good times.

Finally, I want to thank my parents, Marcel and Ursi Lotter. Without their support over almost three decades, I would not have been able to be, where I am today. Not only their financial but also their mental support allowed me to do and finish these studies. For their support and love, I will always be grateful!

iii

II. Table of Contents

I. Acknowledgements i

II. Table of Contents iii

III. Abstract v

IV. Zusammenfassung ix

V. Publications xiii

VI. Presentations xiv

1. Introduction 1

1.1. Biosynthesis of Polyketides 2

1.2. Atropisomeric Compounds 5

1.3. Synthesis of Atropisomeric, Aromatic Compounds 16

1.4. Oligo-Arylenes 25

1.5. Diastereoselectivity and Stereodivergent Synthesis 29

2. Objectives 33

2.1. Oligo-1,2-naphthylenes 33

2.2. Total Synthesis of Naphthacemycin B1 34

3. Oligo-1,2-naphthylenes 35

3.1. Starting Point 35

3.2. Stereodivergent, Diastereoselective Arene-Forming Aldol Condensation 39 3.3. Subsequent Substitution, Applications and Concepts 62

3.4. Conclusion and Outlook 69

4. Total Synthesis of Naphthacemycin B1 71

4.1. Retrosynthetic Analysis 71

4.2. Core-Structure 75

4.3. (Enolendo)-Exo-Trig Cyclization 85

4.4. Preliminary Exploration of a Cross Coupling Strategy 108

4.5. Conclusion and Outlook 109

5. Conclusion 113

6. Supporting Information 115

6.1. General Information 115

6.2. Experimental – Oligo-1,2-naphthylenes 116

6.3. Experimental – Total Synthesis of Naphthacemycin B1 148

iv

6.4. Evaluation of Catalysts and Conditions 189

6.5. X-ray Crystallographic Analysis (by Dr. Markus Neuburger) 211

6.6. HPLC-Data 214

6.7. 2D-NMR Data 217

6.8. NMR Spectra 218

6.9. List of Abbreviations 309

6.10. Literature 312

v

III. Abstract

Nature is taking advantage of the complex polyketide synthase machinery to assemble, cyclize and modify numerous polyketide natural products.^[1] Many of these compounds show antibiotic properties, making them interesting targets for studies of their biosynthesis, bioactivity and also potential total syntheses. Among these natural polyketides many atropisomeric compounds can be found, of which some have a unique way of forming the stereogenic axis. The involved key step in the biosynthesis of this rare class of atropisomeric natural products was proposed to be atroposelective arene-forming aldol condensation rather than the often-found oxidative dimerization (Scheme 1).^[2] A synthetic counterpart of nature’s aldol condensation was developed in our group, enabling the synthesis of various compounds with a configurationally stable axis.^[3] Besides other scaffolds, oligo-1,2-naphthylenes with two consecutive axes were synthesized with an enantioselective and a substrate-controlled diastereoselective arene- forming aldol condensation.^[4] These structurally interesting oligomers can be valuable scaffolds for placing substituents in a precise and predictable spatial relationship.

Scheme 1: Atroposelective arene-forming aldol condensation in nature and its synthetic counterpart from our group.

Stereodivergent Synthesis of Oligo-1,2-naphthylenes

Due to the configurationally stable axes between the subunits of oligo-1,2-naphthylenes, synthetic issues arise from the exponential growth of potential diastereomers. In this thesis, we therefore addressed this issue and aimed for the development of a diastereodivergent, atroposelective arene-forming aldol condensation, enabling the synthesis of different diastereoisomers of oligo-1,2-naphthylenes with different lengths.^[5]

OH

O

S-Enz O O

OH Me OH

O OH HO

OH

OH

Me OH

OH HO

OH O

Proposed key step in the biosynthesis of fasamycin A:

O S-Enz

CHO

R R’ O

CHO R

CHO O

R’

NR₂ R’

CHO

* *

*

*

o

Atroposelective arene-forming aldol condensation developed in the group:

vi

An iterative chain elongation sequence consisting of a C10 building block addition, an in situ double-oxidation and the arene-forming aldol condensation enabled an efficient assembly of the oligomers (Scheme 2).^[6] The key transformation – the arene-forming aldol condensation – was achieved under substrate stereocontrol to obtain a series of oligomers with a non-helical secondary structure. Subsequently, the inherent substrate-stereocontrol was efficiently inverted with amine- or ion-pairing catalysis to obtain helically shaped oligomers with up to four consecutive stereogenic axes. This synthesis represents the first example of a stereodivergent method for the synthesis of compounds with multiple elements of axial chirality. The properties of the corresponding oligomers were subsequently studied and provided the first insight into the structure of configurationally stable oligo-1,2-arylenes. These analyses revealed a densely packed helical secondary structure, confirming low flexibility and high configurational stability and hence making the oligo-1,2-naphthylenes a valuable molecular scaffold for various applications. In different collaborations first applications – as mono-dentate phosphine ligands for gold- and palladium-catalysis, as bridge between a photosensitizer and an electron-donor, and as catalysts for asymmetric epoxidations – have been studied and promising preliminary results have been achieved.

Scheme 2: Stereodivergent synthesis of oligo-1,2-naphthylenes.

R CHO

R

nOH OH

R

nO

CHO R

n OHC

Mg LiO

2 n

n n + 1

building block addition

double oxidation stereodivergent

aldol condensation

R

OHC

O R

OHC

R CHO catalyst

stereocontrol

substrate stereocontrol

(S_a,S_a) (R_a,S_a)

vii Studies Towards the Total Synthesis of Naphthacemycin B1

With the arene-forming aldol condensation being inspired by the biosynthesis of polyketides, we were strongly interested in the structure of fasamycin congeners, which are proposed to be formed in an atroposelective arene-forming aldol condensation.^[2] Applying the synthetic methods developed in our laboratories would give an insight into the biosynthesis of these compounds, and moreover enable the very first stereoselective total synthesis of a member of this class of natural products.^[3] To achieve an efficient total synthesis we aimed to develop an (enolendo)-exo-trig aldol cyclization, hence avoiding an exo-cyclic carbonyl group, as it is formed in the established (enolexo)-exo-trig arene-forming aldol condensation.^[7,8]

Scheme 3: Key step of the envisioned total synthesis of naphthacemycin B1.

The studies towards the total synthesis of the fasamycin congener naphthacemycin B1 led to the development of a new reductive Friedel-Crafts cyclization, enabling an efficient assembly of the dihydro-anthracene core-structure. The desired (enolendo) cyclization was achieved with three different model substrates, though yet unselective. The first substrate, containing an acetonyl side-chain, was successfully cyclized with sodium amide as base and a secondary amine catalyst to obtain the corresponding binaphthalene. The same product was obtained with a substrate with an a-bromo acetonyl side chain through an interesting phosphine mediated titanium enolate formation. Finally, also a substrate containing a b-keto ester side chain was efficiently cyclized to the tetra-ortho-substituted biaryl in a Lewis acid mediated reaction.

Based on these results, the ongoing studies in the group focus on the stereoselective synthesis of naphthacemycin B1.

Me OH

OH OH

HO

OH Me Me

OH

O O

Me OR

OR OR

RO

RO Me Me

O

1) (enolendo)-exo-trig cyclization 2) oxidation/deprotection

Me

naphthacemycin B₁

viii

Scheme 4: Reductive Friedel-Crafts cyclization and successful (enolendo)-exo-trig cyclizations.

OMe MeO

OMe Me Me

R O

Me Me MeO

OMe

OMe

R

TiCl₄, Et₃SiH

O OMe Me O

O OMe O

O OMe O

Br

OMe OH NaNH₂ R

NH N catalyst:

TiCl₄, PPh₃ R = H

CO₂Et

R = H

TiCl₄ R = CO₂Et Reductive Friedel-Crafts cyclization:

(Enolendo)-exo-trig cyclization:

ix

IV. Zusammenfassung

Diverse Polyketid-Naturstoffe werden in der Natur durch komplexe Polyketid-Synthasen zusammengesetzt, cyclisiert und modifiziert.^[1] Viele dieser Verbindungen zeigen antibiotische Eigenschaften, was sie zu interessanten Zielen für Studien von Biosynthesen und Bioaktivitäten sowie chemischen Totalsynthesen macht. Unter diesen Polyketid-Naturstoffen befinden sich viele atropisomere Verbindungen, wovon sich einige durch einen einzigartigen Weg zur Entstehung der konfigurativ-stabilen Achse hervorheben. Eine atroposelektive Aren-bildende Aldolkondensation wurde als Schlüsselschritt in der Biosynthese dieser Verbindungen vorgeschlagen.^[2] Eine synthetische Variante dieser Reaktion wurde in unserer Gruppe entwickelt und ermöglicht die Synthese von verschiedenen Verbindungen mit konfigurativ- stabilen Achsen.^[3] Zu diesen Verbindungen gehören unter anderem Oligo-1,2-naphtylene, welche zwei aufeinanderfolgende stereogene Achsen aufweisen und durch eine Katalysator- kontrollierte, enantioselektive und eine Substrat-kontrollierte, diastereoselektive Aldolkondensation synthetisiert werden.^[4] Diese Oligomere können durch ihre interessante Struktur als Gerüst für die Positionierung von Substituenten in einer vorhersehbaren, genau definierten relativen Orientierung dienen.

Scheme 5: Die atroposelektive arenbildende Aldolkondesation und ihr synthetisches Gegenstück aus unserer Gruppe.

Stereodivergente Synthese von Oligo-1,2-naphthylenen

Aufgrund der konfigurativ-stabilen Achsen zwischen den einzelnen Einheiten des Oligomers und dem daraus resultierenden exponentiellen Anstieg der möglichen Diastereomere entstehen synthetische Herausforderungen. In dieser Dissertation wollten wir mit der Entwicklung einer

OH

O

S-Enz O O

OH OH Me

O OH HO

OH

OH

Me OH

OH HO

OH O

Vorgeschlagener Schlüsselschritt in der Biosynthese von Fasamycin A:

O S-Enz

CHO

R R’ O

CHO R

CHO O

R’

NR₂ R’

CHO

* *

*

*

o

In unserer Gruppe entwickelte atroposelektive arenbildende Aldolkondensation:

CY/AR

catalyst

x

stereodivergenten Methode eine Lösung für diese Probleme bereitstellen, um verschiedene Diastereomere von Oligomeren mit diversen Längen zugänglich zu machen.^[5]

Eine sich wiederholende Sequenz aus einer C10 Baustein-Addition, einer doppelten Oxidation und schließlich einer Aren-bildenden Aldolkondensation erlaubte uns einen effizienten Aufbau der Oligomere.^[6] Der Schlüsselschritt – die Aren-bildende Aldolkondensation – resultierte dabei unter Substrat-Stereokontrolle in nicht-helikalen Diastereomeren. Diese inhärente Substrat-Kontrolle konnte durch Amin- und Ionenpaar-Katalyse umgedreht werden, um die helikalen Diastereomere zu erhalten. Die Synthese dieser Oligomere ist das erste Beispiel einer stereodivergenten Synthese von Multiachsen-Systemen und erlaubte uns des Weiteren, diese Verbindungen zu untersuchen und erste Einblicke in ihre strukturellen Eigenschaften zu erhalten. Dabei stellte sich heraus, dass die Oligomere eine kompakte helikale Sekundärstruktur annehmen, was die Hypothese von geringer struktureller Flexibilität und hoher konfigurativer Stabilität bestätigte. Diese Eigenschaft macht Oligo-1,2-naphthylene zu wertvollen Gerüsten für verschiedene Anwendungen. In mehreren Kollaborationen wurden erste Applikationen – zum Beispiel als mono-dentate Phosphine-Liganden für Gold- und Palladiumkatalyse, als Brücke zwischen einem Photosensibilisator und einem Elektronen-Donor oder als Katalysator für asymmetrische Epoxidierungen – getestet und ergaben vielversprechende erste Resultate.

Scheme 6: Stereodivergente Synthese von Oligo-1,2-naphthylenen.

R

CHO

R

nOH OH

R

nO

CHO R

n OHC

Mg LiO

2 n

n n + 1

Baustein- Addition

Doppelte Oxidation Stereodivergente

Aldolkondensation

R

OHC

O R

OHC

R CHO Stereokontrolle

durch Katalysator

Stereokontrolle durch Substrat

xi Studien zur Totalsynthese von Naphthacemycin B1

Weil die von uns entwickelte Aren-bildende Aldolkondensation massgebend von der Natur inspiriert ist, waren wir in hohem Masse an der Struktur und der Synthese von Fasamycin- Derivaten, interessiert, welche vermutlich durch eine atroposelektive Aldolkondensation entstehen und nicht durch die übliche oxidative Dimerisierung.^[2] Wenn die synthetische Version der Reaktion in einer Totalsynthese angewendet werden könnte, würde dies Einblicke in die Biosynthese der Verbindungen geben und zusätzlich die erste, stereoselektive Synthese dieser Naturstoffe ermöglichen.^[3] Um eine solche Totalsynthese effizient zu gestalten, wollten wir eine (enolendo)-exo-trig Cyclisierung entwickeln, in welcher – im Gegensatz zur etablierten (enolexo)-exo-trig Aren-bildenden Aldolkondensation – keine exo-cyclische Carbonylgruppe entsteht, welche daraufhin modifiziert werden müsste.^[7,8]

Scheme 7: Schlüsselschritt in der geplanten Totalsynthese von Naphthacemycin B1.

Die Studien zur Totalsynthese des Fasamycin-Derivats Naphthacemycin B1 führte zur Entwicklung einer neuen, reduktiven Friedel-Crafts Cyclisierung, welche einen effizienten Aufbau des Dihydroanthracenyl-Gerüsts ermöglicht. Die folgende, bislang unselektive, (enolendo)-Cyclisierung konnte mit drei verschiedenen Modellsubstraten realisiert werden. Das erste Substrat, welches eine Acetonyl-Seitenkette enthält, wurde erfolgreich mit Natriumamid als Base und einem sekundären Amin als Katalysator zu einem Binaphthalen umgesetzt.

Dasselbe Produkt wurde aus einem Substrat mit einer a-Bromoacetonyl Seitenkette durch eine interessante, Phosphin-vermittelte Bildung eines Titanenolats geformt. Schließlich konnte auch ein Substrat mit einer b-Ketoester Seitenkette erfolgreich mit einer Lewis-Säure zum tetra- ortho-substituierten Binaphthalen cyclisiert werden. Basierend auf diesen Resultaten fokussieren sich die aktuellen Studien in der Gruppe auf die stereoselektive Synthese von Naphthacemycin B1.

Naphthacemycin B₁

Me OH

OH OH

HO

OH Me Me

OH

O O

Me OR

OR OR

RO

RO Me Me

O

1) (enolendo)-exo-trig Cyclisierung 2) Oxidation/Entschützung Me

xii

Scheme 8: Reduktive Friedel-Crafts Cyclisierung und erfolgreiche (enolendo)-exo-trig Cyclisierungen.

OMe MeO

OMe Me Me

R O

Me Me MeO

OMe

OMe

R

TiCl₄, Et₃SiH

O OMe Me O

O OMe O

O OMe O

Br

OMe OH NaNH₂ R

NH N Katalysator:

TiCl₄, PPh₃ R = H

CO₂Et

R = H

TiCl₄ R = CO₂Et Reduktive Friedel-Crafts Cyklisierung:

(Enolendo)-exo-trig Cyklisierung:

xiii

V. Publications

Parts of this thesis have been published:

• Stereoselective Arene-Forming Aldol Condensation: Synthesis of Configurationally Stable Oligo-1,2-naphthylenes

D. Lotter, M. Neuburger, M. Rickhaus, D. Häussinger, C. Sparr

Angew. Chem. Int. Ed. 2016, 55, 2920–2923; Angew. Chem. 2016, 128, 2973–2976.

DOI: 10.1002/anie.201510259

• Catalytic Arene-forming Aldol Condensation: Stereoselective Synthesis of Rotationally Restricted Aromatic Compounds

V. C. Fäseke, R. M. Witzig, A. Link, D. Lotter, C. Sparr Chimia 2017, 71, 596–599.

DOI: 10.2533/chimia.2017.596

• Stereoselective Arene-Forming Aldol Condensation: Catalyst-Controlled Synthesis of Axially Chiral Compounds

R. M. Witzig, D. Lotter, V. C. Fäseke, C. Sparr Chem. Eur. J. 2017, 23, 12960–12966.

DOI: 10.1002/chem.201702471

• Diaryl Magnesium Lithium Alkoxides Derived from (Z)-4-(2-Bromophenyl) but-3-en-1-ol

D. Lotter, C. Sparr

in Encyclopedia of Reagents for Organic Synthesis, 2018 DOI: 10.1002/047084289X.rn02170

• Catalyst-Controlled Stereodivergent Synthesis of Atropisomeric Multi-Axis Systems D. Lotter, A. Castrogiovanni, M. Neuburger, C. Sparr

ACS Cent. Sci. 2018, 4, 656–660.

DOI: 10.1021/acscentsci.8b00204

xiv

VI. Presentations

• “Stereoselective Arene-Forming Aldol Condensation: Synthesis of Configurationally Stable Oligo-1,2-naphthylenes” D. Lotter, M. Neuburger, M. Rickhaus, D. Häussinger, C. Sparr, 36^th Regio-Symposium, Mittelwihr (FR), 29^th–31^st August 2016: poster presentation.

• “Stereoselective Arene-Forming Aldol Condensation: Synthesis of Configurationally Stable Oligo-1,2-naphthylenes” D. Lotter, M. Neuburger, M. Rickhaus, D. Häussinger, C. Sparr, SCS Fall Meeting, Zurich, 15^th September 2016: oral presentation (awarded with the “runner up award”).

• “Stereoselective Arene-Forming Aldol Condensation: Synthesis of Configurationally Stable Oligo-1,2-naphthylenes” D. Lotter, M. Neuburger, M. Rickhaus, D. Häussinger, C. Sparr, 1^st Swiss Industrial Chemistry Symposium, Basel, 28^th October 2016: poster presentation.

• “Stereoselective Synthesis of Oligo-1,2-naphthylenes” D. Lotter, M. Neuburger, C. Sparr, 20^th European Symposium on Organic Chemistry, Cologne (DE), 2^nd–6^th July 2017: poster presentation.

• “Diastereoselective Synthesis of Oligo-1,2-naphthylenes” D. Lotter, M. Neuburger, C. Sparr, Hochschule trifft Industrie, Feldberg-Falkau (DE), 4^th–6^th October 2017: poster presentation.

• “Stereodivergent Synthesis of Multiaxis Systems” D. Lotter, M. Neuburger, C. Sparr, Curo-p³, Oxford (UK), 5^th–7^th September 2018: poster presentation.

• “Stereodivergent Synthesis of Multiaxis Systems” D. Lotter, M. Neuburger, C. Sparr, 21^st ORCHEM, Berlin (DE), 10^th–12^th September 2018: oral and poster presentation.

Introduction 1

1. I NTRODUCTION

Natures complexity, evolved over millions of years, is absolutely fascinating, when one is aware that it arises from only very few small building blocks. The four base-pairs of DNA encode for an uncountable number of enzymes – natures molecular machinery and tool box.

The tool box itself is again assembled from a limited number of amino acids and few other building blocks. This, however, does not even closely reflect the amount, diversity and efficiency of reactions nature is capable to catalyze. When seeing every heartbeat or every movement of a finger as an incredibly complex series of chemical reactions, one immediately realizes what diversity of chemical reactions nature is actually capable to perform.^[9]

Combinations of enzymes with different functions are thereby responsible for many multistep transformations, eventually leading to living organisms. The photosynthesis in plants to ultimately form sugars from carbon dioxide, plays a fundamental role in its way of storing energy.^[10] The degradation of these sugars by various enzymes in other organisms provides building blocks for a variety of relevant structures and functions for organisms, such as acetyl- coenzym A (acetyl-CoA).^[11] Energy for other transformations is gained in the citric acid cycle which is utilizing acetyl-CoA. Many of these cascaded reactions in nature, such as the photosynthesis or the catabolism of sugars with the citric acid cycle, can be found over numerous different organisms and set a basis for life.^[12]

Life as we know it would nevertheless not exist, if there would not be tremendous differences between the organisms.^[12] Consequently, they evolved numerous, unique chemical transformation sequences, giving them an evolutionary advantage. The most illustrative example of such unique reaction sequences is the biosynthesis of antibiotics, which can be found in different plants, fungi or bacteria, providing secondary metabolites for protection against other bacteria. Often these antibiotics are polyketides, which are natural products originally derived from acetyl-CoA units.^[13–15]

Biosynthesis of Polyketides 2

1.1. Biosynthesis of Polyketides

Polyketide natural products are synthesized by the complex machinery of different polyketide synthases (PKS) found in bacteria, fungi and plants. These enzymes are strongly related to fatty acid synthases and can similarly be categorized in three different types:^[1,16,17] Type I PKS consist of different, covalently linked modules within one or several multifunctional enzymes.

Type II and type III on the other hand, contain different individual enzymes in a cluster; these enzymes are monofunctional in type II PKS, and acyl carrier protein (ACP) independent in type III.^[18] While type II and III are iterative synthases, meaning an active side can perform more than one transformation, type I can also be non-iterative and consequently each module of the enzyme is catalyzing only one chemical transformation. A type I PKS contains modules with different domains and contains a starting unit with an acyltransferase (AT) and optionally an ACP. The elongation modules contain at least a keto synthase (KS), an AT and an ACP, but can include a variety of other domains such as ketoreductases (KR), dehydratases (DH) or enoylreductases (ER). The natural product is finally cleaved on a terminating unit which contains a thioesterase (TE).

In contrast to type II and III PKS, type I does not only accept malonyl-CoA extender units, but various different building blocks. The immense complexity and variety going along with acceptance of different extender units and the modularity of type I polyketide synthases is nicely exemplified with the well-studied erythromycin polyketide synthase and the biosynthesis of 6-deoxyerythronolide B, the precursor for the antibiotic erythromycin (Scheme 9).^[15,19–21]

The synthase is consisting of three deoxyerythronolide B synthase (DEBS 1–3), each containing two modules of which both are responsible for the elongation by one unit and its modification.

The biosynthesis starts on a starter unit, on which a propionate is bound to the ACP domain.

Each module contains a KS, responsible for transferring one propionate unit from carboxy- propionyl-CoA with simultaneous decarboxylation. After the addition the optional domains (KR, DH, ER) modify the newly introduced part of the chain, until the TE of the terminating unit is cleaving the polyketide to obtain the final product 6-deoxyerythronolide B after macro- lactonization.

Introduction 3

Scheme 9: Organization of erythromycin polyketide synthase and biosynthesis of 6- deoxyerythronolide B. DEBS = deoxyerythronolide B synthase.

In contrast to the type I PKS, type II synthases do often consist of much less functional domains.

A minimal PKS (minPKS) is only consisting of two KS, which build a pocket defining the final chain-length of the polyketide, and an ACP (Scheme 10). The individual proteins that are closely assembled to a polyketide synthase cluster. The synthesis is hence initiated with an activated acyl unit on the KS to which a malonyl unit is added to elongate the chain by one unit under release of carbondioxide. The elongated chain is subsequently transferred back to the KS where a next malonyl unit can be added. The iterative addition of malonyl-units is finally terminated at a chain-length defined by the shape of the synthase. Type III PKS in comparison, do only differ in the absence of an ACP and instead the malonyl-units are transferred directly from coenzyme A.^[22]

AT ACP KS AT KR ACP KS AT KR ACP KS AT ACP KS AT DH ER KR ACP

KS AT KR ACP KS AT EK ACP TE S

Me

O O S

HO Me Me

S O HO Me HO Me Me

S O O Me HO Me HO Me Me

S O

Me O Me

HO Me HO Me

Me

S O

Me Me O Me

HO Me

HO Me

Me HO

S O

Me Me Me

O Me

HO Me

HO

Me

HO Me

HO

O O

Me Me OH Me

OH Me O Me

Me OH Me

S-CoA O

Me O O

Prop-CoA Prop-CoA

Prop-CoA

Prop-CoA

Prop-CoA

Prop-CoA Prop-CoA

load module 1 module 2 module 3 module 4

module 5 module 6 end

DEBS 1 DEBS 2

DEBS 3

6-deoxyerythronolide B

Biosynthesis of Polyketides 4

Scheme 10: Chain elongation of a polyketide by a minimal polyketide synthase type II.

Due to the high reactivity of the formed polyketide chain, diverse aldol addition and condensation reactions as well as lactone or acetal formations could potentially proceed with growing chain length. The polyketide transferase nevertheless can protect the chain from spontaneous, unselective reaction and is able to induce prearrangement of the chain, allowing other enzymes to perform subsequent reactions (Scheme 11).^[1] Mithramycin for example is formed from a decaketide, which is prealigned in a U-shaped arrangement.^[2] A first set of cyclases (CY) and aromatases (AR) leads to the formation of an anthracenyl core. Oxidases (OX) and other cyclases are able to further modify the product which then undergoes several subsequent transformations with methyltransferase (MT), glycotransferases (GT), oxidases and ketoreductases to form the natural product mithramycin. Other minPKS lead to the formation of a bent arrangement of a dodecaketide derived from a hexanoate starting unit and eleven malonyl extender units by a minPKS. The subsequent cyclizations, aromatizations, reductions and methylations transform these simple starting materials into the natural product benastatin A.

KS_α KS_β ACP S

O Me

S O O

O

KS_α KS_β ACP SH O S

O Me

KS_α KS_β ACP S SH

O

Me O

KS_α KS_β ACP S

O

Me O

S O O

O KS_α KS_β ACP

SH O S O

malonyl-CoA –CO₂

O Me KS_α KS_β ACP

SH O S O

O Me

n

–CO₂ n‧malonyl-CoA

– n‧CO₂

Introduction 5

Scheme 11: Biosynthesis of mithramycin and benastatin A. minPKS = minimal PKS; CY = cyclase;

AR = aromatase; OX = oxygenase; MT = methyltransferase; GT = glycotransferase; KR = ketoreductase.

1.2. Atropisomeric Compounds

As exemplified with the biosynthesis of mithramycin and benastatin A, nature is capable of performing a variety of modifications after the polyketide is cyclized, such as methylations, halogenations, oxidations or reductions. These modifications lead to a tremendous number of natural products, which are derived from the polyketide machinery.^[1,15] Next to these versatile post-modifications, dimerizations can also occur, typically through oxidative dimerizations, which often lead to rotationally restricted, atropisomeric natural products.^[23]

O

O O O

S-Enz O

Me O O O O O

OH OH OH HO

OH O

O Me O

O O O

Me O OH

OH HO

OH OH O

O

OH OH Me

OMe

O

Me OH OH

O sugar sugar

Mithramycin

O O O O

O O O O O O Enz-S

O Me 4 O

O OH O OH

OH O

Me 4 O O

Enz-S O

OH O OH OH Me 4

O HO

OH

OH OH O OH

OH Me

O HO

OH Me Me O

O O

SCoA

10 minPKS CY, AR

CY, OX MT, GT

OX, KR

O O O

SCoA 11

minPKS

benastatin A

CY, AR

CY, AR, KR MT

O Me SCoA

4

4

Atropisomeric Compounds 6

1.2.1. Atropisomers, Definition

An atropisomer was initially defined to contain axis with a rotational barrier (DG^‡) of at least 94 kJmol^–1 at room temperature. This laboratory derived definition of the barrier to rotation nevertheless only reflects a practical aspect of a configurationally stable axis corresponding to a half-life of racemization of 1000 s at ambient temperature that enables separation of the enantiomers.^[24] In nature, optically active atropisomers require much larger rotational barriers of >120 kJmol^–1 to maintain optical activity for several hours or days.^[3] The rotational restriction of biaryls is largely dependent on the number of ortho-substituents. Biaryls with only two-ortho-substituents are rotating very fast at room temperature, with a typical half-life of racemization of seconds to minutes (Figure 1). Characteristically, three or four ortho- substituents are necessary to obtain the desired high barrier to rotation.

Figure 1: Barrier to rotation of di-, tri- and tetra-ortho-substituted biaryls.^[3]

1.2.2. Natural Atropisomeric Compounds

Consequently, optically active natural products possess three or four ortho-substituents (Figure 2). Interestingly, in most cases two or more of these ortho-substituents are hydroxy- or methoxy groups.^[23,25,26] An explanation can be found in the biosynthesis of these compounds.

Being derived from polyketide- (e.g. viridotoxin or kotanin) or highly oxidized terpene monomers (e.g. gossypol), most atropisomeric natural products are formed through an oxidative dimerization.^[27]

Figure 2: Configurationally stable, atropisomeric natural products.

Me

MeO CHO OH

OH

ΔG^‡_{332 K} = 101 kJmol^–1 ΔG^‡_{383 K} = 125 kJmol^–1 ΔG^‡_{493 K} = 158 kJmol^–1

O

O MeO

MeO

OH OH OH OH O

OMe O

OMe O O

(–)-viriditoxin (+)-gossypol

OH CHO OH OH i-Pr Me

HO Me

OHC HO

OH i-Pr

OMe OMe O

O O O

OMe

OMe Me

Me (+)-kotanin

Introduction 7 The oxidative dimerizations that can take place as a subsequent step after the polyketide biosynthesis, are typically catalyzed by cytochrome P450 and are not necessarily regio- and enantioselective (e.g. diflaviolins),^[28,29] but normally diastereoselective.^[23,25] The characteristic substitution in the ortho-positions thus is a result of the electronic properties of the polyketide monomers, that are coupled in an oxidative dimerization in the position originating from the a- position of a carbonyl group in the polyketide chain, hence leading to a symmetric connectivity of the two aryl-moieties (Figure 3, left). Nevertheless, in nature a small number of atropisomeric polyketides can be found, which show a typical 1,3-connectivity that is even continued through the stereogenic axis, indicating a different biosynthetic pathway (Figure 3, right). This important and unique asymmetric biaryl motive can for example be found in naphthacemycins, fasamycins, formicamycins, cassibiphenols or cassiarins.^[30–33]

Figure 3: Connectivity of oxidative dimerization products compared with the connectivity of naphthacemycin B1.

1.2.3. Arene-Forming Aldol Condensation in Nature

The 1,3-connectivity through the entire molecular backbone arises when the axis is formed in an aldol condensation reaction rather than a dimerization. This cyclization is nicely exemplified with the proposed biosynthesis of fasamycin A. Brady and his group were able to identify different cyclases and aromatases which are involved in the biosynthesis of fasamycin A and proposed the initial formation of a linear trideca-ketide by a minPKS.^[2] The linear chain is prearranged by a cyclase to putatively form an anthracenone intermediate after aromatization with an aromatase (Scheme 12). In the proceeding step, the ketide chains possess different options to cyclize to the observed product. It is plausible that the bottom ring is formed first, since the terminal ketone group most likely is the best accessible electrophile. The second arene-forming aldol condensation would then stereoselectively form the top ring. Due to the low reactivity of the di-arylketone intermediate, it could nevertheless also be possible, that

O

O OH OH

O

O

HO OH

OH HO

HO

OH O Me Me

OH

OH

Me OH

OH

diflaviolin naphthacemycin B₁

symmetric unsymmetric

1

3 1 3’

3 3’

3 1 3’

3 3’

1 3

2’ 2’

2

Atropisomeric Compounds 8

initially the top ring is formed, then followed by the second one or that a double aldol-addition is taking place, followed by a stereospecific double-dehydration. All possible pathways lead to the same intermediate, which is then decarboxylated and finally substituted by a methyltransferase and a halogenase.^[31]

Scheme 12: Proposed biosynthesis of fasamycin A.

O

O O O

O O S-Enz

O O O

O

O

O

Me O

O OH HO

OH

O

S-Enz O O

O

O

Me O

OH O

O

Me O

OH S-Enz OH O

O

S-Enz O O

OH Me OH

O OH HO

OH

OH S-Enz O

Me OH

OH O

OH HO

OH

OH

Me OH

OH Me Me

Cl

OH

OH S-Enz OHO

O

O HO

*

*

or or

aromatase cyclases

–CO₂ methyltransferase

halogenase initial formation of

the bottom ring

initial formation of the top ring

double aldol-addition

fasamycin A

Introduction 9 (–)-Anthrabenzoxocinone (ABX) and its bis-chlorinated derivative (–)-BABX provide evidence that the bottom ring is cyclized first (Scheme 13).^[34] These natural products are formed when instead of a second cyclization step, a reduction of the di-aryl ketone occurs. After the subsequent decarboxylation, the phenol is forming a hemiacetal with the carbonyl of the acetonyl sidechain. A substitution reaction with the secondary alcohol is then yielding the acetal moiety of the natural product (–)-ABX. The occurrence of this natural product is indicating that the neither the top ring is cyclized first, nor that a double aldol-addition is involved, but that the two rings are formed in separate steps, starting with the bottom ring.

Scheme 13: Proposed biosynthesis of (–)-ABX.

1.2.4. Fasamycins and its Congeners

The most important class of 1,3-connected biaryls are fasamycins and its closely related congeners. This class consists of 33 compounds (related naphthacemycins C1 and C2 seem to be an isolation-artefact) which differ in the oxidation state of ring A3 and the number of chlorinations and methylations.[2,30,31,35]

Figure 4: Structure of fasamycins and its congeners.

Unsurprisingly, not only the structures of fasamycins, naphthacemycins and formicamycins are strongly related, but also the organisms producing them. The first isolation of compounds out of this class was achieved during metagenomic-experiments by Brady and his group.^[2] Gene clusters of unknown Streptomyces isolated from soil in Arizona, were implemented into

OH

O

S-Enz O O

OH OH

Me OH

OH

OH OH Me

Me O

O OH HO

OH

O O Me

H

Me OH

Me Me

(–)-ABX HO

OH O

–CO₂

OH

O OH Me Me HO

OH

X³

OR¹

OR² Me X¹

X²

X⁴ X⁵ R = H or Me

X = H or Cl

fasamycins naphthacemycins B OH

O O

Me Me HO

OH

X¹

OMe

OR Me

X²

naphthacemycins A O

OH

O O

Me Me HO

OH

X³

OR

OMe Me

X¹

X²

X⁴

formicamycins OH

H

A1 A2 A3 A4

B

Atropisomeric Compounds 10

Streptomyces albus, in which fasamycins A and B were subsequently identified as two major metabolites. The first isolation from a natural bacterium was achieved by Hutchings and his group.^[31] From a culture broth of Streptomyces formicae KY5, an actinomycete bacteria cultivated by plant-ants (Tetraponera penzigi),^[35] a large number of metabolites (fasamicins C–E, formicamycins A–M) with a configurationally stable axis were isolated.

Naphthacemycins were isolated from Streptomyces by the groups of Ōmura and Shiomi, completing this set of atropisomeric polyketides.^[30] The strain Streptomyces sp. KB-3346-5, isolated from soil in Okinawa, hence produced 15 identified naphthacemycins.

The isolation of these compounds became particularly interesting, when first bioactivity studies were conducted. Most fasamycins and its congeners show high activity against gram-positive bacteria.^[2,36–38] Especially the activity against methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Enterococcus faecalis (VRE), both wide-spread and potentially lethal infections, thus making them interesting as potential antibiotics.^[39–41] For MRSA, comparable minimum inhibitory concentrations (MIC) (e.g. naphthacemycin A8

against strain KB362, MIC = 1 µg/mL) as with vancomycin have been determined. Similarly low MICs were determined against VRE strains and it was shown, that the anti-MRSA activity of the antibiotic Imipenem was enhanced by the addition of different fasamycin congeners by a factor of up to 512.^[30,42] Moreover, and particularly interesting was the absence of resistant MRSA mutations even after 20 generations when MRSA was treated with ½MIC formicamycin C.^[31]

With the activities going along with fasamycin congeners, it is not surprising that synthetic chemists were interested in the synthesis of these natural products. Even though fasamycins A and B have been isolated already in 2011, it was only in 2017 when the first total synthesis of the congener (±)-naphthacemycin A9 was achieved in the group of Sunazuka.^[37] This first synthesis was followed in 2018 by the preparation of (±)-fasamycins A and (±)-naphthacemycin A9 in the group of Shia^[43] and finally in the same year by the synthesis of (±)-naphthacemycin A9 in the group of Kraus.^[44]

Introduction 11 1.2.5. Sunazuka’s Synthesis of (±)-Naphthacemycin A9

The first successful total synthesis of a fasamycin congener was achieved in the group of Sunazuka.^[37] In their synthesis of (±)-naphthacemycin A9, the key biaryl bond was installed in the first two steps through a borylation of 2-bromo-1,5-dimethoxy-3-methylbenzene and a subsequent Suzuki-Miyaura coupling with 1-bromo-2-iodo-4-methoxybenzene (Scheme 14).

The corresponding biaryl bromide was lithiated and added to chromium hexacarbonyl to obtain a chromium carbene which was directly protected with the Meerwein salt. The subsequent Dötz-reaction can be considered as the key reaction of this total synthesis. Even though the Dötz-reaction with the alkyne yielded the desired naphthalene (Scheme 14, pathway b), however the undesired reaction with the terminal alkene (Scheme 14, pathway a) to form a cyclobutenone was found to be predominant.^[45]

Scheme 14: Dötz-reaction for the construction of the upper part of naphthacemycin A9; MW = microwave.

Br MeO

OMe Me

MeO

OMe Me Br

OMe

OMe

OMe (OC)₅Cr

OMe

OMe (pin)B

Me O

OMe

OMe (pin)B

OMe OMe Me

OMe (pin)B

C O Me

a

b

40%

(via pathway a) 14%

(via pathway b) via:

+ 1) n-BuLi, B(Oi-Pr)₃

2) 4-bromo-3-iodo- anisole, Pd(PPh₃)₄ Cs₂CO₃

76% over two steps

n-BuLi, Cr(CO)₆ then Me₃OBF₄

77%

1) A, MW 2) MeI, Cs₂CO₃

Me B

O O Me Me

Me Me A

Ar Ar

Ar Ar Ar

Atropisomeric Compounds 12

The undesired product was managed to be converted into the desired naphthalene by a thermal rearrangement and subsequent protection (Scheme 15).^[46] The product was then subjected to a Suzuki-Miyaura coupling with benzyl bromide A. After exchanging the protecting groups, an acid mediated Friedel-Crafts alkylation yielded the spiro-cyclic intermediate.^[45]

Scheme 15: Thermal rearrangement and Friedel-Crafts cyclization in Sunazuka’s total synthesis.

The synthesis was concluded with the oxidation of the dimethoxynaphthyl moiety to the corresponding naphthalene-dione, followed by a Lewis acid mediated dienone-phenol rearrangement yielding the fused tetracyclic core-structure (Scheme 16). After protection of the phenol, the dibenzylic position was oxidized in two steps to the protected derivative of (±)-naphthacemycin A9. Finally, deprotection of the acetyl protecting group and a remarkable, selective mono-deprotection of one methoxy group yielded the natural product.^[45]

Scheme 16: Oxidation steps of Sunazuka’s synthesis of (±)-naphthacemycin A9.

OMe

OMe (pin)B

Me O

OMe

OMe (pin)B

OMe Me

Ar Ar

OMe

OMe (pin)B

Ar Me

MW irradiation

OH

MeI, Cs₂CO₃ 42% over two steps

OMe

OMe OMe Me

Ar TsO

OTs OMe

OMe OMe

Ar Me Me

O

OMe Br

TsO

OTs A

A, Pd(PPh₃)₄ aq. Cs₂CO₃, 75%

1) Na_(s), MeOH 2) MeI, Cs₂CO₃ 63% over two steps 3) H₂SO₄, AcOH 83%

OMe

OMe OMe

Ar Me Me

O

OMe

OMe

O O

Ar AcO

OMe Me Me

OH

OMe

O O HO

OH Me Me

O Me

OMe OMe 1) CAN, 76%

2) TiCl₄, 44%

3) Ac₂O, pyridine 4) CAN

84% over two steps

1) PCC, celite MS, NaOAc, 59%

2) MeOH, aq. NaHCO₃, 63%

3) CeCl₃‧7H₂O, NaI, 64%

(±)-naphthacemycin A₉

Introduction 13 This first total synthesis proved the accessibility of the natural products and provided first insights into possible late-stage transformations. The overall yield of 0.47% over 18 linear steps is, compared to the complexity of the natural product, nevertheless relatively low.^[45] This can partially be associated with the amount of protection and deprotection steps and with low yielding key steps such as the Dötz-reaction, the thermal isomerization and the Lewis acid mediated dienone-phenol rearrangement.

1.2.6. Shia’s Synthesis of (±)-Naphthacemycin A9

The group of Shia addressed the issue of the low yield, by a strategy based on a late-stage Hauser-Kraus annulation with building blocks of similar complexity.^[38,47] The acceptor for the annulation was prepared in a radical, iron(III) catalyzed, reductive olefin coupling previously described by Baran (Scheme 17).^[48] An iron(III) hydride is reducing the olefin to form a radical, which is then reacting with an electron-deficient olefin. Upon single-electron transfer from the iron(II), the intermediate is protonated to release the desired product. The corresponding ester was subsequently hydrolyzed to undergo a trifluoroacetic acid anhydride mediated Friedel- Crafts cyclization. The corresponding ketone was found to react further to the enol- trifluoroacetate, which is then cleaved again under basic conditions and finally oxidized with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) to the desired Hauser-Kraus-acceptor.

Scheme 17: Preparation of the Hauser-Kraus-acceptor by Shia.

The preparation of the Hauser-Kraus-donor started with the synthesis of two building blocks for a Suzuki-Miyaura coupling (Scheme 18). Bromination of 1,3-dimethoxytoluene followed by lithiation and addition to an oxoborolane gave the pinacol-boronic ester building block. The coupling partner was prepared from (3,5-dimethoxyphenyl)methanol with a Vilsmeier-Haack formylation, a selective mono-deprotection, a Lindgren-type oxidation with direct lactonization and followed by a triflation of the phenol. After the Suzuki-Miyaura coupling of the two building blocks, the corresponding biaryl-lactone product was opened to the amide-alcohol and oxidized to the amide-aldehyde. Addition of TMS-CN led to the formation of (trimethylsilyl)cyanohydrin, which was finally deprotected with acetic acid to obtain the cyanohydrin, which rapidly underwent lactonization to the Hauser-Kraus-donor.^[38]

BnO

OBn Me

BnO

OBn O

OMe Me Me

BnO

OBn Me Me

O methyl acrylate

PhSiH₃, Fe(acac)₃ 91%

1) aq. NaOH, MeOH 2) TFAA

3) aq. NaOH, MeOH 95% over three steps 4) DDQ, 82%