THE STRUCTURE ELUCIDATION AND SYNTHESIS OF SELECTED NATURAL PRODUCTS

by

WILHELMINA MARAIS

Thesis presented in fulfillment

of the requirements for the degree

PHILOSOPHIAE DOCTOR

in

CHEMISTRY

in the

FACULTY OF SCIENCE

of the

RAND AFRIKAANS UNIVERSITY

Supervisor: Professor C.W. Holzapfel

December 2000 LIST OF ABBREVIATIONS

Ac acetyl acac acetylacetonate AHR asymmetric Heck reaction Bn benzyl CO carbon monoxide dba dibenzylidene acetone dipp 1,3-bis(diisopropylphosphine)propane DME 1,2-dimethoxyethane DoM directed ortho metallation EI-MS electron impact mass spectromety ES-MS electronspray mass spectrometry EtOAc ethyl acetate FAB-MS fast atom bombardment mass spectrometry GC-MS gas chromatography-mass spectrometry HPLC high pressure liquid chromatography m-CPBA meta-chloroperbenzoic acid Me methyl NMR nuclear magnetic resonance nOe nuclear Overhauser effect OTf triflate PHMS poly(methylhydrosiloxane) Piv pivaloyl SQHC sesquiterpene hydrocarbons TBAB tetrabutylammonium bromide TBAI tetrabutylammonium iodide THE tetrahydrofuran TLC thin layer chromatography TONs total catalyst turnover numbers triflates trifluoromethanesulfonates VNS vicarious nucleophilic substitution SYNOPSIS

The objective of the research described in the first part of this thesis was to develop a general method utilising palladium catalysed reactions for the synthesis of the anti-cancer compound, lavendamycin and analogues thereof. Therefore, the development of a general route to synthetic equivalents of the lavendamycin AB quinoline system, 2 -hydroxyquinolines, with potential for coupling to the CDE or CD moiety, was addressed. The first protocol for the synthesis of 2-hydroxyquinolines involved the use of appropriately substituted o- nitrophenyltriflates (readily prepared from phenols) in a Heck reaction under neutral conditions followed by a one-pot reduction and cyclisation step. The synthetic potential of such an approach was demonstrated by the preparation of a suitably substituted lavendamycin AB synthon from commercially available guiacol. A second general strategy towards the synthesis of the AB synthon utilising a preformed ring system such as commercially available 8- hydroxyquinoline has been successfully developed. This approach requiring the introduction of a suitable leaving group in the 2-position involved the following sequence of reactions: protection of the 8-hydroxyl group, N-oxidation, and a rearrangement step. This methodology yielded five different key intermediates all possessing suitable functionality in the 2 position which would allow further cross-coupling to an appropriate CDE ring equivalent.

The next part of this research revolved around approaches to the preparation of a suitable CDE ring system. However, it was realised that although constituting a convergent approach and employing modern catalytic processes the cross-coupling strategy would be very long. Therefore, an alternative approach to lavendamycin via a one-pot palladium catalysed carbonylation of a suitable 2-chloroquinoline derivative with tryptophan as nucleophile to yield an amide containing all the skeletal atoms, was developed. This was followed by a Bischler- Napieralski cyclisation to give the targeted pentacyclic system. Completion of the synthesis will entail a sequence of nitration, reduction and oxidation to complete the left hand side while hydrolysis, lithiation and introduction of a suitable electrophile will complete the right hand side. Although the synthetic approach for lavendamycin described in this thesis is not shorter than any of the published methods, it is unique in that the same methodology will allow access to a range of lavendamycin analogues required for the study of physiological structure-activity relationships. The next part of the thesis involved the development of methods for cyclic enamide synthesis. One of the methods for the preparation of cyclic enamides involved dehydration of the corresponding N-acyl-carbinolamines which was easily prepared from readily available pyrrolidone. The successful application of these compounds in the Fischer indole synthesis on route to the neurohormone, melatonin and other derivatives e.g. a tryptophan analogue was demonstrated.

In the final part of the thesis, the isolation and characterisation of some natural products are described. This research forms part of an ongoing collaboration with chemotaxonomists of the Department of Botany. The chemotaxonomic survey resulted not only in the isolation of several new compounds, one of them the unique compound plicatoloside 4.1 from Aloe plicatilis, but it also successfully demonstrated the application of HPLC coupled to ES-MS in identifying such compounds e.g. from A. africana, A. speciosa and A. broomii. The types of compounds isolated in this study can be regarded as typical of the genus Aloe and may prove to have some chemotaxonomical significance. In addition, the major constituent of Siphonochilus aethiopicus (Zingiberaceae), commonly known as wild ginger, was characterised. It is the only indigenous member of the family in South Africa and has a unique and distinctive morphology reflected in the obvious chemical and thus, biogenetic differences in the structure of the terpenoid constituents of S. aethiopicus and Zingiber officinale Roscoe. SAMEVATTING

Die doel van die ondersoek, soos bespreek in die eerste gedeelte van hierdie proefskrif, behels die ontwikkeling van 'n algemene sintese vir die anti-kanker verbinding, lavendamisien, deur gebruik te maak van 'n reeks palladium gekataliseerde reaksies. Aangesien 2-hidroksikinoliene as 'n geskikte sintetiese ekwivalent vir die lavendamisien AB kinolien ringstruktuur beskou is en ook geredelik aan CDE of CD gedeelte gekoppel kan word, is die ontwikkeling van 'n algemene roete na sulke verbindings, geloods. Die eerste protokol vir die bereiding van 2- hidroksikinoliene benut o-nitrofenieltriflate (maklik bereibaar vanaf fenole) as uitgangstof in 'n Heck reaksie (neutrale kondisies) gevolg deur 'n 'eenpot' reduksie en silklisering. Die suksesvolle bereiding van 'n geskikte lavendamisien AB ringstruktuur vanaf die kommersieel beskikbare, guiacol bevestig die sintetiese potensiaal van so 'n benadering. Die daaropvolgende strategie vir die sintese van 'n AB ringstruktuur maak gebruik van die modifikasie van 'n bestaande kinolien, die kommersieel beskikbare 8-hidroksikinolien. Die vereiste invoeging van 'n geskikte verlatende groep in C-2 van 8-hidroksikinolien is verkry deur die volgende transformasies, nl.: beskerming van die 8-hidroksigroep, N-oksidasie en herrangskikking. Hierdie proses het vyf verskillende sleutelverbindings met die verlangde funksionaliteit in die 2- posisie en gereed vir verdere kruiskoppeling, opgelewer.

In die volgende gedeelte van die ondersoek is verskeie benaderings vir die bereiding van 'n geskikte CDE ringstruktuur oorweeg. Hieruit was dit egter duidelik dat die konvergerende strategie wat op 'n moderne katalitiese proses soos kruiskoppeling gebaseer is, te veel stappe sou behels. Dus is die ontwikkeling van 'n alternatiewe benadering vir die bereiding van lavendamisien deur karbonilering in te span, ontwikkel. 'Eenpot' palladiumgekataliseerde karbonileringsreaksie van 'n geskikte 2-chlorokinolien met triptofaan as nukleofiel het 'n amied gelewer wat deur 'n Bischler-Napieralski siklisering gevolg is, om sodoende die verlangde pentasikliese lavendamisien voorloper te verskaf. Die voltooiing van die sintese behels opeenvolgende reaksies bestaande uit nitrering, reduksie en oksidasie om die linkerkant te voltooi terwyl die regterkant afgehandel sal word deur hidrolise, litiering en invoeging van 'n geskikte elektrofiel. Alhoewel hierdie benadering vir lavendamisien nie veel korter is as die bestaande metodes nie, is dit uniek in die opsig dat dit toegang verleen tot 'n reeks lavendamisien analoe wat noodsaaklik is vir die bestudering van struktuur aktiwiteits verwantskappe. In die volgende gedeelte van die tesis is die ontwikkeling van verskeie metodes vir die bereiding van sikliese eenamiede beskryf. Een van die metodes vir die bereiding van sikliese eenamiede maak gebruik van die dehidrasie van die ooreenstemmende N-asiel-karbinolamiene wat op hul beurt geredelik berei kan word vanaf pirrolidiene. Die suksesvolle toepassing van die verbindings in die Fischer indool sintese vir die bereiding van die neurohormoon, melatonien en ander derivate, o.a. 'n triptofaan analoog, is ook bespreek.

In die finale gedeelte van die tesis word die isolasie en struktuurbepaling van natuurprodukte in samewerking met chemotaksonome van die departement Plantkunde, beskryf. Die resultate van hierdie chemotaksonomiese ondersoek sluit die isolasie van verskeie nuwe verbindings, by. die unieke verbinding plikatalosied uit Aloe plicatilis in, maar dit vestig ook die aandag op die besondere rol wat die gebruik van HPLC gekoppel met ES-MS in die identifisering van sulke verbindings speel, by. die uit A. Africana, A. speciosa en A. broomii. Hierdie verbindings wat tydens die ondersoek gelsoleer is, kan as verteenwoordigend van die genus Aloe beskou word en mag dalk van bepaalde chemotaksonomiese belang wees. In aansluting by bogenoemde is die hoofverbinding van Siphonochilus aethiopicus (Zingiberaceae), algemeen bekend as wilde gemmer, gekarakteriseer. Hierdie spesie is die enigste inheemse lid van die bepaalde familie in Suid-Afrika en word gekenmerk deur 'n unieke en karakteristieke morfologie wat waarskynlik weerspieel word deur die klaarblyklike chemiese en dus, biogenetiese verskille in die strukture van die terpendiedagtige verbindings wat in S. aethiopicus en Zingiber officinale Roscoe voorkom. CONTENTS

List of abbreviations Synopsis Samevatting

CHAPTER 1: SELECTED PALLADIUM CATALYSED REACTIONS IN ORGANIC SYNTHESIS: A BRIEF OVERVIEW

1.1 Introduction 1 1.2 The Heck reaction 1 1.2.1 Background 1 1.2.2 Mechanism of the Heck reaction 2 1.2.3 Improvements and applications of the Heck reaction 6 1.3 Palladium assisted cross-coupling reactions 10 1.3.1 Mechanistic considerations 10 1.3.2 Application of variants of cross-coupling reactions 12 1.4 Palladium(0) catalysed reactions of allylic compounds 15 1.5 Carbonylation 16 1.5.1 History 16 1.5.2 Mechanistic considerations of carbonylation 17 1.5.3 Carbonylation of alkenes, alkynes and allylic substrates 22 1.5.4 Applications: from the synthesis of aldehydes to cascade reactions 23 1.6 Carbonylation of nitro groups 28 1.7 Summary and conclusions 29 1.8 References 30

CHAPTER 2: LAVENDAMYCIN AS A SYNTHETIC TARGET 39

2.1 The history of lavendamycin 39 2.2 Approaches to the synthesis of lavendamycin 40 2.3 Cross-coupling reactions in the synthesis of multicycles 48 2.4 Preparation of the lavendamycin moiety 50 2.4.1 The evaluation of ring synthetic strategies 50 2.4.2 The analysis and evaluation of new strategies 55 2.4.3 Methodology for the conversion of nitrophenols to a lavendamycin AB ring synthon 62 2.4.4 A synthetic approach starting from a preformed ring system 74 2.5 Approaches to the CDE moiety of lavendamycin 81 2.6 An alternative approach to lavendamycin via carbonylation 90 2.7 References 100

CHAPTER 3: MELATONIN AS A SYNTHETIC TARGET 105

3.1 The history of melatonin 105 3.2 The synthesis of melatonin and related derivatives 106 3.2.1 Side-chain attachment strategies 106 3.2.2 Application of the Fischer indole reaction 108 3.3 New approaches to melatonin and derivatives 112 3.3.1 The Heck reaction in a new approach to melatonin 113 3.3.2 A modified Fischer-indole reaction 114 3.3 Summary and conclusions 127 3.4 References 127

CHAPTER 4: THE ISOLATION AND CHARACTERISATION OF SECONDARY METABOLITES FROM SELECTED SOUTH AFRICAN MEDICINAL PLANTS 131

4.1 Introduction 131 4.2 The chemistry of a few Aloe species 131 4.2.1 Background 131 4.2.2 Plicatoloside, an O,O-diglycosylated napthalene derivative from Aloe plicatilis 132 4.2.3 Chromone and aloin derivatives from Aloe africana, A. speciosa and A. broomii 139 4.2.3.1 The identification of A. Africana constituents 139

4.2.3.2 The identification of A. speciosa constituent 141 4.2.3.3 Identification of A. broomii constituents 143 4.3 The study of wild ginger 147 4.4 Summary and conclusions 151 4.5 References 152

CHAPTER 5: EXPERIMENTAL DATA

5.1 General 154 5.2 The naming and numbering of compounds 155 5.3 Preparation of lavendamycin precursors 156 5.3.1 Preparation of 2-hydroxyquinolines 156 5.3.2 Preparation of a lavendamycin AB ring synthon from guiacol 161 5.3.3 Preparation of lavendamycin left hand side from a preformed ring system 164 5.3.4 Nitration of a quinoline derivative 175 5.3.5 Methoxycarbonylation of quinoline derivatives 175 5.3.6 Aminocarbonylation of quinoline derivatives 177

5.3.7 Bischler - Napieralski cyclisation 182 5.3.8 Preparation of lavendamycin pentacyclic system 183 5.4 Preparation of melatonin and derivatives 185 5.4.1 Synthesis of cyclic enamides via oxidation of pyrrolidine 185 5.4.2 Preparation of melatonin and other indoles using enamides 187 5.4.3 Preparation of N-acylcarbinolamines 190 5.4.4 Application of N-acylcarbinolamines in the synthesis of indoles 194 5.5 Structure elucidation of natural products from selected Aloe species 196 5.5.1 Isolation of plicatoloside from A. plicatilis 196 5.5.2 Isolation of new compounds from A. africana, A. speciosa and A. broomii 197 5.5.2.1 Constituents from A. africana 198 5.5.2.2 Constituents from A. speciosa 199 5.5.2.3 Constituents from A. broomii 199 5.6 Isolation of new compounds from wild ginger, Siphonochilus aethiopicus 202 53 References 204

Acknowledgements CHAPTER 1

SELECTED PALLLADIUM CATALYSED REACTIONS IN ORGANIC SYNTHESIS: A BRIEF OVERVIEW

1.1 INTRODUCTION

The renaissance of organometallic catalysis in the 1960's and the subsequent breakthrough of homogeneous organometallic catalysis in laboratory-scale and industrial syntheses have received a major stimulus from palladium coordination chemistry. Palladium is one of the most versatile and efficient catalyst metals in organic synthesis and its complexes are among the most readily available, easily prepared and easily handled of transition metal complexes. 1 "5 The real synthetic utility lies in the wide range of organic transformations promoted by palladium catalysts and in the functionality and functional group tolerance of most of these processes allowing the synthetic chemist a wide choice of starting materials. When utilised with skill and imagination exceptionally efficient total synthesis can be achieved. This chapter will discuss Pd(0) catalysed reactions with synthetic potential, with a special focus on carbonylation chemistry. 63

1.2 THE HECK REACTION

1.2.1 BACKGROUND

Since the pioneering work by Heck and his group on the palladium-catalysed arylation or vinylation of alkenes, the Heck reaction has become one of the most versatile catalytic carbon- carbon bond forming reactions. 2'81°

Pd(0) D i , R1—X + —■ + base.RX Base► Ri = aryl or vinyl, R2 = aryl, alkyl, alkenyl, CO2R (R = alkyl) X = I, Br, OTf SCHEME 1.1: THE HECK-REACTION 2

Despite displaying many of the benefits usually associated with palladium-mediated reactions (for example ease of scale-up and tolerance of water and/or other functional groups), interest in the reaction has been sporadic, largely due to questions regarding the mechanism, the control of substrate selectivity, regio- and stereoselectivity, and the need to improve the efficiency. In recent years, however, the attention paid to the reaction has increased dramatically, as indicated by the large number of publications in this field since the 1980's. Its use as a key step in complex organic synthesis ! 1 '12 as well as the development of enantioselective versions 13-12 are exciting results of recent research with intramolecular i8 and cascade-type l° Heck reactions also offering attractive protocols.

1.2.2 MECHANISM OF THE HECK REACTION

The generally accepted mechanism for the Heck-reaction is shown in Scheme 1.2 and comprises the following steps: (i) oxidative addition, coordination and insertion of the alkene, p-hydride elimination and regeneration of the active palladium(0) species by reductive elimination.

The catalytic cycle starts with the oxidative addition of the organic halide or triflate to the active Pd(0) catalyst, the latter assumed to be a coordinatively unsaturated I4-electron species and normally generated in situ. A 16-electron It 1 -PdL2-X square planar complex in which the Pd-L bond is weak and the Pd-X bond is strong, is formed. It is generally assumed that insertion of palladium into the R I -X bond is the rate determining step and as expected from the C-X bond dissociation energies, the reaction rate increases in the order CI « Br < I, with fluorides being completely unreactive with any of the known catalysts. 2° The scope of the Heck reaction was further extended to carbonyls and phenols by the discovery of trifluoromethanesulphonates (triflates) as effective leaving groups.21'22 3

base.HX oxidative addition reductive elimination PdL R1— X base

g, - hydride H—PdL2-X R1—PdL2-X elimination alkene insertion

R 1..,,,,,,R2 R2

1.3 H PdL2X R 1 PdL2X R1" ')—"< H 141 H Ril 1.2 ' 1"------1.1

C-C bond rotation

R 1 = aryl or vinyl R2 = aryl, alkyl, alkenyl, COOR' etc. X = I, Br, (CI), OTf

SCHEME 1.2: GENERAL MECHANISM FOR THE HECK REACTION

Since chloroarenes do not normally undergo oxidative addition to palladium(0) their activation is of major industrial interest. Activation of the chloroarene as the corresponding electron withdrawing tricarbonyl chromium complexes is promising. 23 Other studies include work on nickel dibromide/sodium iodide assisted activation, 24 immediate generation of the catalytically active species 25 and the use of palladium catalysts sufficiently stable at higher temperatures. 26

The ligand of choice is usually triphenylphosphine, but a variety of other mono- and bidentate phosphines27 and 1,1 0-phenanthroline derivatives 28 have also been employed. High enantioselectivities have been achieved in various asymmetric Heck reactions with 2,2'- bis(diphenylphosphino)-1,1'-binaphthyl (BINAP) and phosphinooxazolines, 29 but the search for more effective catalysts remains a challenge. It is of interest to note that in the case of aryl iodides, ligands are not needed to effect oxidative addition; instead of a Pd-L bond a weak Pd- solvent bond forms.3° 4

The next and most difficult step in the catalytic cycle is association of the alkene and its insertion into the Pd-R 1 bond resulting in the formation of intermediate 1.1. The mechanism for this process remains a matter for conjecture, but recent catalytic and stoichiometric studies 3I have led to deeper insight and postulated two possible mechanisms (see Scheme 1.3): the first involves coordination of the alkene via dissociation of a neutral ligand, while the second involves coordination via dissociation of the counterion X, resulting in the formation of cationic intermediates. After coordination the alkene inserts into the a-aryl or a-alkenyl R 1 -bond via a syn addition.

Lcm

,Pd, ,Pd, RN_/ Pd A R X R X \ X CL

R X ■ ‘ 7—"N @ /Th 1® Ln,LI ® L,Pd,1- L„L ,Pd Pd B R >1

SCHEME 1.3: COORDINATION-INSERTION OF THE ALKENE

The nature of the alkene substrate is important: the initial cationic complex of mechanism B reacts faster with electron-rich alkenes as they are poor rc-acceptors and good a-donors while the neutral complex in mechanism A reacts faster with electron-poor alkenes as they are good r<- acceptors and poor a-donors. The reaction can be directed 32 to follow either route by careful consideration of the nature of the alkene and the leaving group. Aryl and vinyl triflates are generally assumed to follow mechanism B, but addition of halide ions to the reaction mixture will result in replacement of the weak Pd-OTf bond with a strong Pd-I (Br, Cl) bond and thus invoking mechanism A. However, the addition of sequestering silver salts 33 to the reaction of an aryl or vinyl halide results in the replacement of the strong halide with the anionic component of the silver salt to create a labile Pd-Y bond and thus ensuring that mechanism B is followed.

From detailed mechanistic investigations it is evident that the regioselectivity is determined solely by the coordination-insertion step as shown in Scheme 1.3. When mechanism A is followed, 5 steric effects predominate and favour migration of the R group to the less substituted carbon of the alkene, resulting in linear products. However, in the case of olefins bearing heteroatoms (enol ethers, enol amides, vinyl acetates etc.), allylic alcohols and homoallylic alcohols, electronic factors predominate and the reaction proceeds via the cationic mechanism B which leads to the formation of branched products. In fact, coordination of the rc-system results in an increase of the polarisation of the cationic complex with the anionic aryl species migrating selectively onto the carbon with the lower electron density. 34 With cyclic enol ethers and enamides the electronic factors also play an important role, therefore alkylation always occurs regioselectively and a to the heteroatom. 35 It has been successfully applied to various cylic enol ethers and also allows the creation of stereogenic centers by the linking of two sp 2 carbons.36

After the olefin insertion step a rotation around the C-C bond is necessary to bring the p-H atom in close enough proximity to the metal center. Termination of the reaction occurs with 13-hydride elimination to furnish coupled product 1.3 and the hydridopalladium complex, H-PdL2-X. If the alkene does not dissociate fast enough from the Pd(II)-H complex, migration of the double bond will occur via a series of addition-elimination steps resulting in isomerisation of the double bond. This is especially the case for endocyclic alkene products. The competition between 0- and (3'- hydride elimination (see Scheme 1.4) might complicate the regioselectivity, but for the asymmetric Heck-reaction (AHR) to occur 13'-hydride elimination is essential. Since rotation around the C-C bond is necessary for 0-hydride elimination, it might be expected that 0'-hydride elimination is the kinetically more favourable pathway. More significantly, for endocyclic alkenes the a-bond rotation is not sterically feasible, making 0'-hydride elimination the only possibility.9'I7 This explain why most of the reported AHRs forming tertiary centers involves endocyclic alkene substrates. 6

R1>_ 13-hydride R2 \i—R3 elimination product R4 i311 1.4 R1--X + R2"MP—R3 R4 W-hydride 3 elimination product R4

1.5

SCHEME 1.4: HYDRIDE ELIMINATION PRODUCTS

The partial dissociation of the chiral ligand during the neutral mechanism A (Scheme 1.3) would seem to oppose asymmetric induction and, indeed, evidence from most of the asymmetric Heck reactions indicated that conditions favouring the cationic route give the best enantiomeric excesses. In most cases the combination of the triflate leaving group with BINAP as the palladium ligand yielded the best results regarding stereoselectivity, reaction rate and yield. However, to account for the significant exception observed by Overman;' where the neutral pathway is followed in the case of certain special triflates, the possibility that the reaction proceed via a pentacoordinated intermediate, was suggested. 38'39

The catalyst is regenerated after reductive elimination of HX in the presence of base which may typically be trialkylamines or inorganic salts. In some cases good results were obtained using TI(I) or Ag(I) salts with the latter being preferred due to their lower toxicity and double role as enhancers of enantioselectivity. 9

1.2.3 IMPROVEMENTS AND APPLICATIONS OF THE HECK REACTION

The classic conditions for the Heck reaction generally required high temperatures of between 80 and 120 °C. However, many of these reactions can occur under milder conditions, proceeding at room temperature with DMF as solvent and in the presence of quaternary ammonium halides 40 (apparently playing a crucial role in stabilising the palladium(0) complexes 41) or by applying higher pressure.42 Heck reactions of aryldiazonium salts, 43 N-nitroso-N-aryl acetamides 44 and 7 hypervalent iodo compounds45 as starting materials also proceed at room temperature. Recent developments include microwave promoted Heck reactions of p-iodoanisole and methylacrylate carried out with reasonable conversions in just a few minutes (instead of several hours under standard conditions)46 and the use of reverse phase silica as support for the palladium catalysts:" Another protocol employed the thermally stable palladacycles 48a,48b as source for the active palladium(0) catalyst. Their advantages with regard to conventional catalyst mixtures are based on the possible use of more economic aryl halides, high activity at low palladium/ligand ratio (1:1) and improved thermal stability and life time in solution. Scientific and industrial applications include utilising a palladacycle as catalyst in the total syntheses of estrone steroid derivatives,48` the anti-leukemia alkaloid cephalatoxin, 48d the blood coagulation inhibitor DX- 9065a48e and vinylation of 2-bromo-6-methoxynaphthalene with ethylene on pilot plant scale as a key step in the synthesis of Naproxen by Hoechst: 2f

Polar aprotic solvents such as acetonitrile, DMF, dimethylsulfoxide and N-methylpyrrolidone were initially employed in the Heck reaction, but since the pioneering work by Beletskaya 49 in aqueous medium, it seems that the use of water will become increasingly popular in future. 5° As salt additives increase the polarity of the solvent, Hermann et al 8,51,52 reasoned that the Heck reaction could also be performed in ionic liquids. Good results were obtained not only with tetrabutyl ammonium bromide as solvent at 130 °C but also with perfluorinated solvents and perfluorinated phosphines as ligands. 53

The rapid development of new and vastly improved reaction protocols has made it possible to apply the Heck reaction in elegant syntheses of various biologically active compounds. Intramolecular versions abound and has emerged as an extremely and efficient way to make both carbocyclic and heterocyclic rings. A short reaction sequence has been demonstrated by Masters et. al. who applied an intramolecular Heck reaction as a key step in the construction of the highly functionalised tricyle 1.6 containing the basic structure of Taxol (Scheme 1.5).54 8

1.6

SCHEME 1.5: INTRAMOLECULAR HECK RECTION IN THE SYNTHESIS OF TAXOL

A novel strategy to form the critical quaternary carbon of the galanthamine-type alkaloid,

lycoramine, relies on the 6 - exo cyclisation of 1.7 under non-classical Heck reaction conditions to furnish the expected a,I3 unsaturated ketone 1.8 in 50 % yield. 55 Without the addition of tetrabutyl ammonium acetate the ketone was isolated in lower yield (30 %) which corroborates the suggestion that acetate ions favour the formation of highly reactive palladium species. 39

0 Pd(0A02, dppe, PMP 3 steps Bu4NOAc, toluene ,. (t)-lycoramine 70 °C, 5h 0 0 0 0 OMe 50 % OMe 1.7 1.8

SCHEME 1.6: SYNTHESIS OF LYCORAMINE

An example of a new intermolecular application is the vinylation of tertiary allyl alcohols 1.9 giving access to isoprenoid aldehydes 1.10 from the laboratories of Rh8ne-Poulenc, once again underlining revived interest in industry in this catalytic transformation.56

9

Pd(OAc)2 , DMF ORS OR 1 R A92CO3 or TIOAc or AnAc HO + KVZ‘h")ThR1 58-67° C Fi cT1Wmj0R 1 1•9 0.5 - 48 h 33 - 82 %

R = CH3, A./ \ ./* eLerf teCHO

1.10 \51%* 72 - 88 % = Et

SCHEME 1.7: PREPARATION OF ISOPRENOID ALDEHYDES

Since the first reports on the enantioselective version of the Heck reaction it has successfully been developed up to the point where both tertiary and quaternary centers can be generated with ee > 80 %.

OTBDPSpd(OAc)2, (R)-BINAP, K2CO3 THE MeO Me0 60°C, 72h Me 'Iv OTBDPS Me 90 %, 90 %ee, E:Z 21:3

6 steps

MeO Me (-)-epitazocine

SCHEME 1.8: THE ASYMMETRIC HECK REACTION APPLIED TO (-)-EPITAZOCINE 1 0

The bulk of the myriad of examples involve intramolecular reactions which have the advantage of allowing relatively easy control of alkene regiochemistry. I7 The synthesis of (-)-epitazocine57 involve, as a key step, the formation of a benzylic quaternary center by arylation of a trisubstituted alkene and BINAP as the preferred ligand (see Scheme 1.8).

Without a doubt the Heck reaction will remain as one of the several most important and widely used organopalladium reactions along with palladium-catalysed cross-coupling and carbonylation.

1.3 PALLADIUM ASSISTED CROSS-COUPLING REACTIONS

The palladium(0) catalysed cross-coupling reaction s, 59 ' 60 is an extremely powerful tool for generating new carbon-carbon bonds in organic synthesis, being particularly useful for the preparation of biaryls.

1 Pd(0) R—X + M—R R—R1 M—X

R, R1 = alkyl, alkenyl, aryl, alkynyl X = halide, triflate M = main group metal

SCHEME 1.9: THE CROSS-COUPLING REACTION

The unrivaled versatility and applicability of this reaction (Scheme 1.9) in organic synthesis is evidenced by the vast amount of recent publications and reviews. 58-6I

1.3.1 MECHANISTIC CONSIDERATIONS

The generally accepted theory regarding the mechanism of the palladium(0) catalysed cross- coupling reaction includes a sequence of oxidative addition, transmetallation and reductive elimination steps (see Scheme 1.10). R—R 1 Pd(0) reductive R—X elimination oxidative addition

R—pro R—Pd(II)—X Ri (cis) R 1—M

R—Pd(II)—R1 transmetallation (trans) M—X

R, R1 = alkyl, alkenyl, aryl, alkynyl M = main group metal X = halide, triflate

SCHEME 1.10: MECHANISM OF PALLADIUM CATALYSED CROSS-COUPLING REACTION

The oxidative addition of the active palladium(0) catalyst to the electrophile R-X to generate an adduct R-Pd(II)-X experience the same limitations that applies to this step in the Heck reaction i.e. the order of reactivity follows I > OTf > Br >> Cl while [3-hydrogens in the substrate are not well tolerated due to competing 0-hydride elimination. This adduct then undergoes transmetallation with a main group organometallic, M-12 1 resulting in the transfer of the R I group of the latter to palladium(II) in exchange for the halide or triflate, to afford the corresponding dialkylpalladium(II) complex and the main group metal halide or triflate. Although a wide variety of main group metals such as Li, Mg, Zn, Zr, B, Al, Sn, Si, etc., are able to transmetallate to palladium(II), the following discussion will center around those that have been more useful, i.e. the Stille-,58 Suzuki-59 and Negishi-reactions 69 utilising Sn, B and Zn, respectively, as the corresponding main group metal. This step of the reaction, although little understood, is almost invariably the rate determining step and takes place with retention of configuration. It has been postulated that transmetallation is favoured from the more electropositive metal of R I-M to the less electropositive palladium, occurring via a four-centered a-bond metathesis (intermediate 1.11 in Scheme 1.11) which is promoted by an empty orbital on one or both metals ° and being driven to the right by the irreversible conversion of 1.12 to the coupled product R-121.

12

R1 ) L "ta \ R-Pd(II)-X + R1-K4 R P cd1 = R-Pd ' R Pd(II)-R 1 + M-X \ X' L

1.11 1.12

SCHEME 1.11: TRANSMETALLATION AND ISOMERISATION

Once transmetallation has occurred, rearrangement 62 of the trans-diorganopalladium(II) complex 1.12 to the cis complex 1.13 followed by irreversible reductive elimination to afford the cross coupled product and regeneration of the palladium(0) catalyst, takes place rapidly. Since reductive elimination is faster than (3-hydride elimination from the diorganopalladium(II) complex the presence of 13-hydrogens may be present in the R I group transferred from the main group organometallic.

1.3.2 APPLICATIONS OF VARIANTS OF CROSS-COUPLING REACTIONS

Since the first cross-coupling of arylzinc derivatives with aryl halides employing palladium(0) catalysts was reported by Negishi in 1977, 63 the Negishi reaction has established itself as a valuable tool in the construction of unsymmetrical biaryls. The organozinc reagents are among the most efficient reagents for transmetallation to palladium and are tolerant of both steric hindrance64 and functionality,65 including carbonyls. However, the inability to isolate and characterise the arylzinc intermediates which cannot be prepared on a large scale and stored for later use, remains a shortcoming. Despite this disadvantage numerous heterobiaryls, 64d'e'r phenylpyridine derivatives66 and even polymers67 have been prepared via the Negishi reaction More recently, two noteworthy palladium catalysed cross-couplings were reported: i.e. coupling of a-bromoester 1.14 with the organozinc 1.15 on route to gadain 68 ( Scheme 1.12) and coupling 13 of the vinyl nonaflate 1.16 with alkylzinc chloride 1.17 to prepare the perfume ingredient, dihydrojasmone. 69

0

0 0 ZnCI pd( pph3)4 + 0 THE COOMe 0 0 20 °C 1.14 1.15 89 %

0 > 0

gadain

0 O

6(ONf Pd (dba )2, d ppf

WZnCI THE Me Me 65 °C, 20 h 1.16 1.17 dihydrojasmone (62 %)

SCHEME 1.12: NEGISHI REACTIONS IN THE SYNTHESIS OF GADAIN AND DIHYDROJASMONE

The organotin cross-coupling reaction, known as the Stille reaction, has a much broader scope than the Negishi reaction due to the thermal and hydrolytic stability of the organostannanes. Unfortunately this stability also elicits a lower reactivity towards transmetallation and thus, requires other catalysts or additives to increase the rate of transmetallation. Dramatic increases in reaction rate has been observed by changing to poorer donor ligands such as tris(2- furyl)phosphine and triphenylarsine, 7° keeping in mind that if a ligand is too poor a donor it may lead to catalyst decomposition and the precipitation of metallic palladium. The Stille reaction has been used extensively in the preparation of natural products, 71 pharmaceuticals 72 and modem

14 synthetic materials. 73 A novel application74 is found in the synthesis of substituted isocoumarins in excellent yield via a one pot palladium catalysed coupling of 1,1-dibromoalkenes and organostannanes ( Scheme 1.13). The first step involves the Stille reaction of the "E" bromide with the stannane, followed by the oxidative palladium insertion to the "Z" bromide to give intermediate 1.18a. The coordination of palladium to the ester in 1.18a promotes the elimination of methyl bromide to 1.18b and is finally followed by reductive elimination to afford the corresponding isocoumarin 1.19. However, the use of organotin compounds in industry is hampered by their notorious toxicity.

MeO Br PhSnMe3, toluene MeO Ph MeO Br Br —.. CO2Me Pd2(dba)3, TFP 100°C, 20 h CO2Me

MeO Ph MeO

0 0 1.19 (81 %) 1.18b

SCHEME 1.13: SYNTHESIS OF ISOCOUMARINS VIA A STILLE REACTION

The powerful Suzuki coupling employing organoboron compounds as the organometallic counterpart generally proceeds regio- and stereoselective. 75 The availability of the reagents, the mild reaction conditions and the toleration of a broad range of functional groups contribute to the versatility of the reaction. Moreover, the inorganic byproduct of the reaction is non-toxic and easily removed from the reaction, thereby rendering it suitable for laboratory and industrial processes. The reaction has been used extensively in the preparation of a wide variety of biaryls, 76 various aromatic polymers, 77 pharmaceuticals78 and natural products 79 such as the first synthesis ever of the alkaloids, buflavine and 8-0-demethylbuflavine." Timari gi developed a facile five 15 step synthesis of furostifoline based on a Suzuki coupling and regioselective ring closure of the nitrene intermediate generated from the corresponding nitro-compound 1.20 by deoxygenation.

CH3

OH BrCH2CH(OEt) 2, K2CO3 H3PO4, P205, PhCI CH(0E02 140 DMF, 100°C Br °C, 3 h 75 % 51%

THF, -70 °C n-Bu Li, B(OBu)3, H2O CH3 1 0 Br P(OEt)3, reflux NO2 4h Pd(0) NO2 Na2CO3, DME-H20 (HO)2B reflux, 5 h furostifoline (42 %) 1.20 (72 %) 85 %

SCHEME 1.14: TIMARTS APPROACH TO FUROSTIFOLINE

1.4 PALLADIUM(0) CATALYSED REACTION OF ALLYLIC COMPOUNDS

Palladium(0) catalysis is not only restricted to those substrates featuring oxidative addition as the initial step. A wide range of allylic substrates also has the ability to undergo palladium(0) catalysed reactions with nucleophiles, main group organometallics and small molecules such as olefins and carbon monoxide, resulting in a number of synthetically useful processes.' -5

The first step of all of these reactions is the oxidative addition of the allylic C-X system to Pd(0), a step that proceed with clean inversion of configuration at the allylic carbon and produces a a- allylpalladium(II) complex which rapidly collapses to the it-ally1 complex. In the presence of ligands these n-allyl complexes undergo reaction with a wide variety of nucleophiles, most commonly amines or stabilized carbanions. Nucleophilic attack on the n-allylpalladium complex is a very stereoselective process which proceeds with overall retention of configuration, the result of two consecutive inversions. If the process is carried out in the presence of a main group organometallic compound (usually a tin derivative) transmetallation of the n-allyl complex occurs which is followed by reductive elimination to yield the coupled product with overall inversion of 16 stereochemistry. Finally, rc-ally1 complexes undergo both CO (carbon monoxide) and olefin insertion, making the carbonylation (see section 1.5) and olefin chemistry also accessible to allylic substrates. The utilisation of allylic substrates in synthetic chemistry has been extensively reviewed82'83 describing a multitude of novel syntheses of biologically active compounds. It has been shown that the combination of 'metallo-ene'cyclisations with carbonylation reactions" are particularly attractive since they permit the stereoselective formation of a multitude of carbon- carbon bonds in a single operation. The initial insertion of an olefin into the a-allyl palladium intermediate produces a a-alkyl palladium complex 1.21 which inserts CO competitively with f3- elimination, resulting in overall formation of ester 1.22 (Scheme 1.1 5). 85

Ts Ts Pd(dba)2, PPh3 II1 atm CO Ac0 AcOH, 45°C Pd---- Pd 1.21 1CO TS

Me0H

CO2Me 1.22 Pd 77 %

SCHEME 1.15

1.5 CARBONYLATION

1.5.1 HISTORY

The first well-defined carbonylation reaction, aptly named hydroformylation, was discovered sixty years ago by the German chemist, Otto Roelen 6'7'86 while studying the high pressure, cobalt catalysed Fischer-Tropsch synthesis of hydrocarbons from carbon monoxide and hydrogen. He 17 observed that the addition of ethene to the usual feed-gas mixture of carbon monoxide and hydrogen led to the formation of propanal in high yield.

CHO R _CO/H2.. HO

SCHEME 1.16: HYDROFORMYLATION

Several years of extensive research followed which resulted in many industrial applications, e.g. the synthesis of 1-butanal from propene and acrylic acid from acetylene. However, most of the reactions were demanding, proceeding at high temperatures (100 - 300 °C) and pressures (100 - 1000 bar), using expensive autoclave equipment and large quantities of dangerously toxic, volatile and unstable catalysts (Ni(CO)4, Fe(CO)5 and HCo(C0)4). A further complication was the formation of a complex mixture of compounds which required separation. Consequently, carbonylation was scarcely contemplated as a viable synthetic tool in organic chemistry. Fortunately, dramatic changes occurred with the advent of stable, but extremely active catalysts based on organophosphine complexes of rhodium and palladium. Thus, it is now possible to do carbonylation reactions at temperatures below 100 °C, at atmospheric pressure and with small quantities of involatile, air-stable catalyst precursors such as Pd(PPh3)2C12 which are converted to the active catalytic specie in situ. The versatility and increasing sophistication of carbonylation has lead to a wide application in natural product synthesis e.g. in the total synthesis of echinasporin,87 gelsemine,88 glycinoeclepin A89 and many more. 90 The development is still continuing with new applications being published regularly. 91

1.5.2 MECHANISTIC CONSIDERATIONS OF CARBONYLATION

The generally accepted mechanism for carbonylation 8a involves a sequence of reactions starting with the formation of a palladium carbon a-bond, CO insertion to form an acyl palladium complex followed by either reductive elimination or reaction of the acyl complex with an appropriate nucleophile yielding the carbonylated product and the hydridometal complex and finally, elimination of HX to regenerate the original catalyst. 18

The first step i.e. the formation of the cr-organopalladium complex, usually employs one of the following general methods:

coordination and insertion of an alkene into a palladium(0) complex,

H—Pd—X --ow RC H2CH2-Pd—X

transmetallation of palladium(II) salts or organopalladium reagents, 92

_= RPd 4, SnBu3 PdR

ligand directed ortho-palladation 93 - this is a very general process with the only requirement a lone pair of electrons to precoordinate to the metal and conditions conducive to electrophilic aromatic substitution and

PdC12 L 2 NC

/ C1 2

oxidative addition94 of aryl and vinyl halides or triflates which is by far the method of choice.

Pd(0) R—X R—Pd—X X = I, Br, Off

As experienced with the Heck and cross-coupling reactions, the rate of oxidative addition decreases along the sequence I > OTf > Br >> CI >> F (see paragraph 1.2.2.) with the rate being enhanced by the presence of electron-withdrawing groups or additional coordination sites on the aryl ring. 19

base.HX

reductive elimination Pd L2 RcX base- oxidative addition

H—Pd L2 X RiCONu R 1—PdL2-X

NuH R1CO—PdL2-X

CO insertion

SCHEME 1.17: CARBONYLATION MECHANISM VIA OXIDATIVE ADDITION

Despite the divalent character of carbon monoxide, it is not particularly reactive and reaction conditions usually require extreme conditions, energetic reagents or some form of catalysis. However, its ability to coordinate to transition metals increases the susceptibility of carbon monoxide to nucleophilic attack, either by another ligand or a non-coordinated nucleophile. 6 Perhaps the most characteristic reaction of coordinated carbon monoxide is the so-called insertion step in which a carbonyl ligand undergoes concerted intramolecular attack by another ligand. 95 The concerted mechanism of ligand migration requires a cis configuration of the combining ligands and is normally also highly stereospecific.

u R u “.„4...Ri HI ! .... L , L—Pd=C —II' /Pd—C L 0 0

SCHEME 1.18: CARBON MONOXIDE INSERTION

Carbonylation has been shown to proceed with complete retention of configuration (Scheme 1.18) at the migrating carbon. 96 Although the fine detail of the mechanism remains controversia1,97 20 carbon monoxide inserts very readily into the palladium-carbon a-bond, no matter how this a- bond originated, to furnish an acyl palladium complex. Direct reductive elimination of this acyl palladium complex is the simplest of all carbonylations (RX + CO -> RCOX), and although an important feature of the low pressure Monsanto process for acetic acid from methanol, it is rarely observed.98 However, nucleophilic attack on the acyl palladium complex with whichever nucleophile R2Wor RC00") is present, to furnish a variety of acids, esters and amides, is much more frequently encountered. In the case of iodo- and activated bromo-aromatics the rate determining step is the nucleophilic attack on the acyl palladium intermediate. Therefore, the better nucleophilicity of amines renders amidation faster than esterification under similar conditions.99 Stereospecificity in carbonylations of cis or trans vinyl halides is also greater for amidations (100% retention) compared to esterifications (90 - 95 % retention).

Another detailed mechanistic investigation of the palladium catalysed methoxycarbonylations of bromobenzene relied on Fourier transform (FT) infrared spectroscopy. loo Despite the occurrence of carboalkoxy palladium species (L2(X)PdCOOR) in other contexts, no evidence was found for their participation in this transformation. The dominant palladium complex was determined as one without a carbonyl group and the alkoxide ion as the active nucleophilic species. This is consistent with a rate limiting step involving oxidative addition of the halide to palladium(0) followed by CO-insertion and finally, nucleophilic cleavage of the acyl complex by the alkoxide. The reactivity pattern for alkoxy carbonylations, correlating with the dissociation energies of the R-X bond, is: allyl > benzyl > phenyl = methyl > vinyl > propyl > ethyl.

The isolation of doubly carbonylated products under elevated CO pressures could be rationalised by extensive mechanistic studies which indicated that double CO insertion into the palladium- carbon bond does not occur directly. m The key steps are formation of cationic acyl/carbonyl complex 1.23 ( Scheme 1.19), attack by an external nucleophile on the coordinated carbon monoxide in 1.23 followed by reductive elimination to yield the keto-acyl derivative and regeneration of palladium(0). Selectivity for double carbonylation is maximised by high pressures, solvents of low polarity and sterically hindered alcohols and amines.

21

x xe co co jo e CONu RCO—pc1(11) RCO—pd(io U RCO—pc1(11) Ln L n Ln 1.23

SCHEME 1.19: DOUBLE CARBONYLATION

It is of interest to note that metal complexes which catal yse carbonylation also catal yses its reverse, namely decarbonylation, under certain conditions.1°2 Despite operatin g under heterogeneous conditions and requiring high temperatures, palladium on charcoal has been extensively used as decarbonylating agent for aromatic aldeh ydes. One interestin g decarbonylation reaction that needs mentionin g is that of aroylchlorides - an arylpalladium intermediate 1.24 is readily formed and this is sufficientl y longliving to be trapped by insertion of an alkene, followed by 13-hydride elimination to furnish an arylated alkene (see Scheme 1.20). Thus, "Heck arylation" can be achieved using aroyl chlorides in stead of the less accessible iodoarenes.

0 COCI co Pd—CI Pd, + Pd(0) CI

1.24

R2 = HCI -11— CI Pd H \ CO

.4--

SCHEME 1.20: DECARBONYLATION AND HECK-ARYLATION 22

1.5.3 CARBONYLATION OF ALKENES, ALKYNES AND ALLYLIC SUBSTRATES

The most well-known example of carbonylation is the hydroformylation of alkenes. A plethora of metal complexes have been stated to catalyse the hydroformylation reaction with cobalt, rhodium and platinum considered to be the most active. However, very little on the palladium catalysed hydroformylation of alkenes has been published in the open literature as it is mostly described in patents by She11. 163 The catalytic cycle starts with insertion of the alkene into the metal hydride followed by migration of the resulting alkyl ligand to coordinated carbon monoxide. The direction of insertion determines the ratio of linear to branched aldehydes and is directly dependent on the reaction conditions, such as temperature, kinds and amount of phosphine ligands and pressure. 6

When it comes to the palladium catalysed hydrocarboxylation and alkoxycarbonylation of alkenes two different mechanisms, depending on the substrates and reaction conditions, are operative. In the first case the sequence comprises: insertion of the alkene into H-Pd-X, CO insertion and finally, nucleophilic attack to give an acid or ester with regeneration of H-Pd-X while the second cycle starts with CO insertion into a X-Pd-OR' complex followed by alkene insertion and finally, protonolysis to furnish an ester and regenerating the palladium alkoxide. 3 Considerable attention 104(a) has been paid to the control of regioselectivity e.g., the carbonylation of styrene in alcohol affords a linear ester with a bidentate ligand, a branched ester with triphenyl phosphine (creating a new chiral center), while using the chiral ligand, neomenthyl-diphenyldiphosphine leads to reasonable asymmetric induction (ee > 50 %). Alkoxycarbonylations of styrenes are also sensitive to the nature of the counter ion. 1646 Despite all the generalisations it seems clear that every reaction needs its own optimisations.

Alkynes are reactive compounds for palladium(0) catalysed carbonylations with the regioselectivity depending on the reaction conditions. Addition of the ester function to terminal alkynes frequently occurs at the substituted carbon atom to give branched products as illustrated by the synthesis of the carbanepem derivative 1.25 (Scheme 1.21). 6 Carbonylation of ethynyl alcohols also proved to be a very attractive route to a-methylene-y-lactones, 163 a structure widely distributed in natural products. 23

OH Pd, HI, 20 atm CO2Me 65°C, 16 h

1.25 (70 %)

SCHEME 1.21: CARBONYLATION TO AFFORD A CARBAPENEM DERIVATIVE

Although allylic halides are less reactive than aryl or alkenyl halides, they can be successfully carbonylated as shown by the conversion of allyl chloride into 3-butenoic acid in a biphasic system at room temperature and atmospheric pressure using a water soluble ligand. In the presence of a double bond, the carbonylation of the allylic chloride 1.26 is followed by intramolecular insertion of the double bond into the acyl palladium bond, affording the cyclopentenone derivative 1.27 (Scheme 1.22). 106 However, the difficulty experienced with the carbonylation of allylic acetates could be ascribed to faster reductive elimination than CO insertion of the allylpalladium intermediate. 107

Hi iC5.,c_ PdC12(PPh3)2, E13N u c ) "13 6 CO2Me CI Me0H, MaCN CO, 600 psi 0 1.26 1.27 (90%)

SCHEME 1.22: CARBONYLATION OF ALYLLIC CHLORIDES

1.5.4 APPLICATIONS: FROM THE SYNTHESIS OF ALDEHYDES TO CASCADE REACTIONS

The carbonylation of readily available starting materials such as aryl, heteroaryl and vinylic halides represents a large family of related reactions with prominent members being hydrido, hydroxy, alkoxy, amino and double carbonylations. It allows the selective and high-yielding preparation of aldehydes under moderately drastic conditions (100 bar CO/H2 and 100 °C), especially with the iodo compounds owing to the greater ease of oxidative addition of the C-I bond. In the presence of the unique electron-rich, chelating ligand 1,3-bis(di- 24 isopropylphosphine)propane (dippp) even low-pressure palladium catalysed formylation of aryl chlorides to aldehydes occurred. Various other hydride donors such as poly(methylhydrosiloxane) (PMHS), 1°8 sodium formate !" and tributyltin hydride ! I° have been employed in the hydroformylation of various halides and triflates and since the hydride donors are all more reactive than hydrogen, carbonylations proceed at lower pressures and temperatures (typically at 3 bar and 50 - 80 °C). However, in the case of the mild metal hydride reducing agent, tributyltin hydride, direct reduction of the halide or triflate has occurred as a competitive side reaction, but fortunately this could be circumvented with higher CO pressures and the slower addition of the tributyltin hydride. Substituted aryl halides were transformed into the corresponding aldehydes regardless of the electronic nature of substituents. A significant reduction in yield was experienced with ortho nitro substition, presumably due to steric hindrance. Regioselective formylation of allyl halides occur at the less substituted allylic position while retention of geometry at the carbon bond is normally observed.

When the proton source is replaced with a carbanion or a carbonium ion source it is possible to furnish a range of ketones. A potentially useful route to unsymmetrical ketones is the palladium catalysed carbonylative coupling reaction of organic halides with organotin compounds. Since transmetallation is rate limiting, the CO inserts into the oxidative addition product prior to transmetallation resulting in carbonylative coupling. The high yield synthesis of egomaketone 1.28 from prenyl chloride and 3-furanyltrimethyltin (Scheme 1.23) 111 is a demonstration of the viability of this methodology.

SnMe3 Pd CI CO 0 egomaketone 1.28

SCHEME 1.23: SYNTHESIS OF EGOMAKETONE

One of the single most useful features of carbonylation chemistry is the ease of preparation of closely related amides, esters and carboxylic acids from a single substrate utilising a single 25 catalyst precursor. The synthesis of carboxylic acids via a palladium(0) catalysed carbonylation involving a water soluble palladium complex in aqueous base with heptane or anisole as the organic phase is applied in the conversion of benzyl chloride to phenylacetic acid at 50 °C and I bar in a yield of >90 %. In an industrial application of the palladium catalysed carbonylation of 1- (4-isobutylphenyl)ethanol, the nonsteroidal anti-inflammatory pharmaceutical, ibuprofen 1.29 (Scheme 1.24) has been produced on a 3500 ton scale. Recycling of the palladium catalyst is of paramount importance to the economic feasibility of the process, but with careful reaction design, total catalyst turnover numbers (TONs) high above 10 000 have been reached. 112

PdC12(PM13)2 Pd/C, Me0H MEK 5 bar H2 H2O, HCI 30 °C 50 bar CO, 130 °C

ibuprofen 1.29

SCHEME 1.24: INDUSTRIAL SYNTHESIS OF IBUPROFEN

The synthesis of esters and amides via carbonylation has been well studied and a wide variety of improvements on the initial reaction conditions have been claimed using different solvents, bases and special ligands. 113 Carbonylations of aryl-, vinyl-, benzyl and ally] halides as starting materials have been used frequently to generate esters in high yield. The alkoxy- and aminocarbonylations are used most frequently in an intramolecular manner and is applicable to virtually any system having an appropriately situated nucleophile to trap a a-acylpalladium complex and subsequently giving access to a wide variety of heterocycles, 114 such as five-, six- and seven-membered a-methylene lactones, [3- and y-lactones, cyclohexenones, spiro-compounds lactams, aryl-oxazoles, -imidazoles -thiazoles, anhydrides, imides and many more. This carbonylation methodology has been successfully applied as key steps in the synthesis of a number of natural compounds: intramolecularly in the case of the phtalideisoquinoline alkaloid,

26 cordrastine H 114a (Scheme 1.26) and intermolecularly for zearalenone" 5 and curvularin 116 with yields around 70 % for the carbonylation steps in the latter two cases.

Me0 Me0

N\ Me0 Pd(OAc )2 , PP h3 Me0 Me Toluene, reflux - OH K2c03,Tmsci Br CO, 24 h

OMe OMe (t)-cordrastine-II 98 %

OMe OMe 0 HO 0 0 CO 0 zearalenone Me0 Me0 0 SPh SPh

SCHEME 1.26: CARBONYLATION CHEMISTRY APPLIED INTER- AND INTRAMOLECULARLY

The intramolecular palladium catalysed aminocarbonylation reaction has been extended to the synthesis of various antibiotics. A key step in the route to the antitumour antiviral antibiotic, tomaymycin, 117 for example, involves the carbonylation of the intermediate amino-amide 1.30 as shown in Scheme 1.27. It has also been applied successfully to the synthesis of various other azepinone and berbin-8-one derivatives. 118

OMe R2 OMe Pd(OAc)2 OR3 1 , tomaymycin CO N R Br OR3 1.30

SCHEME 1.27: AMINOCARBONYLATION TO TOMAYMYCIN 27

As already demonstrated, a-aryl- and vinylpalladium complexes exhibits a diverse array of reactions, undergoing carbonylation, transmetallation or olefin insertion. In principle these processes can be combined in any order and since many of them generate another a-bonded • organopalladium complex, a sequence of these processes forming several bonds, could be envisaged. The only requirement for these cascade reactions to proceed, is that the next step is more facile than reductive elimination or (3-hydride elimination which terminate the process. The conversion of 1-iodo-1,4-dienes 119 (Scheme 1.28) into the unsaturated cyclic keto-ester 1.31 is a splendid example of the ability of palladium(0) complexes to promote a complex sequence of palladium centered reactions yielding a single product with a high degree of specificity. Two carbonylation steps are involved, the first leading to the formation of a cyclic keto-alkyl ligand and the second to the introduction of an ester function and release of the product from palladium.

0 R CO R 1 CO2Me PdL Me0H R 1.31

SCHEME 1.28: CARBONYLATION OF 1-10D0-1,4-DIENES

In a pentamolecular queuing process designed by Negishi et. a1. 12° a succession of oxidative addition and insertion steps resulted in the formation of five carbon-carbon bonds per catalyst cycle (Scheme 1.29).

5% P dC I2(P Ph3)2

Me0H/CH3CN/PhH 0 Et3N 40 atm H 1000 R CO2Me 40 -70 %

SCHEME 1.29: MULTICASCADE REACTION BY NEGISHI 28

The relative rate of CO insertion and intramolecular carbopalladation is dependent on CO pressure and ring size of the incipient ring in the cyclisation-carbopalladation step. At a CO pressure of 40 atm., carbonylation is faster than 5-exo-trig cyclisation and 13-hydride elimination, as evidenced by the triple carbonylation achieved by Negishi (Scheme 1.29).

Grigg et. a/. 121 recently reported a novel and flexible route to hydroxamic acids (metalloenzyme inhibitors) employing a catalytic termolecular cascade reaction in which the hydroxamic acid functionality is introduced via the palladium catalysed carbonylation of aryl iodides and subsequent trapping of the acyl palladium intermediate with a hydroxylamine derivative. There are endless variations on this theme and it is likely that all of them will be tried.

1.6 CARBONYLATION OF NITRO GROUPS

For the sake of completeness it should be mentioned that the carbonylation of nitro compounds may result in reduction. The carbamic acid derivatives namely, isocyanates, urethanes and urea's are all obtained from the reductive carbonylation of nitroarenes: in the absence of nucleophilic species isocyanates are isolated, in the presence of an alcohol the corresponding urethane and with an amine the appropriate urea's (Scheme 1.30). 122

R1—NHCO2R

R1—NO2 + CO R1—NCO

R1—NHCONHR

SCHEME 1.30: CARBONYLATION OF NITRO GROUPS

Although very useful and already developed more than twenty years ago very little is known about the mechanistic aspects of this area of carbonylation chemistry. Palladium complexes with heterocyclic ligands such as pyridines, isoquinolines and lately, palladium(1,10-phenanthro- line)2(hexafluorophophate)2, have proved to be very efficient catalysts for the high pressure 29 formation of urethanes from nitroarenes as illustrated by the viable two step production process for toluene-2,4-diisocyanate. 123

Application of the reductive carbonylation methodology in organic synthesis resulted in a short, mild and economical route to substituted indoles. If 2-nitrostyrenes are subjected to reductive carbonylation, N-heteroannulation occurs to furnish the corresponding indoles 124 in moderate to excellent yield. It is still not known whether the reaction is catalysed by palladium(0) or palladium(II) as it has been shown that both oxidation states transform 2-nitrostyrenes into the corresponding indoles with almost equal success. Another application is the development of a novel route to fused indoles via two consecutive palladium catalysed reactions: an intramolecular Heck-reaction followed by reductive N-heteroannulation (Scheme 1.31) as reported by Soderberg et al. 1"

Br Pd(OAc)2 Br Pto-to1103, No 2NaH Et3N NO2 NO 48 % 66 % 2

Pd(OAc)2 dppp, DMF CO (60 psi) 120 t

0

N H 78%

SCHEME 1.31: NOVEL ROUTE TO FUSED INDOLES

1.7 SUMMARY AND CONCLUSIONS

The examples discussed in this chapter serve to highlight the potential of selected palladium catalysed transformations such as the Heck, cross-coupling and carbonylation reactions and 30 combinations thereof. It has been shown that palladium promotes inter- and intramolecular reactions in a chemo-, regio- and stereoselective manner to give access to a wide variety of natural products, pharmaceuticals and other modern synthetic materials. The benefit that derives from palladium chemistry in terms of selectivity and functional group toleration makes them interesting and powerful alternatives to many standard organic reactions. Further improvements will be achieved by applying novel low-cost multifunctional building blocks, developing more environmentally friendly reaction conditions such as ambient temperature for the coupling and the development of more active and reusable catalysts. Finally, to quote B. Trost 126 : "Palladium is a catalyst for all seasons."

1.8 REFERENCES

For comprehensive reviews, see: (a) L.S. Hegedus in Organometallics in Synthesis- A Manual; M. Schlosser, Ed.; Wiley: New York, 1994, Chapter 5. (b) L.S. Hegedus, Transition Metals in the Synthesis of Complex Organic Molecules; University Science Books: Mill Valley, 1994. (c) J.P. Collman, L.S. Hegedus, J.R. Norton and R.G. Finke, Principles and Applications of Organotransition Metal Chemistry; University Science Books: Mill Valley, 1987. (d) A. Yamamoto, Organotransition Metal Chemistry; John Wiley: New York, 1986. (e) J. Tsuji, Organic Synthesis by Means of Transition Metal Complexes; Springer-Verlag: Berlin, 1975. R.F. Heck, Palladium Reagents in Organic Synthesis; Academic: New York, 1985. (a) J. Tsuji, Organic Synthesis with Palladium Compounds; Springer-Verlag: Berlin, 1980. (b) J. Tsuji, Synthesis, 1984, 369. 4 J. Tsuji, Synthesis, 1990, 739. (a) J. Tsuji, Acc. Chem. Res., 1969, 2, 144. (b) B.M. Trost and T.R. Verhoeven in Comprehensive Organometallic Chemistry; G. Wilkinson, F.G.A. Stone and E.A. Abel, Eds.; Pergamon: Oxford, 1982; Vol. 8. H.M. Colquhoun, D.J. Thompson and M.V. Twigg in Carbonylation - Direct synthesis of carbonyl compounds; Plenum Press: New York and London, 1991. 31

B. Comills and W.A. Hermann, Eds., Applied Homogeneous Catalysis with Organometallic Compounds, A Comprehensive Handbook in Two Volumes; VCH: Weinheim, 1996. A. de Meijere and F.E. Meyer, Angew.Chem. Int. Ed. Eng., 1994, 33, 2379. W. Cabri and I. Candiani, Acc. Chem. Res., 1995, 28, 2. R.F. Heck, J. Am. Chem. Soc., 1968, 90, 5518. B.M. Trost and J. Jebaratnam, Tetrahedron Lett., 1987, 28, 1611. L.E. Overman, Pure Appl. Chem., 1994, 66, 1423. R. Noyori, Asymmetric Catalysis in Organic Synthesis; I. Ojima, Ed; VCH: Weinheim, 1993. M. Wills, Chem. Soc. Rev., 1995, 177. K. Kondo, M. Sodeoka and M. Shibasaki, J. Org. Chem., 1995, 60, 4322. 0. Loiseleur, P. Meier, A. Pfaitz, Angew. Chem. Int. Ed. Engl., 1996, 35, 200. M. Shibasaki and E.M. Vogl, J. Organomet. Chem., 1999, 576, 1. (a) D.L. Comins, S.P. Joseph and Y.-m. Zhang, Tetrahedron Lett., 1996, 37, 793. (b) D. Bankston, F. Fang, E. Huie and S. Xie, J. Org. Chem., 1999, 64, 3461. G. Dyker and P. Grundt, Tetrahedron Lett., 1996, 37, 619. (a) M. Julia, M. Duteil, C. Grard and E. Kuntz, BuL Soc. Chim. Fr., 1973, 2791. (b) A. Spencer, J. Organomet. Chem., 1984, 270, 115. S. Cacchi, E. Morera and G. Ortar, Tetrahedron Lett., 1984, 25, 2271 and 4821. W.J. Scott, M.R. Pella, K. Sward, S.J. Stoessel and J.K. Stille, J. Org. Chem., 1985, 50, 2302. (a) W.J. Scott, .J. Chem. Soc. Chem. Commun., 1987, 1755. (b) B. Crousse, L.H. Xu, G. Bernadelli and E.P. Kandig, Synlett, 1998, 658. 24 J.J. Bozell and C.E Vogt, J. Am. Chem. Soc., 1988, 110, 2655. W.A. Herrmann, C. Brossmer, K. Ofele, C.P. Reisinger, T. Priermeier, M. Beller and H. Fischer, Angew. Chem. Int. Ed. Engl., 1995, 34, 1844. (a) M.T. Reetz, K. Wanninger and M. Hermes, Chem. Commun., 1997, 535. (b) M.T. Reetz, G. Lohmer and R. Schwickardi, Angew. Chem., 1997, 109, 1559. (c) M.T. Reetz, G. Lohmer and R. Schwickardi, Angew. Chem., 1998, 110, 492. W. Cabri, I. Candiani, A. Bedeschi and R. Santi, J .Org .Chem., 1990, 55, 3654. 32

W. Cabri, I. Candiani, A. Bedeschi, and R. Santi, J.Org.Chem., 1993, 58, 7421. (a) 0. Loiseleur, M. Hayashi, M. Keenan, N. Schmees and A. Pfaltz, J. Organomet. Chem., 1999, 576, 16. (b) M.-J. Jin, J.-A. Jung and S.-H. Kim., Tetrahedron Lett., 1999, 40, 5197. C.-M. Andersson, K. Karabelas, A. Hallberg and C. Andersson, J. Org. Chem., 1985, 50, 3891. (a) W. Cabri, I. Candiani, A. Bedeschi, S. Penco and R. Santi, J. Org. Chem., 1992, 57, 1481. (b) M. Shibasaki, C.D.J. Boden and A. Kojima, Tetrahedron, 1997, 53, 7371. T. Jeffery and M. David, Tetrahedron Lett., 1998, 39, 5751. R.C. Larock, W.H. Gong and E. Baker, Tetrahedron Lett., 1989, 30, 2603. G.D. Daves Jr. and A. Hallberg, Chem. Rev., 1989, 89, 1433. K. Ono, K. Fugami, S. Tanaka and Y. Tamaru, Tetrahedron Lett., 1994, 35, 4133. R.C. Larock and W.H. Gong, J. Org. Chem., 1990, 55, 407. A. Ashimori, T. Matsuura, L.E. Overman and D.J. Poon, J. Org. Chem., 1993, 58, 6949. E.G. Samsel and J.R. Norton, J. Am. Chem. Soc., 1984, 106, 5505. C. Amatore and A. Jutand, J. Organomet. Chem., 1999, 576, 254. (a) T. Jeffery, Tetrahedron, 1985, 26, 2667, J. Chem. Soc. Chem. Commun., 1984, 1287, Synthesis, 1987, 70. (b) R.C. Larock and B.E. Baker, Tetrahedron Lett.,1988, 29, 905. C. Amatore, M. Azzabi and A. Jutland, J. Am. Chem. Soc., 1991, 113, 8375. K. Voigt, U. Schick, F.E. Meyer and A. de Meijere, Synlett, 1994, 189. W. Yong, P. Yi, Z. Zhuangyu and H. Hongwen, Synthesis, 1991, 967. K. Kikukawa, M. Naritomi, G.-X. He, F. Wada and T. Matsuda, J. Org. Chem., 1985, 50, 299. 45 (a) R.M. Moriarty, W.R. Epa and A.K Awasthi,. J. Am. Chem. Soc., 1991, 113, 6315. (b) S.- K. Kang, H.-W. Lee, S.-B. Jang, T.-H. Kim and S.-J Pyun, J. Org. Chem.,1 996, 61, 2604. M. Larhed and A. Halberg, I Org. Chem., 1996, 61, 9582. M. S. Anson, A.R. Mirza, L. Tonics and J.M. J. Williams, Tetrahedron Lett., 1999, 40, 7147. (a) W.A. Hermann, V.P.W. Ham and C.-P. Reisinger, J. Organomet. Chem., 1999, 576, 23 and reference cited therein. (b) F. Miyazaki, K. Yamaguchi and M. Shibasaki, 33

Tetrahedron Lett, 1999, 40, 7379. (c) L.F. Tietze, T. Nobel and M. Spescha, Angew. Chem. Int. Ed. Engl., 1996, 35, 2259. (d) L.F. Tietze, and H. Shirok, Angew. Chem. Int. Ed. Engl., 1997, 36, 1124. (e) C. Kelm-, R. Neidlein, R.A. Engli, H. Brandsetter, R. Kucznierz, H. Leinert, K. Marzenell, K. Strein and W. von der Saal, Hely. Chim. Acta, 1997, 80, 892. (f) M. Beller, A. Tafesh and W.A. Hermann, German Patent, 1996, DE 19503/119 N.A. Bumagin, P.G. More and I.P. Beletskaya, J. Organomet.Chem., 1989, 371, 397. J.P. Genet and M. Savignac, J. Organomet. Chem., 1999, 576, 305. D.E. Kaufmann, M. Nouroozian and H. Henze, Synlett., 1997, 11, 1091. W.A. Hermann and V.P.W. B6hm, J. Organomet. Chem., 1999, 576, 141. J. Moineau, G. Pozzi, S. Quici and D. Sinou, Tetrahedron Lett., 1999, 40, 7683. J.J. Masters, D.K. Kung, W.G. Bornmann and S.J. Danishefsky, Tetrahedron Lett., 1993, 34, 7253. E. Gras, C. Guillou and C. Thal, Tetrahedron Lett., 1999, 40, 9243. H. Bienayme, C. Yezegulian, Tetrahedron, 1994, 50, 3389. T. Takemoto, M. Sodeoka, H. Sasai and M. Shibasaki, J Am. Chem. Soc., 1993, 115, 8477. J.K. Stille, Angew. Chem. Int. Ed. Engl., 1986, 25, 508. (a) A. Suzuki, I Organomet. Chem., 1999, 576, 147. (b) N. Miyaura and A. Suzuki, Chem. Rev., 1995, 2547. (a) E. Negishi, J. Organomet. Chem., 1999, 576, 179. (b) S. Baba and E. Negishi, J. Am. Chem. Soc., 1976, 98, 6729.

F.Diederich and P.J. Stang, Eds., Metal - catalyzed cross - coupling reactions, Wiley-VCH: Weinheim, 1998. J.M. Brown and N.A. Cooley, Chem. Rev., 1988, 88, 1031. E. Negishi, A.O. King and N. Okukado, J. Org. Chem., 1977, 42, 1821. (a) U. Lauk, P. Skrabal and H. Zollinger, Hely. Chim. Acta, 1981, 64, 1847. (b) E.R. Larson and R.A. Raphael, J Chem. Soc. Perkin Trans. 1, 1982, 521. (c) T.R. Hoye and M. Chen, J. Org. Chem., 1996, 61, 7940. (d) A. Minato, K. Tamao, T. Hayashi, K. Suzuki and M. Kumada, Tetrahedron Lett., 1981, 22, 5319. (e) E. Negishi, F. Luo, R. Frisbee and H. Matsushita, Heterocycles, 1982, 18, 117. (f) P.Rocca, F. Marsais, 34

A. Godard, G.Queguiner, L. Adams and B. Alo, Tetrahedron Lett., 1995, 36, 708. (g) S. Gronowitz, A.-B. Hornveldt, V. Kristjansson and T. Musil, Chem. Scripta, 1986, 26, 305. E. Erdik, Tetrahedron, 1992, 48, 9577. (a) T. Sakamoto, Y. Kondo, N. Murata and H. Yamanaka, Tetrahedron Lett.,1992, 33, 5373. (b) T. Sakamoto, Y. Kondo, N. Murata and H. Yamanaka, Tetrahedron, 1993, 49, 9713. (c) C.A. Quesnelle, O.B. Familoni and V. Snieckus, Synlett, 1994, 349. A. Pelter, M. Rowlands and I.H. Jenkins, Tetrahedron Lett., 1987, 28, 5213. (a) R. Rossi, F. Bellina, C. Bechini, L. Mannina and P. Vergamini, Tetrahedron, 1998, 54, 135. (b) R. Rossi, F. Bellina, C. Bechini and L. Mannina, Synlett., 1997, 1061. F. Bellina, D. Ciucci, R. Rossi and P. Vergamini, Tetrahedron, 1999, 55, 2103. (a) V. Farina, Pure & Appl. Chem., 1996, 68, 73. (b) V. Farina and B. Krishnan, J. Am. Chem. Soc., 1991, 113, 9585. (c) E. Shirakawa and T. Hiyama, J. Organomet. Chem., 1999, 576, 169 . (a) A.M. Echavarren and J.K. Stille, J. Am. Chem. Soc., 1987, 109, 5478. (b) A.M. Echavarren and J.K. Stille, J. Am. Chem. Soc., 1988, 110, 4051. (c) M. Alvarez, D. Fernandez and J.A. Joule, Synthesis, 1999, 615. (a) T.R. Kelly and F. Lang, J. Org. Chem., 1996, 61, 4623. (b) A.M. Echavarren, N. Tamayo and D.J. Cardenas, J. Org. Chem., 1994, 59, 6075. (c) A. Batch and R.H. Dodd. Heterocycles, 1999, 50, 875. (a) M. Bochmann and K. Kelley, J. Chem. Soc., Chem. Commun., 1989, 532. (b). T. Yamamoto, Z.-h. Zhou, T. Kanbara, M. Shimura, K. Kizu, T. Maruyama, Y. Nakamura, T. Fukada, B.L. Lee, N. Ooba, S. Tomaru, T. Kurihara, T. Kaino, K. Kubota and S. Sasaki, J. Am. Chem. Soc., 1996, 118, 10389. (c) W.K. Chan, Y. Chen, Z. Peng and L. Yu, J. Am. Chem. Soc., 1993, 115, 11735. L. Wang and W. Shen, Tetrahedron Lett., 1998, 39, 7625. (a) N. Miyaura and A. Suzuki, Chem. Rev., 1995, 95, 2457. (b) X. Bei, H.W. Turner, H.W. Weinberg and A.S. Guram, J. Org. Chem., 1999, 64, 6797. (a) S. Gronowitz and A. Svensson, Isr. J. Chem., 1986, 27, 25. (b) M.P. Cruskie Jr., J.A. Zoltewicz and K.A. Abboud, J Org. Chem., 1995, 60, 7491. (c) R. B. Millar and S. Dugar, Organometallics, 1984, 3, 1261. (d) H. Zhang, F. Xue, T.C.W. Mak and K.S. 35

Chan, J. Org. Chem., 1996, 61, 8002. (e) I. Yokoe, Y. Sugita and Y. Shirataki, Chem. Pharm. Bull,. 1989, 37, 529. (f) S. Chowdhury and P.E. Georghiou, Tetrahedron Lett., 1999, 40, 7599. (a) M. Rehahn, A.D. Schluter, G. Wegner and W.J. Feast, Polymer, 1989, 30, 1054. (b) Y.H. Kim and O.W. Webster, J. Am. Chem. Soc., 1990, 112, 4592. (c) T.M. Miller, T.X. Neenan, R. Zayas and H.E. Bair, J. Am. Chem. Soc,. 1992, 114, 1018. (d) T.I. Wallow and B.M. Novak, J. Am. Chem. Soc., 1991, 113, 7411. (a) T.R. Kelly, A. Garcia, F. Lang, J.J. Walsh, K.V. Bhaskar, M.R. Boyd, R. Gotz, P.A. Keller, R. Walter and G. Bringmann, Tetrahedron Lett., 1994, 35, 7621. (b) T.R. Hoye, M. Chen, L. Mi and O.P. Priest, Tetrahedron Lett., 1994, 35, 8747. (c) V. Upender, D.J. Pollan, J. Liu, P.D. Hobbs, C. Olsen, W. Chao, B. Bowden, J.L. Crase, D.W. Thomas, A. Pandey, J.A. Lawson and Mk Dawson, J. Heterocycl. Chem., 1996, 33, 1371. (e) R.B. Millar and S. Dugar, Tetrahedron Lett., 1989, 30, 297. (a) B. Zhao and V. Snieckus, Tetrahedron Lett., 1991, 32, 5277. (b) X. Wang and V. Snieckus, Tetrahedron Lett., 1991, 32, 4883. P.A. Patil and V. Snieckus, Tetrahedron Lett., 1998, 39, 1325. T. Soos, G. Timari and G. Haj6s, Tetrahedron Lett., 1999, 40, 8607. (a) B.M. Trost, Acc. Chem. Res., 1996, 29, 355. (b) B.M. Trost and D.L. van Vranken, Chem. Rev., 1996, 96, 395 and cited references. A. de Meijere and S. Brase, J. Organomet. Chem., 1999, 576, 88. (a) W. Oppolzer, Pure Appl. Chem., 1990, 62, 1941. (b) W. Oppolzer, Angew. Chem. Int. Ed. Engl., 1989, 28, 38. (c) W. Oppolzer In Comprehensive Organic Synthesis; B.M. Trost, I. Fleming and L.A. Paquette, Eds.; Pergamon: Oxford, 1991; Vol. 5, Chapter 1.2, p. 29. (a) W. Oppolzer, T.H. Keller, M. Bedoya-Zurita and C. Stone, Tetrahedron Lett., 1989, 30, 5883. (b) W. Oppolzer, J.M. Gaudin and T.N. Birkinshaw, Tetrahedron Lett., 1988, 29, 4705. (c) W. Oppolzer and J.M. Gaudin, Hely. Chim. Acta, 1987, 70, 1477. 0. Roelen, Angew. Chem., 1948, A60, 62. A.B. Smith III, G.A. Sulikowski, M..M. Sulikowski and K.J. Fujimoto, J. Am. Chem. Soc., 1992, 114, 2567. W.G. Early, T. Oh and L.E. Overman, Tetrahedron Lett., 1988, 29, 3785. 36

A. Murai, N. Tanimoto, N. Sakamoto and T. Masamune, J. Am. Chem. Soc., 1988, 110, 1985. K. Ritter, Synthesis, 1993, 735. R. Grigg and V. Sridharan, J. Organomet. Chem., 1999, 576, 65. (a) H. Kuhn and W.P. Neumann, Synlett, 1994, 123. (b) M.W. Shair, T. Yoon and S.J. Danishefsky, J. Org. Chem., 1994, 59, 3755. (a) H. Jendrella, Liebigs Ann. Chem., 1991, 295. (b) S.J. Tremont and H.U. Rahman, J Am. Chem. Soc., 1984, 106, 5759. (c) J.C. Clinet and G. Balavione, J. Organomet. Chem., 1991, 405, C29. (d) P. Ballester. M. Capo, X. Garcias and J.M. Saa, J. Org. Chem., 1993, 58, 328. Y. Hatanaka and T. Hiyama, J. Org. Chem., 1988, 53, 918. G.K. Anderson and R.J. Cross, Acc. Chem. Res., 1984, 17, 765. F. Calderazzo and K. Noack, Coord. Chem. Rev., 1966, 1, 118. K.C. Brinkman and J. A. Gladysz, Organometallics, 1984, 3, 147. (a) D. Forster, J. Am. Chem. Soc.,1976, 98, 846. (b) J.F. Roth, J.H. Cradock, A. Hershman and F.E. Paulik, Chem. Technol., 1971, 600. A. Schoenberg and R.F. Heck, J Org. Chem., 1974, 39, 3327. (a) W.R. Moser, A.W. Wang and N.K. Kildahl, J. Am. Chem. Soc., 1988, 110, 2816. (b) W.R. Moser, A.W. Wang and N.K. Kildahl, Chem. Ind., 1990, 40, 137. (a) F. Ozawa, N. Kawasaki, H. Okamoto, T. Yamamoto and A. Yamamoto, Organometallics, 1987, 6, 1640. (b) M. Tanaka, T. Kobayashi, T. Sakakura, H. Itatani, S. Danno and K. Zushi, J. Mol. Catat, 1985, 32, 115. H.U. Blaser and A. Spencer, J Organomet. Chem., 1982, 233, 267. E. Drent, European Patent App.,1987, No. 218,282 A 1 , to Shell. (a) G. Cometti and G.P. Chiusoli, I Organomet. Chem., 1982, 236, C31. (b) M.C. Bonnett, A.L. Monteiro and I. Tkatchenko, J. Mol. Catal. A. Chem., 1999, 143, 131. (a) T.F. Murray, E.G. Samsel, V. Varma and J.R. Norton, J Am. Chem. Soc., 1981, 103, 7520. (b) S. Cacchi, J. Organomet. Chem., 1999, 576, 42. E. Negishi, G. Wu and J.M. Tour, Tetrahedron Lett., 1989, 29, 6745. Y. Yamamoto, M. Akimoto, 0. Saito and A. Yamamoto, Organometallics, 1986, 5, 1559. I. Pri-bar and 0. Buchman, J. Org. Chem., 1984, 49, 4009. Y. Ben-David, M. Portnoy and D. Milstein, J Chem. Soc. Chem. Commun., 1989, 1816. 37

V.P. Baillargeon and J.K. Stine, J. Am. Chem. Soc., 1986, 108, 452. (a) J.K. Stille, Angew.Chem. Int. Ed. Engl., 1986, 25, 508. (b) F.K. Sheffy, J.P. Godschalx and J.K. Stille, J. Am. Chem. Soc., 1984, 106, 4833. (a) J.N. Armor, Appl. Catal., 1991, 78, 141. (b) V. Elango, M.A. Murphy, G.L. Moss, B.L. Smith, K.G. Davenport and G.N. Mott, European Patent App., 1988, No.00 284 310, to Hoechst Celanese Corporation. (c) S. Rittner, A. Schmidt, L.O. Wheeler, G.L. Moss, and E.G. Zey, European Patent App., 1989, No.00 326 027, to Hoechst Celanese Corporation. (a) G.G. Wu, Y. Wong and M. Poirier, Org. Lett., 1999, 1, 745. (b) Y. Kayaki, Y. Noguchi, S. Iwasa, T. Ikariya and R. Noyori, Chem. Commun., 1999, 13, 1235. (c) T. Takahashi, H. Inoue, S. Tomida, T. Doi and A.M. Bray, Tetrahedron Lett., 1999, 40, 7843. (d) A. El-ghyoury and R. Ziessel, Tetrahedron Lett., 1998, 39, 4473. (e) M.A. Ciufolini, J.W. Mitchell and F. Roschangar, Tetrahedron Lett., 1996, 37, 8281. (f) C.S. Cho, L.H. Jiang and S.C. Shim, Chem. Commun., 1999, 29, 2695. (a) K. Orito, M. Miyazawa, R. Kanbayashi, M. Tokuda and H. Suginome, J. Org. Chem., 1999, 64, 6583 and references cited therein. (b) R.J. Perry, B.D. Wilson and R.J. Miller, J. Org. Chem., 1992, 57, 2883. (c) W.Xiao, G. Vaspallo and H. Alper, J. Org. Chem., 1999, 64, 2080. T. Takahashi, T. Nagashima and J. Tsuji, Chem Lett., 1980, 369. T. Takahashi, H. Ikeda and J. Tsuji, Tetrahedron Lett., 1980, 21, 3885. M.M. Mori, Y. Uozomi, M. Kimura and Y. Ban, Tetrahedron, 1986, 42, 3793. (a) D. Roberto and H. Alper, J. Am. Chem. Soc., 1989, 111, 7539. (b) L.S. Trifonof and A.S. Orahovats, Tetrahedron Lett., 1985, 26, 3159. (c) G.D. Pandey and K.P. Tiwari, Tetrahedron, 1981, 37, 1213. J.M. Tour and E. Negishi, J. Am. Chem. Soc., 1985, 107, 8289. C. Coperet, S. Ma and E. Negishi, Angew. Chem. Int. Ed. Engl., 1996, 35, 2125. R. Grigg, J.P. Major, F..M. Martin and M. Whittaker, Tetrahedron Lett., 1999, 40, 7709. M. Toatsu, German Patent, 1976, No. 2 555 557. (a) S.P. Gupte and R.V. Chaudhari, J. Mol. Catal., 1984, 24, 197. (b) R. Santi, A.M. Romano, F. Panella, G. Mestroni, 0. Sessanta and A. Santi, J. Mot Catal. A. Chem., 1999, 144, 41. 38

B.C. S8derberg and J.A. Shriver, J. Org. Chem., 1997, 62, 5838. B.C. Soderberg, S.R. Rector and S.N. O'Neil, Tetrahedron Lett., 1999, 40, 3657. B.M. Trost, a one day course titled, Palladium complexes: catalysts for all seasons, presented on 10 July 2000 at BOSS-8 in Ghent, Belgium. CHAPTER 2

LAVENDAMYCIN AS A SYNTHETIC TARGET

2.1 THE HISTORY OF LAVENDAMYCIN

Streptonigrin l 2.1, Iavendamycin2 2.2 and streptonigrone 3 2.3 are three structurally and biosynthetically related antitumour antibiotics isolated from Streptomyces flocculus, Streptomyces lavenduale and an unidentified Streptomyces species (IA-CAS isolate no. 114), respectively. Each possesses a characteristic, highly functionalised 7-amino-5,8-quinone AB ring system and a fully substituted pyridyl C ring central to its structure. The discovery that streptonigrin and congeners are inhibitors of HIV reverse transcriptase 4 and that lavendamycin displays significant activity against topoisomerase 1 5 has renewed interest in these compounds and has reinforced the need for additional investigation of their medicinal chemistry.

Me0 Me0 N 0 H2N H2N H2N 0 H2N C H3 OH

OCH3 OCH3 OCH3 lavendamycin streptonigrone streptonigrin 2.1 2.2 2.3

SCHEME 2.1: STRUCTURES OF THREE RELATED STREPTONIGR1NOIDS

The structure elucidation of lavendamycin, which was disclosed in 1981, relied exclusively on extensive spectroscopic studies on a limited supply of naturally occurring dark red material. It had a limited solubility in organic solvents which frustrated efforts to grow a single crystal for X- ray analysis. The 1H and 13C NMR (including comparisons with model quinones), IR, high resolution mass and UV spectroscopic studies as well as biosynthetic considerations permitted 40

Doyle et. al. 2 to assign the pentacyclic quinone structure 2.2 to lavendamycin. Thus, in contrast to the earlier structure elucidation of streptonigrin, classical chemical degradative studies played little role in the structure identification of lavendamycin. Unambiguous proof of the proposed lavendamycin structure was accomplished by subsequent total syntheses. 6'7

2.2 APPROACHES TO THE SYNTHESIS OF LAVENDAMYCIN

The first total synthesis of lavendamycin was achieved by Kende and Ebetino 6 in 1984. The route comprises 8 steps from 'readily available' starting materials with an overall yield of 2.3%. A retrosynthetic view of the route is shown in Scheme 2.2. The key synthon 2.4, representing the AB fragment, was obtained by a straightforward preparation from 8-methoxyquinaldic acid by sequential nitration at C-5 and bromination at C-7.

NO, NO

N COOCH3 22 > B N > Br OCH3 OCH3 0 CH3 HN HN

V CH3 NO 0000H3. NH2 N COON OCH3

2.4

SCHEME 2.2: RETROSYNTHETIC APPROACH OF KENDE TO LAVENDAMYCIN

The synthetic strategy revolved around the Bischler-Napieralski reaction between the presumptive biogenetic precursor, (3-methyl tryptophan methyl ester and the quinaldic acid derivative 2.4 to 41 construct ring C, and in the process bridging the two subunits. This was followed by cyclisation to the pentacyclic intermediate as present in the skeleton of lavendamycin and finally, the desired functionalisation of ring A and oxidation to obtain the lavendamycin methyl ester. The same convergent strategy utilising the Bischler-Napieralski reaction was also applied successfully by the group of Rao. 7's However, this procedure has serious drawbacks including the low yields of 30 % and 31 %, respectively, for the construction of the C ring via a Bischler-Napieralski reaction and the introduction of the C-7 amino group in the quinoline ring. The latter step occurs, after oxidation of ring A to a quinone by displacement of the bromine with sodium azide followed by reduction of the resultant azidoquinone to the desired aminoquinone. The apparent use of 'readily available' starting materials is also misleading since the preparation of both 8-methoxyquinaldic acid and n-methyl tryptophan methyl ester adds several additional steps to the synthesis and thereby lowers the overall yield.

Several other groups" -12 also based their total syntheses of lavendamycin on the construction of the C ring. In the first case Behforouz 9'1° and Hibino l 1'12 replaced the Bischler-Napieralski reaction to build the p-carboline moiety with the closely related Pictet-Spengler cyclisation (see Scheme 2.3). This reaction between an aldehyde (instead of an acid) possessing the AB ring functionalities and 13-methyl tryptophan methyl ester furnished the required pentacyclic derivative. As before, final introduction and modification of functional groups by Hibino l 1'12 afforded the target compound in 10 steps. Since the authors did not report yields for the preparation of the starting material, it is difficult to determine the overall yield. The same difficulties regarding the introduction of the C-7 amino group and the so-called 'readily available' starting materials apply to this procedure.

A distinct difference between the approach by Hibino and that of Behforouz is that the last group constructed the quinolinedione AB ring system with the protected C-7 amino group introduced prior to CDE ring formation. A sequence of nitration, reduction, protection and oxidation of 8- hydroxy-2-methylquinoline made it possible to solve the problems experienced by other groups with the introduction of the C-7 amino group later in the synthesis. However, the synthesis of Behforouz utilise reactive and unstable quinone species. The high yield (79 %) reported for the Pictet-Spengler cyclisation is also of interest since the fully aromatic [3-carboline was apparently 42 formed directly, without the need of the additional oxidant as usually required. Though (3-methyl tryptophan methyl ester could be prepared via a 3 step transformation, one of the key reagents, methyl nitroacetate, is quite expensive. Another limitation is a methyl group oxidation with toxic SeQ2 to obtain the appropriate AB ring aldehyde precursor.

r

N COOCH3

C H3 ) HN Ph

,Hibino 2.2 Behforouz % 0

N COOCH3 RCONH >RCONH 0 C H3 HN

N CH3 RCONH N CHO OH 0

SCHEME 2.3: PICTET-SPENGLER CYCLISATION IN THE SYNTHESIS OF LAVENDAMYCIN

In the second approach to construct the C ring, Ciufolini 4 employed a modified Knoevenagel- Stobbe pyridine ring construction and a thermolytic nitrene insertion as key steps in a new

43 synthesis of lavendamycin methyl ester. A retrosynthetic analysis of their proposal is presented in Scheme 2.4. Condensation of quinoline 2.5 with 2-azidobenzaldehyde provided the chalcone 2.6 followed by a Michael addition to furnish adduct 2.7 exclusively. Treatment of this adduct 2.7 with hydroxylamine hydrochloride resulted in the formation of the pyridine derivative 2.8 which was subjected to a selenium dioxide oxidation to give 2.9 followed by a thermolytic nitrene insertion to yield the corresponding f3-carboline. The complete molecular framework of lavendamycin was thus assembled from three building blocks in five steps and a reasonable yield of 62 %. The 13-carboline intermediate could be converted into the final product by functional group transformations on rings A and C. In contrast to the present focus on convergent syntheses, this protocol comprises a linear approach. Other objections to the viability of this route would include, amongst others, the introduction of the C-7- amino group, the use of a toxic selenium dioxide oxidation and the so-called availability of the starting materials.

22 > Br C H3

2.8 Z = CH3 2.7 2.9 Z = CHO

MeCH=C(Me)0Etl. 2:1 CH2=C(Et)OEt Br

2.5 2.6

SCHEME 2.4: CIUFOLINIS STRATEGY TO LAVENDAMYCIN

In a third illustration of C ring formation Molina n et. al. assembled the lavendamycin carbon skeleton in an elegant procedure involving reaction of the aldehyde 2.10 with iminophosphorane 44

2.11 (generated in situ from the corresponding azide and tributylphosphine) in a sealed tube at 165 °C in the presence of Pd/C (see Scheme 2.5 for a retrosynthetic approach). The conversion from 2.10 to the 1-substituted 13-carboline (45 % yield) involves an initial aza-Wittig reaction to give a 2-azahexatriene which subsequently undergoes electrocyclic ring closure followed by further dehydrogenation under the reaction conditions. The presence of the bromine group at C-7 is subsequently exploited to introduce functionality to eventually render the aminoquinone A ring as previously described by Boger. This constitutes a simple and direct formal synthesis of lavendamycin ethyl ester, but despite all the advantages by employing an aza-Wittig reaction as a key step, a major stumbling block remains the introduction of the 7-amino group and the availability of the starting material.

CH3 COOCH2CH3 N=PPh3 22 >Br B OBn CH3 2.10 2.11

V CH3 COOCH2CH3

N3

C H3

SCHEME 2.5: RETROSYNTHESIS VIA IMINOPHOSPHORANE

Bogeris 14 group preferred to extend the inverse electron demand Diels-Alder reaction, successfully demonstrated for streptonigrin, to the synthesis of lavendamycin as shown in Scheme 2.6. Thus, formation of the pyridine C ring and attachment of the E ring was achieved in one step via a [4+2] cycloaddition between pyrrolidine enamine 2.12 and the electron deficient 1,2,4-triazine 2.13. The notable feature of this strategy is the palladium(0) catalysed oxidative insertion into the aryl halide bond and subsequent cyclisation to provide the P-carboline derivative 2.14. This was utilised in a Friedlander reaction with the appropriately substituted aldehyde 2.15 to assemble the

45

carbon skeleton of lavendamycin. As before, final introduction and modification of functional groups afforded the target compound in 15 steps and an overall yield of 3.2 %. However, this strategy rely on a linear synthesis with the main contributors to its inefficiency arising from the construction of the C ring and the manipulations to introduce the C-7 amino group.

B CHO 22 > > + B NH2 OBn

2.15 2.14

V O EtO2C N CO 2Et CH3 EtO2C...N..,,,,,,..0O2Et + I NI < HOC . CH3 EtO2C-M(r Br

2.12 2.13

SCHEME 2.6: BOGERS TACTICS TO LAVENDAMYCIN

A totally different retrosynthetic analysis (Scheme 2.7) employing modern advances in synthetic chemistry such as directed ortho metalation (DoM) and transition metal catalysed cross-coupling reactions was proposed by a French group. 15'16 They suggested a convergent synthesis obtaining lavendamycin from benzene, pyridine and quinoline building blocks in a few key steps: a Suzuki cross-coupling between 2.16 and 2.17, a Stille cross-coupling of the previously formed 4- arylpyridine intermediate with stannane 2.18 and an indole cyclisation followed by oxidation of the dimethoxyquinoline moiety to furnish the pentacyclic framework of lavendamycin 2.19 in a yield of 39 %, starting from stannane 2.18. The introduction and modification of functional groups in the appropriate positions are still required to complete the synthesis. This strategy emphasise the importance of the utilisation of preformed and largely prefunctionalised aromatic rings to form complex polycyclic systems. However, the difficulty with the introduction of the amino group in position C-7 still takes place at the end of the synthesis as reported previously. 46

OCH3

N CH3 22 OCH3 F CH3 tBuOCHN

2.19 V

OCH3 tBuOCHN CH3

CH3 N SnMe3 0 C H3 2.16 2.17 2.18

SCHEME2.7: RETROSYNTHETIC VIEW OF THE FRENCH GROUP'S SYNTHESIS

These five formal syntheses as well as the other approaches to lavendamycin, 4"16 are certainly ingenious and definitely highlight the challenges and the intrinsic difficulties involved in the synthesis of such a highly functionalised molecule. The application of these strategies on large scale is still being hampered by the fact that the discussed protocols require far too many steps, involve complex ring-forming reactions and suffer from low overall yields. In addition to the low yields encountered the authors claim their yields form 'readily available' starting materials. Although the preparation of these starting materials require several steps of simple chemistry, many are low yielding, thereby lowering the overall yield. A critical analysis of the different strategies employed, revealed a few imperfections which could be summarised as follows: (i) construction of the C ring with all its functional groups, (ii) introduction of the amino group into the AB quinoline-5,8-dione structure and (iii) a major concern: the lack in versatility and generality of the various strategies to give access to other lavendamycin analogues. Ease of access to other lavendamycin derivatives would allow researchers to establish structure-activity relationships, hopefully leading to the design of an antibiotic with minimum chemical 47 functionality, optimal activity and negligible toxicity. Preclusion of the potential clinical use of the streptonigrinoids due to their potent toxicity necessitates the development of such a general synthetic strategy even though a chemical synthesis of such a complex molecule is unlikely to be competitive with a biochemical synthesis.

The fascinating features and interesting biological activity of lavendamycin led us to undertake the formidable challenge of the synthesis of such a complex molecule with a view to develop a general route to the parent compound and its analogues. In designing a new strategy we had to consider the problems experienced by previous researchers (vida supra). This allowed us to suggest that the main difficulties to be addressed were the preparation of an appropriate AB quinoline equivalent, which could readily be converted into the desired quinone and the development of a simple route to the CDE ring structure. The initially proposed strategy also took cognisance of the existing expertise in palladium chemistry at the Rand Afrikaans University as exemplified in the approach to the synthesis of streptonigrin.' 7 Thus, the basic lavendamycin framework can be assembled involving the independent elaboration of simple building blocks via a series of palladium catalysed key reactions as indicated in Scheme 2.8.

Heck reaction Cross-coupling

6

7 N • COX

5 Amino or hydroxy carbonylation

Cross-coupling Pd(0) catalyse cyclisation

SCHEME 2.8 : Pd-CATALYSED REACTIONS IN THE BASIC STRATEGY FOR LAVENDAMYCIN SYNTHESIS

The basic strategy therefore involves four key features : firstly, a Heck reaction of an appropriate A ring precursor with methyl acrylate, followed by cyclisation to form the AB ring system; secondly, the construction of the CDE ring system comprising the following steps: (a) a cross 48 coupling between appropriate C and E ring precursors, (b) a suitable cyclisation step to construct the D ring and (c) amino or hydroxy carbonylation to introduce a suitable DoM group into the D ring, thereby completing the CDE ring structure. The third feature will involve a second cross- coupling reaction between the AB and CDE moieties which will be followed by the fourth feature of the synthesis i.e. the final introduction and/or modification of functional groups to afford the target compound. Since the final oxidation procedures to the 5,8-quinone have been well described in the literature, 6'7'14 our overall aim regarding a target molecule could be described as the preparation of a pentacyclic lavendamycin derivative containing the required 7-amino substituent and with functionality at either or both the 5 and 8 positions to convert to the 5,8- quinone. The presence of a suitable DoM group at the 6' position will allow the possibility to introduce a wide variety of groups in the 5' position giving access to several lavendamycin analogues.

2.3 CROSS-COUPLING REACTIONS IN THE SYNTHESIS OF MULTICYCLES

A pivotal step in the proposed strategy towards lavendamycin is the linking of the AB fragment with a CE or CDE fragment via a cross-coupling reaction. Fortunately, good results were achieved in our laboratory' s with the Stille cross-coupling of 2-trimethylstannylquinoline to 2-chloro-3- nitropyridine to afford 2.19, suggesting that the desired reaction between the appropriate AB and CD equivalents to create the B-C biaryl axis of streptonigrin would be successful.

Pd2dba3.CHCI3, P(04003

dioxane, 95° C, 24 h

2.19 ( 67 %)

SCHEME 2.9: AN EXAMPLE OF A STILLE CROSS-COUPLING 49

However, the above reaction and other model work indicated that each cross-coupling reaction required individual optimisation. Godard et. a/. 15'16 have also reported similar experiences with their couplings on a variety of substituted quinolines and pyridines. The extensive use of palladium catalysed coupling reactions to produce substituted 13-carbolines from the corresponding halogenated derivatives were illustrated by Bracher 19 e.g. 1-chloro-13-carboline 2.21a was coupled with phenylboronic acid to produce 2.20 in excellent yield (see Scheme 2.10). It has also been shown that the bromine analog, 2.21b can be used to prepare the lithiated-8-carboline 2.22 which in turn can be reacted with any number of electrophiles. Transmetallation of 2.22 with zinc chloride produces an organozinc species which can undergo palladium catalysed coupling with 2- chloroquinoline to provide nitramarine 19c 2.23 in 53 % yield, the latter being a natural product with hypnotic and hypotensive properties.

Ph X = CI PhB(OH)2 Pd(PPh3)4 2.20 (87 %)

N

'1 X H 2.21a X = CI \ =Br( 1/4 2.21b X = Br KH, t- BuLi

CI

Pd(P Ph3)4

2.22

2.23 (53 %) nitramarine

SCHEME 2.10: THE UTILISATION OF CROSS-COUPLING BY BRACHER 50

Therefore, since the important coupling of the AB and CDE or CE ring systems appears to pose no obvious problems, further attention towards the synthesis of lavendamycin will focus on the synthesis of suitable AB and CDE ring fragments, respectively.

2.4 PREPARATION OF THE LAVENDAMYCIN AB MOIETY

2.4.1 THE EVALUATION OF RING SYNTHETIC STRATEGIES

As previously discussed the structural dissection of the lavendamycin molecule revealed a quinolinequinone (AB) as part of its framework. Since quinolines are historically known and many exhibit significant biological activity they have received much interest over the years which resulted in the development of numerous methodologies to synthesise substituted quinolines. 2° These old and established methods include, amongst others, the widely used Skraup synthesis (1880), the Friedlander reaction (1882), the Pfitzinger synthesis (1886) and the cyclisation of cinnamanilides (1927) and ortho-aminocinnamic acids (1852), respectively. Using these standard procedures several approaches to the lavendamycin AB ring structure have been reported as illustrated by Kende and Ebetino's 6 employment of 2-amino-3-methoxybenzaldehyde in a Friedlander condensation with pyruvic acid to furnish the required 8-methoxy quinaldic acid 2.24 in a yield of 86 % (Scheme 8.11).

0 CHO )COON

NH2 Na001-13, CH3OH COOH OMe reflux, 16 h OMe 2.24 (86 %)

SCHEME 2.11: FRIEDLaNDER CONDENSATION TO 8-METHOXYQUINALDIC ACID

The Friedlander condensation was also employed by Boger 14 and Molina° in their approaches to the lavendamycin AB ring basing their selection of 2-amino-3-(benzyloxy)-4-bromobenzaldehyde 2.25 as the appropriate starting material on recognised observations that substrates possessing two 51 substituents ortho to the aryl aldehyde suffer from low to modest yields in this particular reaction. Thus, the choice of 2.25 represents one in which the 2-aminobenzaldehyde possesses suitable functionality for penultimate 7-aminoquinoline-5,8-quinone introduction and one in which the Friedlander condensation could be anticipated to proceed uneventfully. However, the amino aldehyde 2.25 had to be synthesised in a seven step protocol starting from 3-hydroxy benzoic acid (Scheme 2.12). A modified version of the above route to construct the required AB ring system was also applied by Ciufolini. 4

COOH COOCHt COOC H3 Br2, HOPI, HNO3 NaH, DMF, 0°C, 10 min HCI, CH3OH 1.1 PhCH2Br, 0 -25°C, 20 h Br NO2 OH OCH2Ph 85% 3 LiBH4, THF, 21 h 1.5 PDC, CH 2Cl2 11h CHO Friedlander CHO 5 Na2S204 condensation H20, THF NO2 NH2 60 °C, 0.5 h Br OCH2Ph OCH2Ph 2.25 77 - 80 % 0c02043 93 % R = CH3 ,

SCHEME 2.12: UTILISATION OF THE FRIEDLaNDER CONDENSATION

One of the strategies utilised by Behforouz l°a in a lavendamycin AB moiety synthesis relied on a novel intermolecular Diels-Alder condensation of a siloxy-activated 1-azadiene 2.26 with bromoquinone 2.27, the latter being prepared in three steps from the commercially available 2,4- dibromo-6-nitrophenol. The final step to the desired aldehyde 2.28 requires an oxidation with the toxic substance, selenium dioxide. 52

C6H5CI reflux, 22 h

02N Br AcHN AcHN OH 0 0 0 2.27 66% Sr

2.26 SeO2 reflux 9h

AcHN CHO 0 2.28 91%

SCHEME 2. 13: THE DIELS-ALDER APPROACH OF BEHFOROUZ TO QUINOLINEDIONES

In the other strategies towards the AB ring system of lavendamycin the researchers opted to employ commercially available preformed rings which could be modified by a sequence of appropriate transformations to the suitable AB ring precursor. The approach by Rao 8 focused on the use of the available 8-hydroxyquinoline as starting material by following a tedious procedure of 11 steps involving several oxidation and reduction reactions, as illustrated in Scheme 2.14. The required quinaldic acid 2.29 was obtained in an overall yield of 10.8 %. 53

H202 (CH30)2S0 2 AcOH NJ No OH OH Os OH OCH3 83% OSO3CH3

1NaCN NHAc NO Na2S2O4 tAs2E)

OCH3 Na2S2O4 (CH30)2S02, N Br N CN K2CO3 0 OCH3 61 % 53%

1 KOH

OCH3

N CO2H OCH3 2.29 86%

SCHEME 2.14: THE APPROACH OF RAO TO A SUBSTITUTED QUINALDIC ACID 54

Although Behforouz ieb managed to shorten the route to 8 steps by employing the commercially available 8-hydroxy-2-methylquinoline and using existing literature 21 methods for similar systems, the overall yield was still only 10 %. He bypassed the problems revolving around the introduction of the amino group at C-7 by the introduction of a dinitration step followed by reduction early in the synthesis. However, the final oxidation to the 2-formylquinoline derivative 2.30 still involves the utilisation of the toxic selenium dioxide as oxidant.

NO2 NHCOR

HNO3 ,H2S041 (i) H 2/Pd-C, HCI-H 20 (n) (RCO)20, RCO2Na N CH3 02N N CH3 Na2S 03 ROCHN OH OH (iii) Me0H- H2O OH 75 - 100% 72 % K2Cr207 1 HOAc

C H3 .p...„...1%.„,.....12904, H2O

ROCHN N CH3 0 73 - 94 %

4e02 Dioxane, H2O reflux H2N N CHO 0 2.30 91 %

SCHEME 2.15: THE STRATEGY OF BEHFOROUZ USING A PREFORMED AB RING

The examples discussed above clearly demonstrates that the preparation of such highly functionalised aromatic compounds is not only tedious and difficult, but also suffer from low overall yields. Although the strategies utilising preformed ring systems experienced similar shortcomings than the approaches based on the construction of the B ring, a major improvement appeared to be the introduction of the required amino group via a combination of dinitration and 55 subsequent reduction steps (see Scheme 2.15). Even the introduction of functionality at position 2 appeared to be problematic and/or indirect thereby confirming the need for an improved synthesis of substituted quinolines.

Thus, our initial aim was to construct an appropriately substituted quinoline ring featuring the following: (i) functionality at either or both the 5 and 8 positions of the quinoline ring which would give access to the 5,8 quinone, the presence of the 7-amino substituent and functionality such as a halide or a metal at the 2 position ready for cross- coupling with the CDE ring structure.

2.4.2 THE ANALYSIS AND EVALUATION OF NEW STRATEGIES

All approaches to the total synthesis of natural products and other complex structures are characterised by the necessity to abandon many promising approaches due to unforeseen complications with some envisaged chemical transformations. This perennial problem of synthesis is due to the fact that every molecule is more than the sum of its parts which results in the unpredictability of the exact chemical behaviour of a multifunctional compound. It is therefore understandable that, similarly, in our own work, many attempted chemical transformations from a synthetic point of view, failed, giving either the wrong compound or gave the sought after product in a very low yield or yielded no identifiable products. We deliberately decided not to include reports on such failed reactions in the experimental part of this thesis.

It was envisaged that a substituted 2-hydroxyquinoline would be the ideal synthetic precursor, because the hydroxy group could be readily converted to either the corresponding halide or to the corresponding quinolylmetal derivative, the latter via a lithium-halogen exchange. It is known that 2-hydroxyquinolines can be obtained by the cyclisation of cinnamanilides and o-aminocinnamates while the latter can be readily prepared via a Heck reaction22 between aryl halides containing an

ortho- amino or nitro group, and methylacrylate. Work in the RAU laboratories expanded the 56 approach to include o-nitro-aryltriflates. 17b However, the Heck reactions of compounds containing amino groups are inferior to those of the corresponding nitro compounds in which case the reaction rate is enhanced by the presence of the electron withdrawing nitro group. The application of this methodology was aptly demonstrated in our laboratories at RAU which resulted in the preparation of the substituted 2-hydroxyquinoline 2.32, the required AB synthon of streptonigrin 2.1. 17b This protocol utilised 3,5-dimethoxyphenol 2.31a as starting material and involved the successful nitration of the electron rich aromatic followed by triflation, a Heck reaction under neutral conditions and finally, reduction and subsequent cyclisation of 2.3 lb to furnish the desired 2- hydroxyquinoline 2.32 as a suitable AB synthon for streptonigrin (as illustrated in Scheme 2.16).

MeO Me0 OH Me0 OTf OH2 No2BF4 1120, Et3N DME CH2Cl2, 0 °C -50° C NO2 NO2 NO2 NO2 OMe OMe OMe 2.31a 65 % 82 % Pd(PPh3)4

epichicro- hydrin, 95°C MeO MeO CO2Me SICb.2H20 abs EtOH H2N N OH NO2 NO2 70 °C OMe OMe 2.32 (92 %) 2.31b (73 %)

SCHEME 2.16: PREPARATION OF STREPTONIGRIN AB RING SYSTEM PRECURSOR

This facile synthetic methodology was also applied to the other substrates 2.33a-e to furnish 2- hydroxyquinolines 2.34 as shown in Scheme 2.17. 57

R1 CO21vb, Pd(0), Et3N (H) SnC12.2H20, abs EtOH

70°C NO2 R2 N OH (iii) 3M HCI, reflux 2.33 2.34 (42 - 62 `)/0) X= Br, R 1 = R2 = H X= Br, R 1 = Me, R2 = H X = OTf, R1 = R2 = H X = OTf, R 1 = H, R2 = Pri X = OTf, R 1 = H, R2 = CO2Me

SCHEME 2.17: SYNTHESIS OF 2-HYDROXYQUINOLINES

Although the preparation of 2-hydroxyquinoline 2.32 represented the only example of a dinitro aromatic subjected to this protocol, it was envisaged that extension of the methodology to an appropriate dinitro starting substrate would readily provide the required 2-hydroxyquinoline as AB ring precursor of lavendamycin. The selection of suitable precursors revolved around the following requirements: (i) the presence of one or two methoxy and/or hydroxy groups which would end up in the 5- and/or 8-positions in the quinoline ring which would be readily oxidisable to the desired 5,8-quinone moiety, two nitro substituents meta orientated towards each other or the presence of substituents which would direct nitration to the required positions and the presence of a suitable leaving group such as a halide or a triflate to take part in a Heck reaction or the possibility of introducing such a leaving group.

Various substrates which were either commercially available or could be easily prepared were considered as possible starting materials for the preparation of the AB ring structure of lavendamycin by taking the above considerations into account (see Scheme 2.18).

58

OMe OH OMe OH 02N NO2 OH OH CI

02N NO2 02N NO2 02N NO 2 02N NO2 OMe OMe OMe 2.35 2.36 2.37 2.38 2.39

SCHEME 2.18: SELECTION OF POTENTIAL LAVENDAMYCIN SYNTHETIC PRECURSORS

It was reasoned that the dinitrated phenol 2.35 would be the ideal substrate for our purposes. However, the required precursor 2,5-dimethoxyphenol is not commercially available, while the inherent preference for oxidation of such compounds to the corresponding quinones under nitrating conditions, could cause additional problems. Another possible candidate 2.36 could not be prepared since the attempted dinitration of 4-methoxyphenol furnished only 4-methoxy-2- nitrophenol (62 - 67 %) and oxidation products."'

OH OH OH OH NO2 NHCOR NHCOR Nitration Acylation DoM- Reduction methodology X OMe OMe OMe OMe X = Br or I (i) Heck reaction 1 (ii) H4- OH OH (i) Nitration OH AcHN CI - CI (10 Acetylation POCI3 OH 4E- (iii) Reduction

OMe OMe OMe

SCHEME 2.19: A POSSIBLE UTILISATION OF 4-METHOXYPHENOL AS SUBSTRATE

59

Although it would be possible to employ this mono nitro compound on route to a suitably substituted 2-hydroxyquinoline as illustrated in Scheme 2.19, it requires too many steps with the added drawback that the introduction of the C-7 amino group at a later stage in the synthesis requires additional nitration and reduction steps. This supports our proposal that both the nitro groups should preferably be present in the initial substrate which allows for a shorter and more elegant route to the desired AB ring system of lavendamycin.

It was envisaged that the difficulties, as discussed above, could be circumvented by utilising the dominant activating properties of the hydroxy substituent of 4-methoxyphenol by directly brominating it to 2-bromo-4-methoxyphenol. The subsequent protection of the latter compound as the corresponding O-pivaloyl or 0-acetyl derivative would diminish the dominating effect of the hydroxyl group to such an extent that dinitration would most probably occur ortho to the methoxy group. It was expected that a Heck reaction followed by protection, reduction and concomitant cyclisation would furnish the 2-hydroxyquinoline 2.40 in 7 steps as shown in Scheme 2.20 thereby providing a first possible substrate to the AB fragment of our target molecule. However, experience has shown the impracticality of preparing sterically crowded dinitro compounds by nitration of electron rich aromatics.

OH OH OCOR OCOR Br Bromination (i) Protection Heck reaction (ii) Dinitration 02N NO2 02N NO2 OMe OMe OMe OMe 0) Reduction 1 (ii) Cyclisation OCOR

H2N N OH OMe 2.40

SCHEME 2.20: 4-METHOXYPHENOL AS POSSIBLE SUBSTRATE FOR THE LAVENDAMYCIN LEFT HAND SIDE

60

The dinitrated phenol 2.37 (see Scheme 2.18) was considered as alternative starting material Unfortunately, the dinitration of 3-methoxyphenol favours the formation of 5-methoxy-2,4- dinitrophenol im rather than the desired product 2.37 due to the dominant ortho-directing effect of the hydroxy substituent in conjunction with steric constraints. In contrast, attempts to prepare 2.38 by dinitration of guiacol resulted mainly in oxidation. It is possible that the more expensive 2,6- dinitrophenol could act as an alternative starting material, provided that a suitable halide (required for the Heck reaction) could be introduced into the 3-position. It was envisaged that protection of 2,6-dinitrophenol as the corresponding 0-acetyl or O-pivaloyl derivative followed by selective reduction of one of the nitro groups would result in the formation of 2.41 due to transfer of the acetyl to the amine under the reaction conditions. The O-acylation of the deprotected hydroxyl group would then be followed by direct bromination expected to take place para to the amide group due to its stronger directing properties and steric crowding at the ortho position. The next step would involve a Heck reaction followed by reduction and subsequent cyclisation steps to furnish the required 2-hydroxyquinoline (see Scheme 2.21) in seven steps.

OH OCOR OH OR2 NO2 NO2 NO2 NO2 NHCOR (0 NHCOR Selective 02N Protection 02N Protection reduction (ii) Halogenation Br 2.41 1 Heck reaction

0 R2 OR2 0 R2 02N NHCOR NHCOR H2N NHCOR Reduction

Me00C Me00C

SCHEME 2.21: PROPOSED PREPARATION OF 2-HYDROXYQUINOLINES FROM 2,6- DINITROPHENOL

The dinitroguiacol 2.38 was considered as a possible substrate for the AB ring system. Dinitration of guiacol was expected not to present difficulties, since Kametani and Ogasawara 23 subjected

61 guiacol 2.42 to the method of Hindmarsch et. a/.24 and managed to prepare 5-bromo-4,6- dinitroguiacol 2.43 in 4 steps (Scheme 2.22).

OMe OMe OH ( i ) Ac20, pyridine °M (i) NaOH, Et0H (ii) Br2, CHCI3 (ii) HNO3/AcOH, CCI4 02N Br 2.42 2.43

SCHEME 2.22: DINITRATION OF 5-BROMOGUIACOL

It was thus envisaged that an appropriate 2-hydroxyquinoline 2.44 could be prepared from guiacol 2.42 in five steps. These steps would comprise dinitration, triflation, a Heck reaction with methyl acrylate, reduction and subsequent cyclisation. It may also be possible to prepare quinoline 2.44 from the commercially available 3,5-dinitroanisole by direct halogenation to introduce a suitable halide followed by a Heck reaction, reduction and cyclisation (Scheme 2.23). However, it was decided that the high cost of 3,5-dinitroanisole precluded its use as a starting material in a multistep synthesis. OMe OMe OMe CO2Me Halogenation Heck --).reaction 02N NO2 02N NO2 02N NO2 Reduction and 1 cyclisation OMe

H2N v N OH 2.44

SCHEME 2.23: A POSSIBLE ROUTE TO QUINOLINE 2.44 FROM 3,5-DINITRO ANISOLE 62

The inexpensive 2,4-dinitrochlorobenzene 2.39 was also considered as a possible substrate. However, hydroxylation of nitroarenes with alkyl hydroperoxide anions via vicarious nucleophilic substitution(VNS)25 of hydrogen usually afford products of hydroxylation in the ortho and para positions. Although the resultant VNS reaction product of 2.39, 2,4-dinitro-5-chlorophenol, was reported to be obtained in high yield the substituents are orientated in such a manner that it could not be applied in our strategy.

2.4.3 METHODOLOGY FOR THE CONVERSION OF NITROPHENOLS TO A LAVENDAMYCIN AB RING SYNTHON

On the basis of the above research and analysis the starting materials which could still be considered for our proposed protocol for the synthesis of a lavendamycin AB ring synthon are 2,6- dinitrophenol, an appropriate dinitro derivative and guiacol 2.42. In this regard 2,6-dinitrophenol was submitted to this protocol. Its 0-acetyl derivative 2.45a was readily obtained in an isolated yield of 92 % by treatment of the substrate 2,6-dinitrophenol at room temperature with acetic anhydride in the presence of pyridine as base. The Ili NMR spectrum of the symmetrical compound 2.45a displayed the characteristic signals as a doublet, triplet and singlet at 8 8.29, 7.57 and 2.14, respectively. The use of tin dichloride dihydrate 26 in absolute ethanol has been reported as a mild and selective reducing agent tolerant to a variety of functional groups. Good results were obtained with its application in the case of streptonigrin (see Scheme 2.16). However, treatment of the acetyl derivative 2.45a with SnC12.2H20 in absolute ethanol at 70 °C followed by acetylation resulted in the isolation of a penta-acetylated product 2.46 which could be easily hydrolysed in a near quantitative yield to the diacetylated derivative 2.47, thereby confirming the reduction of both nitro groups. It is of interest to note that the unexpected N,N-diacetylation that occurred under the mild reaction conditions has also been previously reported in similar cases. I8 When subjecting the readily prepared O-pivaloyl derivative 2.45b to the same reductive conditions the desired product 2.48 was isolated only in a poor yield of 22.4 %. A further complication appeared with the isolation of the side product 2.49. Its formation was ascribed to the intermediacy of a hydroxylamine in the reduction process. Further reduction of this hydroxylamine may be so slow that that it becomes susceptible to competitive side reactions such as the acid catalysed Bamberger 63 rearrangement. 27 In terms of our objective, the best results were obtained by reduction of the unprotected 2,6-dinitrophenol with slightly less than the stoichiometric quantity for the reduction of one nitro group. The reduction was immediately followed by acetylation to furnish the tri- acetylated compound 2.50 as the major product in a yield of 62 %. A small amount of the product 2.47, corresponding to the reduction of both the nitro groups, was also isolated. However, control of the reaction on a scale required for synthesis proved to be extremely difficult and this route was, therefore, abandoned.

OCOCH3 OH (CH3C0)2 N N(COCH3)2 RHN NHR

(i) 5 SnC12.2H2,0 N(Et)3, Me0H abs. Et0H, 711°C OH Ac20 or OCOR (ii) Ac20, N(Et) 2.46 (6.5 %) R = COCH3 pivaloly NO2 CH2Cl2 NO2 NO 2 c hloride NO2 2.47(98 %) pyridine ► 5 SnC12.2H20 CH2Cl2, RT abs. EtOH, 70°C 2.45a R = CH3 (92 %) 2.45b R = C(CH3)3 ( 93 %) (i) 5 SnC12.2H20 OH OH abs. EtOH, 70°C 02N NHCOR 02N NHCOR (ii) Ac20, N(Et)3 1 CH2Cl2 HO OCOCH3 R = C(CH3)3 R = C(CH3)3 02N N(COCH3)2 2.48 (22 .4 %) 2.49 (10.2 %) 2.47(9.9 %)

2.50 (62 %)

SCHEME 2.24: ATTEMPTED REDUCTIONS OF 2,6-DINITROPHENOL AND DERIVATIVES

Another possibility employing 2,6-dinitrophenol would comprise triflation, a Heck reaction, reduction and cyclisation to a 2-hydroxyquinoline derivative. The introduction of the other substituents could follow later in the synthesis by utilising the mew directing properties of the remaining nitro substituent. However, presumably due to steric crowding, the Heck reaction of the triflate readily prepared from 2,6-dinitrophenol did not proceed at all, not even in the presence of 1 equivalent of the palladium(0) catalyst indicating that insertion of palladium into the C-OTf bond 64 does not take place. These disappointing results prompted us to investigate one of the other selected substrates as a suitable precursor to 2-hydroxyquinolines.

The dinitration of guiacol 2.42 was attempted using 2 equivalents of nitronium tetrafluoroborate in 1,2-dimethoxy ethane (DME) at a temperature of -50 °C. The colourless reaction solution quickly turned into a redbrown colour while thin layer chromatography (TLC) indicated complete conversion of starting material after 10 minutes and the formation of three compounds, two of them close to each other on the chromatogram. These products were isolated by column chromatography and identified as the desired dinitro compound 2.51 (60.2 %) and a mixture of the two mono nitro compounds (16.1 %), respectively. The simplicity of the 'H NMR spectrum of 2.51 in the aromatic region, showing only 2 doublets meta coupled to one another, indicated that substitution had indeed occurred ortho and para to the hydroxy group. The mass spectrum exhibiting a M ± of m/z 214 was also consistent with the structure of 2.51. The nitration of guiacol 2.42 with a 1:1 HNO3/AcOH mixture in ether and at a temperature of -10 °C also proceeded smoothly with the results similar to those of the previous experiment. However, when the latter reaction was repeated on a bigger scale, formation of the two mono nitro compounds were favoured to that of the dinitro compound

2.51 (yield: 12.4 % vs 67.6 % for mono nitration). It appears that the ether is such a good solvent that it extracts the mono nitro products from the biphasic system before a second nitration can take place. Therefore, by replacing the ether with a solvent such as dichloromethane the required product 2.51 could be obtained in a reasonable yield of 68 %. This yield is quite acceptable since both the starting material, guiacol 2.42 and the reagents, nitric and acetic acid are relatively cheap. The tendency of 2.51 to decompose on chromatography could be ascribed to acid catalysed denitration,28 but fortunately this problem could be solved by deactivating the column with 5 % Et3N and washing the product from the column with methanol.

Triflation of the substrate 2.51 was achieved in a yield of 70 - 75 % utilising Tf20 with either Et3N or 2,4,6-collidine as base in dichloromethane at 0°C. The 1H-NMR spectrum of the product 2.52 was in accord with the assigned structure with the aromatic protons resonating as a set of meta- coupled doublets at 8 8.51 and 8.14, respectively . The structure was confirmed by DC NMR and MS. Since it has been reported that dinitrotriflates 29 are unstable under basic conditions, it was not

65 unexpected that subjection of the triflate 2.52 to high temperature basic Heck conditions resulted in its rapid decomposition to the corresponding phenol and other products. Therefore, it was decided to perform the Heck reaction under neutral conditions while using epichlorohydrin as the proton scavenger. This modification of the Heck reaction which increased the applicability of the reaction significantly has been developed in the RAU laboratories. 17b Thus, treating triflate 2.52 with methyl acrylate, 2 equivalents of epichlorohydrin and 15 mole % of the catalyst Pd(PPh3)4, furnished the desired cinnamate 2.53 in a yield of 78 %. The 'H NMR spectrum of the compound 2.53 showed the characteristic alkene doublets resonating at 8 6.66 and 8 7.65, respectively with coupling constants of 16.1 Hz, indicating a trans orientation, while the MS spectrum displayed the required molecular ion at miz 282.

OMe OMe OMe OTf OH 1:1 HNO3/AcOH OH Tf20, base, 0°C), CH2C12, 0 °C, 1 h 02N CH2Cl2, 0° C, 3 h NO2 02N NO2 2.42 2.51 2.52 68 % 75 % ...... %,.. CO 2Me Pd(PPh3)4 epichloro- hydrin

OMe CO2Me

02N NO2

2.53 78 %

SCHEME 2.25: NITRATION, TRIFLATION AND HECK REACTIONS OF GUIACOL

However, the next step in the synthetic sequence, namely the reduction of the cinnamate 2.53 with SnC12.2H20 in absolute ethanol presented some difficulties. TLC of the reaction mixture indicated 66 a plethora of reaction products from which almost none of the desired 2-hydroxyquinoline 2.44 could be isolated. This observation was in sharp contrast to that obtained in the synthesis of streptonigrin, 17b where reduction of the corresponding cinnamate provided the cyclised product 2.32 in one step and not required additional acid to effect isomerisation and cyclisation (see Scheme 2.16). In spite of extensive variation in reduction conditions such as concentration, time, the amount of reducing agent, inverse addition and slow addition compound 2.44 could not be obtained in a satisfactory yield. Detailed analysis of the reaction products suggested that the Bamberger rearrangement of an intermediate hydroxylamine was at least partly responsible for the problem.

Therefore, it was decided to focus on the reduction of the model compounds 2.55a and 2.55b, easily prepared from 2-nitrobromobenzene 2.54a and 2,4-dinitrophenol 2.54b, respectively (see Scheme 2.26). The corresponding NMR spectra of the cinnamates 2.55a and 2.55b exhibited the additional sets of doublets for the alkene moiety and the required singlet for the methoxy signal of the methyl ester. The choice regarding the use of other reducing agents 30' e.g. samarium diiodide is rather limited since the substrate contains a double bond which is easily reduced. This tendency to reduce the double bond was confirmed by employing hydrazine and graphite mb as the reducing agent in the reduction of 2.55a which resulted in the isolation of 2.56a as the major product. The 1H NMR of this compound 2.56a displayed two sets of triplets at 5 3.22 and 8 3.72 in the place of the alkene doublets of the cinnamate, thereby confirming the reduction of the double bond. Small amounts of the amine derivative 2.56b with the double bond still intact and 2.56c resulting from reduction of both the double bond and the nitro group followed by cyclisation, were also isolated from the reaction mixture. The structures of 2.56b and 2.56c were confirmed by NMR and MS spectra. This, once again, demonstrates the difficulty in selecting a suitable reducing agent for aromatic nitro groups, a problem highlighted by the vast array of publications still appearing in this field."'" f

Our next option revolved around a one pot procedure which comprised treating the model compound 2.55a with tin dichloride dihydrate in ethanol until TLC indicated complete consumption of the starting material. The solvent was removed under reduced pressure (thereby cutting out the tedious work-up procedure of basifying the acid reaction mixture) followed by the addition of 3M 67

HCl and boiling of the resultant reaction mixture for 24 hours. This procedure afforded the required 2-hydroxyquinoline 2.57a after a basic work up procedure and chromatography in 60 % overall yield. Although not optimised, this yield compares favourably with the 65 % obtained with an alternative protocol involving two separate work-up and chromatography procedures. The 11-1 NMR of 2.57a showed the absence of a methyl ester signal and the hydroxy proton resonating at 8 12.41 ppm while the original alkene doublets became part of the quinoline ring giving rise to characteristic doublets at 8 6.71 and 7.81 (J3,4 = J4,3 = 9.5 Hz), respectively.

This so-called 'one pot' procedure was also successfully applied to the substrate 2.55b containing two nitro groups and furnished the 2-hydroxyquinoline 2.57b quantitatively with both the NMR and MS spectra consistent with the expected structure.

CO2Me

N2H4 R R NH 0 graphite 2.56a R 1 = H, R2 = NO2 2.56c R1 = H 2.56b RI = H, R2 = NH2

;Me CO2Me Pd(PPh3)4 epichloro- R NO2 hydrin SnC12.2H20 EtOH, 70 0C 2.54a R 1 = H, X =Br 2.55a R1 = H (98 %) 3M MCI, reflux 2.54b RI = NO2, X = OTf 2.55b R1 = NO2 (84.8 %)

R N OH 2.57a R1 = H (60 %) 2.57b R1 = NH2 (100 %)

SCHEME 2.26: ONE POT PROCEDURE TO 2-HYDROXYQUINOLINES

It was anticipated that application of the above protocol to substrate 2.53 would provide the proposed AB synthon of lavendamycin. However, when subjecting the cinnamate 2.53 to the one pot protocol described above, the desired 2-hydroxyquinoline 2.44 was furnished in the

68 disappointingly low yield of 45 %. This compound 2.44, representing our first success with respect to the AB fragment of lavendamycin, displayed in its 'H NMR spectrum the hydroxyl and amine protons as characteristic broad singlets at 8 11.21 and 8 8.52 ppm, respectively. The MS analysis indicated a molecular ion base peak at m/z 190 as well as a strong fragment corresponding to CO loss (M+-28).

The difficulty with reduction could be ascribed to the inherent tendency of the intermediate hydroxylamines to undergo acid catalysed Bamberger type rearrangements under the reaction conditions giving rise to, amongst others, the byproducts 2.58a and 2.58b.

OMe OMe OMe COOMe (I)Sn012.2H20, Et0H 70 °C v (ii) 3M HCI, reflux C)21\1 NO2 H2N v N OH H2N

2.53 2.44 R = H ( 45 %) 2.58a R = OEt 2.58b

SCHEME 2.27: FIRST PREPARATION OF AN AB SYNTHON OF LAVENDAMYCIN FROM GUIACOL

The structure of the unusual product 2.58b was indicated by its 'H NMR spectrum exhibiting the following signals i.e. three singlets in the aromatic region, a methoxy singlet and a doublet and triplet corresponding to the ethoxy group, respectively. Its formation could be rationalised in terms of an unusual variation on the Bamberger rearrangement (Scheme 2.28). Since methanol is generated during the reaction it also acts as a competing nucleophile resulting in the isolation of a mixture of corresponding methoxy and ethoxy derivatives.

69

OMe OMe

- H2O 0 H28—HN NH 0 H2O—HN OH OH

OMe OMe OMe OR H OR ROH H2N OH OH

SCHEME 2.28: INTERESTING VARIATION ON THE BAMBERGER REARRANGEMENT

This problems regarding the Bamberger rearrangement in the case of substrate 2.53 did not occur in the case of the streptonigrin precursor 2.31b (see Scheme 2.16). Here, the Bamberger rearrangement of a hydroxylamine reduction intermediate cannot occur since both positions para to the nitro group are blocked to nucleophilic attack by substituents. In order to establish whether substrates para to the nitro group ensured successful reduction and cyclisation reactions, in-cresol 2.59 and 3-methoxyphenol 2.60a were selected as starting materials for our 2-hydroxyquinoline synthesis. However, the dinitration of m-cresol with NO2BF4 provided an interesting, but unexpected result. On the basis of the meta directing effect of the first nitro group introduced in conjunction with steric constraints, 4,6-dinitro-m-cresol was the expected product. However, the only dinitro compound isolated was the 2,4-dinitro-m-cresol 2.61 showing a preferential formation of a product with four contiguous centers. This is contrary to the expectation based on steric and directing effects, but as has been previously reported with the dinitration of 1,4-dimethoxybenzene, it can be explained if the reaction is taking place via a radical mechanism. 31 In the case of 2.60a the whole sequence of proposed reactions proceeded smoothly as illustrated in Scheme 2.29. The final reduction and cyclisation step furnished the substituted 2-hydroxyquinoline 2.62 in a quantitative yield. 70

(01120, collidine CH2Cl2 0 °C OH 2.2 NO2BF4 I R OR2 Me0 COOMe 1,2-DME, -50 ° C Pd(PPh3)4 02N NO2 epichbrohydrin 02N NO2 2.59 R1 = CH3 2.60a R 1 = OCH3 2.60b R 1 = OCH3, R2 = H (66 %) 2.60d 85 %) 2.2 NO2SE4 2.60c R1 = OCH3, R2 = Tf (85 %) (i)SnC12.2H20, EtOH 1,2-DME, -50 ° C 70 °C (11)3M HCI, reflux NO2 1 OH Me0

NO2 H2N N OH 2.61 R1 = CH3 (35 %) 2.62 (quantitative)

SCHEME 2.29: ANOTHER EXAMPLE OF THE SYNTHESIS OF 2-HYDROXYQUINOLINES

The protocol developed for the synthesis of 2-hydroxyquinolines from the O-triflate derivatives of nitrophenols thus appears to have considerable synthetic potential. In addition, a suitably substituted lavendamycin AB synthon has been prepared from commercially available guiacol 2.42 in 5 steps in an overall yield of 17.9 %. Although unoptimised, the yield was considered too low in the context of a multistep synthesis.

It was considered unlikely that the detrimental effect of the Bamberger rearrangement could be overcome. However, it was realised that it might be to our advantage in a new synthetic strategy. The Bamberger rearrangement is the most convenient and economical method for the synthesis of para-aminophenols directly from nitrobenzenes, 32 particularly on an industrial scale. This process has been described in several patents 33 and is normally carried out by catalytic hydrogenation in the presence of precious metal catalysts under highly acidic conditions. In the case of nitrobenzene the intermediate has been shown to be N-phenylhydroxylamine. A modified version of the Bamberger reaction reported by Patrick and coworkers 34 included a reasonable synthesis for fluoroaromatic amines by treating the appropriate N-phenylhydroxylamine with anhydrous hydrogen fluoride. 71

In our case it was envisaged that selective reduction of the nitro compound 2.63 to the corresponding hydroxylamine 2.64 followed by a Bamberger rearrangement and subsequent cyclisation under the highly acidic conditions would afford quinoline 2.65, thereby providing an alternative approach to the streptonigrin skeleton starting from the commercially available guiacol 2.42 (see Scheme 2.30). The introduction of the required amino group in the C-7 position later in the synthesis would follow the same procedure as previously described in the literature. 6

OMe OMe OMe OH COOMe reduction COOMe

NHOH 2.42 2.63 2.64 Bamberger rearrangement I ROH OMe OMe OMe RO RO RO a COOMe cyclisation 4 H2N OH NH2 2.65

SCHEME 2.30: A POSSIBLE APPROACH TO THE STREPTONIGRIN AB RING SYSTEM FROM GUIACOL VIA A BAMBERGER REARRANGEMENT

The general application of such an approach is highly dependent on the availability of the required N-hydroxylamine while the synthesis of the latter is determined by the ease of selectively reducing the required nitro compound to the corresponding N-hydroxylamine. Selective reduction to N- hydroxylamines can be achieved in a variety of ways, 35 but the most widely applicable systems utilise zinc and ammonium chloride in an aqueous and/or alcoholic medium. 72

Some initial model reduction reactions were undertaken to test the viability of such a proposal. Good results were obtained by the selective reduction of nitrobenzene with zinc dust-ammonium chloride in aqueous medium to the corresponding N-hydroxylamine 2.66 which in the presence of sulphuric acid rearranged to product 2.67a (see Scheme 2.31). If the rearrangement was performed in methanol the corresponding para-methoxyphenol 2.67b was isolated. Similar results with respect to reduction were achieved when employing the rapid catalytic transfer reduction methodology using a rhodium-charcoal catalyst and hydrazine as hydrogen donor, as reported by Entwistle et. a1. 35a

Therefore, the reduction of the model compound, the nitrocinnamate 2.55a, in the hope of demonstrating a consecutive partial reduction, Bamberger rearrangement and cyclisation, was investigated. However, its reduction with the zinc dust-ammonium chloride system resulted in a mixture of products with overreduction to the unwanted amine 2.56b occurring frequently. The 1 1-1 NMR spectrum of 2.56b displayed, apart from the two sets of alkene proton doublets, a methoxy singlet and two sets of triplets and doublets of the aromatic protons and the characteristic broad amine two proton singlet at S 3.97. The mass spectrum exhibited the required molecular ion at m/z 177 and the expected fragments at m/z 146 (M +-OMe) and 118(M+- OMe - CO), respectively. The suggested35a use of a two-phase ether-water system with careful control of both the temperature and the pH (between 6.5 - 7) was expected to alleviate the overreduction to the amine 2.56b, but it was still obtained in a 24 % yield. The major product, isolated in a yield of 45 %, was identified as the desired hydroxylamine 2.68 after extensive spectroscopic studies. The MS of the compound showed the expected M + peak at m/z 193 and its composition was confirmed by high resolution mass spectrometry (HR-MS). Although the i ll NMR spectrum of 2.68 showed the expected two sets of alkene proton doublets ( J = 15.9 Hz), a methoxy singlet and two broad singlets at 8 6.14 and 8 6.99 (which disappeared on the addition of D20), the aromatic protons was, surprisingly enough, characterised by two second order multiplets in a 3:1 relation. A third reaction product, azoxyarene 2.69 was isolated in minute quantities and most probably originated from the known reaction between an N-hydroxylamine and a nitroarene. 35b This dimeric 2.69 compound was fully characterised by Ili and BC NMR and MS data. 73

The problem regarding overreduction could not be solved utilising other alternative and mild reducing agents such as palladium with phosphinic acid or its sodium salt as the hydrogen donors, respectively. 35a The same difficulties were also experienced with other model compounds such as o-nitrotoluene and 2,6-dinitrotoluene (Scheme 2.31). Another difficulty arose when treatment of the N-hydroxylamine 2.68 with acid in order to facilitate the Bamberger rearrangement furnished at least 4 different reaction products. Thus, it is clear that both the preparation of the extremely labile hydroxylamines as well as the Bamberger rearrangement of such compounds are not straightforward and that the methodology to control these reactions, is still lacking. It led us to the conclusion that this approach to a streptonigrin AB synthon is (for the time being) not synthetically viable.

OR

NH2 NH2 RI = CH 3, R2 = NO R2RI = R2 = H (i) (iii)or (i) or (iv) CH3 CH3 R NHOH NO2 NHOH NO2 NH2 2.66 2.67a R = H (34 %) (15%) 2.67b R = Me

R1 = CH=CHCO2CH3 , R2 = H 2.55a (i) or (iii) or (iv) CO2Me O CO2Me CO2Me N=N

NH2 Me02C 2.68 (45%) 2.56b (24 %) 2.69 (<5 %)

Reagents: (i) 2Zn, NH 4CI H2SO4 Rhodium on charcoal/hydrazine Pd/phosphinite

SCHEME 2.31: ATTEMPTED PREPARATIONS OF VARIOUS HYDROXYLAMINES 74

2.4.4 A SYNTHETIC APPROACH STARTING FROM A PREFORMED RING SYSTEM

The focus of the synthesis for the AB ring system of lavendamycin now changed from the construction of a substituted 2-hydroxyquinoline via a Heck reaction of ortho-nitrotriflates to the utilisation of a preformed ring system such as the commercially available 8-hydroxyquinoline. It was envisaged that the introduction of a suitable leaving group in the 2-position would involve the following sequence of reactions: protection of the 8-hydroxyl group, N-oxidation and a rearrangement step to introduce the required functionality in the 2 position (Scheme 2.32). The protecting group should be easily removable as the directing properties of the hydroxy group would play a vital role later on in the synthesis to introduce the required amino functionality via nitration and reduction.

The first protecting group to be considered was an acetyl group as illustrated by the ease of acetylation of 8-hydroxyquinoline to furnish 2.70 in a yield of 85 %. During the following N- oxidation reaction of the acetate 2.70 with m-CPBA, m-chlorobenzoic acid is formed which is removed from the reaction mixture by reaction with solid potassium carbonate present in the reaction mixture. Unfortunately, this caused hydrolysis of the acetate moiety resulting in the isolation of a 1:1 mixture of the acetate 2.70 and the N-oxide 2.71. This problem was easily solved by interchanging the 2 reactions, i.e. the first step is the N-oxidation of 8-hydroxyquinoline to the N-oxide 2.71 followed by acetylation to the diacetate 2.72. The characteristic low field signal of the phenolic proton in the N-oxide 2.71, strongly hydrogen bonded to the oxygen of the N-oxide, resonated at 5 15.01 ppm compared to the phenolic proton of 8-hydroxyquinoline resonating at 5 9.90 ppm. The subsequent diacetylation of 2.71 to 2.72 was slow and required 3 days under reflux to go to completion. This can probably be ascribed to the previously mentioned strong hydrogen bonding rendering the phenolic proton less acidic. It should be mentioned that it is reported in the literature that the acetylation of 2.71 was completed within 5 hours and that a basic workup furnished the 8-acetoxycarbostyril 2.74.36 The Ili NMR spectrum of the diacetate 2.72 showed the expected five aromatic protons compared to six found in 2.71 and the presence of the expected two acetyl groups resonating at 5 2.45 and 5 2.36, respectively. The structure was further consistent with 13C NMR and MS data, the latter exhibiting the required M + at tn/z 245 and two consecutive

75 losses of 42 to tn/z 203 (M+ - CH2CO) and tn/z 161 (M+ - 2xCH2CO), respectively. Hydrolysis of both of the acetate groups of the diacetate 2.72 was effected by treatment with triethylamine in methanol. This furnished the polar 2,8-dihydroxyquinoline 2.73 in quantitative yield, thereby giving access to the desired quinoline derivative substituted in the 2 and 8 positions in 3 steps. Once again, the 'H NMR, the I3C NMR and MS data were in complete agreement with the assigned structure of 2.73. If required, the selective alkylation of the 8-hydroxyl group could be carried out as described in the literature 37 which could then be followed by either a triflation step or a halogen exchange of the 2-hydroxyl group, to eventually allow cross-coupling to a suitable CDE moiety.

m-CPBA, CH2Cl2' Ac20, reflux

OH OH 0 - OAc 2.71 (75 %) 2.72 (78 %) Ac20,NEt3 (lit.) CH2Cl2 NEt3' MeO 1

OH OAc OH 2.70 (85 %) 2.73 (quantitative) 2.74 1 m-CPBA, CH2Cl2

+ 2.70 (1:1)

OH 0 2.71

SCHEME 2.32: THE POSSIBLE USE OF ACETYL AS PROTECTING GROUP

The investigation was then extended to other protecting groups which would still meet the requirements of being easily deprotected while able to withstand the reaction conditions later on in

76 the synthesis. Thus, following the standard procedures employing potassium carbonate as base, alkylation with either allyl bromide, benzyl bromide or methyl iodide in refluxing acetone, 8- hydroxyquinoline furnished the corresponding allyl, benzyl and methyl derivatives 2.75 - 2.77, respectively (Scheme 2.34). The corresponding yields for the three compounds were, 85 %, 58 % and 71 %, respectively. In the benzylation reaction a higher overall yield could be obtained by recycling unreacted 8-hydroxyquinoline. In addition to the six protons of the quinoline ring, the presence of the three different protecting groups followed from their respective II-I NMR, 13 C NMR and MS spectra.

+

2.75 2.78 (M+ 226) 2.79 de, Claisen rearrangement

OH (11 2.80 (M+ 226)

SCHEME 2.33: SUGGESTED REACTIONS OF 2.75 IN THE MASS SPECTROMETER

It is of interest to note that the allyl species 2.75 did not show the required molecular ion at m/z 185 in its MS, but exhibited strong ions at m/z 226, m/z 208 and m/z 184 with high intensities in the spectrum. This is tentatively rationalised (see Scheme 2.33) in terms of an intermolecular reaction under the high temperature conditions in the MS to furnish the di-allylated species 2.78 which should readily undergo the Claisen rearrangement to 2.80. The suggestion is not unreasonable since 77 nitrogen is a good nucleophile and 2.79 is a good leaving group. The fragment ions at tniz 208 and m/z 184 suggest the loss of elemental water and propylene from the molecular ion of 2.80, respectively.

The following step in the synthetic sequence, namely the N-oxidation with m-CPBA 38 to the corresponding allyl, benzyl and methyl N-oxides 2.81 - 2.83 proceeded uneventfully (see Scheme 2.34). Usually the work up procedure involves decomposition of the remaining m-CPBA by treatment with solid sodium bisulfite, a filtering step, removal of the generated m-chlorobenzoic acid by stirring the solution in the presence of solid potassium carbonate and again, filtration. However, the yields of the various reactions could be improved significantly by replacing this tedious work up process with a rapid chromatography step. Thus, the reaction mixture is immediately filtered through a thick, short silica gel column packed in 100 % ethyl acetate. The polar N-oxides are then washed from the column with a 3:1 ethyl acetate-methanol eluent. This procedure yields excellent results, except in the case of the benzyl derivative where the yield remained low at 52 %. The aromatic proton resonances in the 'H NMR spectrum of 2.81- 2.83 had identical splitting patterns to those of 2.75 - 2.77, and changes in the chemical shifts of these protons were similar to those observed between pyridine and pyridine N-oxide as reported by Pretsch. 39 In all three cases the M + peaks corresponding to the required molecular masses confirmed the presence of an additional oxygen atom.

The N-oxidation was followed by a rearrangement step to introduce a suitable substituent ready for cross-coupling into the 2 position of the quinoline ring. However, attempted rearrangement by treatment with acetic anhydride under reflux resulted in complex, difficult to separate reaction mixtures. The main problem was that rearrangement to the 4-acetoxyquinolines competed with the desired rearrangement to the 2-acetoxyquinolines. Similar rearrangements have been reported in the literature 4°a

It was therefore decided to focus on the utilisation of phosphorous oxychloride and the expensive phosphorous oxybromide as reagents to facilitate the required transformation 4 0b These reagents coordinate at the N-oxide oxygen and activates the ring towards nucleophilic attack at C-2 or C-4

78 by chloride or bromide ions. Elimination of HOPOCl2 or HOPOBr2 should then afford the deoxygenated, halogenated products.

K2CO3, acetone reflux m-CPBA, CH 2Cl2 Mel or allyI bromide or OH benzyl bromide OR 2.75 R = ally' (85 %) 2.81 R = ally! (78 %) 2.76 R = Bn (58 %) 2.82 R = Bn (52 %) 2.77 R= Me (71 °hi) 2.83 R = Me (74%)

POCI3 or POBr3 or Tf2O/NI-14 BuX with X= Br or I

OR

2.84 R = allyl, X = CI (66 %) 2.85 R = Bn, X = CI (58 %) 2.86 R= Me, X= Cl (77 %) 2.87 R = allyl, X = Br (42 %; 2.88 R= Me, X= Br (47 %)

SCHEME 2.34: ALTERNATIVE PREPARATION OF LAVENDAMYCIN AB SYNTHON

Thus, treatment of N-oxides 2.81 -2.83 with phosphorous oxychloride at room temperature afforded the expected 2-chloroquinolines 2.84 -2.86 within a few hours. Although TLC indicated total consumption of the starting material, the yields were only moderate and it was suspected that loss of material occurred in the work up procedure comprising neutralisation of the acid reaction mixture followed by extraction. A slight improvement was observed when the reaction mixture was distilled under reduced pressure to remove the excess phosphorous oxychloride. 79

Initially the reaction was performed in chloroform as solvent, but since the chloroform contains small amounts of ethanol, this resulted in contamination of the product with triethylphosphate. Fortunately, this could be circumvented by using 1,2-dichloroethane as solvent. The I I-I NMR spectra of the desired 2-chloroquinolines confirmed the presence of the appropriate protecting groups and five aromatic protons resonating as four doublets and one triplet. The EI-MS spectra confirmed the presence of chloride, as the M ± base peaks exhibited the characteristic C1 35/C137 isotope pattern.

In the case of the conversion of the N-oxides 2.81 and 2.83 to the corresponding bromine containing derivatives 2.87 and 2.88 the yields (Scheme 2.34) were considerably lower (< 50 %) even on treating the N-oxide with an excess of the expensive phosphorous oxybromide at 70 °C. This is, at least, in part, due to competitive nucleophilic attack at C-4 since about 20 % of the 4-bromo

derivative were also isolated. Just as in the case of the corresponding chloro derivatives 2.84 - 2.86, the presence of the bromide could be deduced from the characteristic 1:1 Br 79/Br8I isotope peaks in the MS spectra. The II-I and 13C NMR spectra of these two compounds were in full agreement with the proposed structures.

Since the yields of these bromo derivatives 2.87 and 2.88 were disappointingly low, another route to these compounds was proposed. It was envisaged that treatment of the N-oxide with triflic anhydride in the presence of a quaternary ammonium bromide salt would result in triflation of the N-oxide and since the triflic anion is essentially non-nucleophilic, attack by the bromide ion would take place with concomitant loss of triflic acid to afford the required bromo compound 2.87. This reaction did indeed occur at room temperature, but once again, the yield of the bromo compound 2.88 was low (2-- 30 %). Although the competitive reaction to the 4-bromo derivative was eliminated with this protocol, another byproduct, i.e. dimer 2.89 appeared. Formation of this dimer 2.89 is believed to occur as a result of attack by unreacted N-oxide 2.81 on the N- 0c triflyloxyquinolinium salt 2.90 with subsequent rearrangement (Scheme 3.35).4 80

N H OTf 6 H

2.90 2.81

2.89

SCHEME 2.35: FORMATION OF DIMER 2.89

This second approach to the preparation of an AB rin g synthon of lavendamycin meets the required criteria as previousl y set with the exception that the desired amino substituent in C-7 is not present. However, since the protectin g groups were chosen to be easil y removed, the directin g properties of the hydroxy substituent would be used to introduce the re quired amino functionality in the C-7 position following a sequence of nitration, protection and reduction. This was confirmed b y the dinitration of 8-hydroxyquinoline41 to 2.91 which proceeded smoothl y and with the expected structure confirmed b y its NMR spectra. The reduction of the nitro groups and the final oxidation to the quinone system would follow the same methodolo gy employing H2/Pd on carbon as previously described by Behforouz.9•1°

81

HNO 3/H2SO4 0°C 02N OH OH 2.91 (88 %)

SCHEME 2.36: DINITRATION OF 8-HYDROXYQUINOLINE

Thus, two general strategies towards the synthesis of the AB ring system have been successfully developed. The five different key intermediates i.e. 2-hydroxyquinoline 2.44, the two bromo compounds 2.87 and 2.88 ( in an unoptimised yield) and the three chloro derivatives 2.84 -2.86, all possess suitable functionality in the 2 position which would allow further cross-coupling to an appropriate CDE ring equivalent. Therefore, the following part of this chapter will revolve around the existing as well as new approaches to the preparation of a suitable CDE ring system.

2.5 APPROACHES TO THE CDE MOIETY OF LAVENDAMYCIN

From the structure of lavendamycin 2.2 it follows that its CDE ring synthon possesses a substituted P-carboline as the basic structural unit. This important ring system has received much attention since B-carbolines are widely distributed in nature with new natural ones continuously being discovered, and the biological activity of a wide range of natural and synthetic 13-carbolines likewise being explored. 19° A great many elegant syntheses to complex structures containing the p- carboline structural unit (including lavendamycin) have been reported. However, on evaluation of the available literature methods to 13-carbolines, it had to be taken into account that the proposed strategy for lavendamycin involves a cross-coupling reaction between the respective AB and CDE moieties. To allow such a cross-coupling, functionality was required at C-1 of the pyridine C ring of the p-carboline (see Scheme 2.37 for an example of the numbering of the (3-carboline structural unit). If a halogen could be introduced at the required position, it would allow access to the desired metallated product via halogen-lithium exchange, or enable direct cross-coupling to an appropriate 82 organometallic derivative of the AB quinoline ring system. Therefore, a selection of those methodologies to ft-carbolines viable to the desired halogenation, will be discussed.

The traditional synthetic methods such as the Pictet-Spengler ca'b' and Bischler-Napieralski 42d'e reactions are by far the most commonly employed methods to synthesise ft-carbolines. These cyclisation methods usually start with a preformed indole ring with annulation of the pyridine ring taking place during the reaction. As the Pictet-Spengler reaction initially yields the tetrahydro-ft- carbolines, these compounds are usually converted to the fully aromatic systems with a variety of oxidants. The synthesis of ft-carboline 2.92 was accomplished via a base-catalysed cyclisation of tryptophan followed by esterification and oxidation (Scheme 2.37). 42c A similar reaction has been successfully performed under neutral conditions while another variation involves the utilisation of glyoxylic acid in stead of formaldehyde. 19a However, the Pictet-Spengler condensation reaction and the subsequent aromatisation are often low yielding steps.

COOH 30 % formalin, NaOH NH2 Me0H, HCI N N H H 87 % (over 2 steps) 1 Pb(0Ac)4, HOAc COOMe 3

7 2 9 N 1 H 2.92 (57 %)

SCHEME 2.37: THE PICTET-SPENGLER REACTION TO A ft-CARBOLINE 83

The Bischler-Napieralski reaction is frequently applied in the synthesis of P-carbolines, but usually in approaches to compounds substituted at C-1 of the p-carboline nucleus. This is also the case with Molina's43 tandem aza-Wittig/electrocyclic ring closure process and all other reported strategies to lavendamycin, except for the approach by the French group 15' 16 (Scheme 2.6). Although other alternative methods to P-carbolines via formation of the pyridine ring exist, it often involves novel cyclisations and/or the use of starting materials prepared by difficult multistep procedures.

Dodd et.al.44 required multigram quantities of 3-carboxy-p-carbolines which was successfully prepared via intramolecular alkylation at C-3 of the indole ring (see Scheme 2.38).

EtO CHO R OR 1 H NH2 2.93 2.94a R = H, R 1 = Et 2.95a R = H, R 1 = Et (98 %) 2.94b R = R 1 = Me 2.95b R = R1 = Me (95 %)

NaBH3CN CO2R1 1 TiCI4 H OEt benzene,A OEt CO2R1 H H 2.97a R= H, R1 = Et (57 %) 2.96a R = H, R1 = Et (94 %) 2.97b R = R1 = Me (64 %) 2.96b R = R1 = Me (95 %)

SCHEME 2.38: DODD'S METHOD TO SUBSTITUTED P-CARBOLINES

Condensation of indole-2-carboxaldehyde 2.93 with amino ester 2.94 produced imines 2.95 in excellent yield as illustrated in Scheme 2.38. Attempts to cyclise 2.95 directly with concomitant oxidation (air) were unsuccessful, therefore the imine functionality was reduced to the amine 2.96 84 which was then cyclised to 2.97 by heating with titanium(IV)tetrachloride. Although this protocol furnishes 13-carbolines in a reasonable overall yield, it is misleading, since both the substrates 2.93 and 2.94 are not readily available. These compounds had to be synthesised via multistep procedures, thereby lowering the overall yield.

The synthesis of p-carbolines may also proceed with formation of the pyrrole ring as illustrated by

Queguiner's 15 ' 16 approach where cyclisation occurred via nucleophilic attack on pyridyl halide 2.98 to form 2.99 (Scheme 2.39). The required aryl pyridine 2.98 was prepared in high yield by a Suzuki coupling with cyclisation to 2.99 effected on heating under reflux in anhydrous pyridinium chloride. This method has been applied to the preparation of other C-1 substituted 6-carbolines such as lavendamycin (a retrosynthesis is shown in Scheme 2.6). However, the employment of the penta substituted pyridine 2.17 in this strategy has two drawbacks: the preparation of the pyridine 2.17 requires a five step synthesis and since being fully substituted, access to lavendamycin analogues is denied.

(i) BuLi B(OH)2 (H) (Me0)3B 0 (Hi) HCI

58%

t LDA IZ Py•HCI NH40 H

t1 H 2.99 (86 %)

SCHEME 2.39: QUeGUINER'S APPROACH TO f3-CARBOLINES

85

Two promising approaches to 13-carbolines possessing hetero-atoms at C-1 have been reported. In a modified version of Molina's 45 tandem aza Wittig/electrocyclic process the amide 2.100 was converted to the corresponding substitued P-carboline 2.101 on treatment with carbon dioxide (Scheme 2.40). 45 This compound is of importance since a "pyridone" (in tautomeric equilibrium with its corresponding 2-hydroxy derivative) could be readily converted into the desired 2- halogenated derivative required for cross-coupling to an AB ring precursor.

CONHCH3 COz s "PP h3

CH3 2.100 2.101

SCHEME 2.40

Bracher i9b managed to prepare 1-chloro-13-carboline in an efficient three step synthesis from tryptamine in an overall yield of 49 %. This approach entailed treatment of the readily available tryptamine with triphosgene to furnish 2.102 which was followed by oxidation with palladium on carbon and subsequently, halogenated to complete this practical synthesis (Scheme 2.41).

Triphosgene N—H N 0 2.102 (74 %)

1Pd/C

POCI3 or P0Br3 X X= CI (70 %) or Br (88 %)

SCHEME 2.41: BRACHER'S PREPARATION OF 1-CHLOR0-13-CARBOLINES 86

The final strategy to f3-carbolines that needs mentioning is Boger's I4 method employing a Pd(0)- catalysed substitutive cyclisation of amino bromide 2.103. This method was developed especially with lavendamycin in mind.

Me CO2Me Pe(PPh3)4

2.103

SCHEME 2.42: BOGER'S Pd(0)-CYCLISATION TO fl-CARBOLINES

From the discussion (vida supra) it follows that although the traditional methods still find considerable use in the preparation of fl-carbolines, newer methods, especially those involving organometallic species, are finding increased utility. Therefore, we based our strategy to p- carbolines on the application of a relatively new palladium catalysed N-heteroannulation reaction, 46 successfully developed for the preparation of indoles from 2-nitrostyrenes, by employing commercially available 4-phenyl pyridine N-oxide 2.104 as a suitable starting material. The introduction of oxygen functionality into the a position of the pyridine with acetic anhydride would involve initial attack of the anhydride by the N-oxide 2.104, addition of the actetate to C-2 to afford the 2-acetoxy compound 2.105a which should be readily hydrolysed to 2.105 b. It is known that, depending on reaction conditions, the nitration of pyridin-2-one could furnish either the 3- or 5- nitro substituted pyridone 47 compound as the predominant product. Therefore, it was expected that nitration of the 4-phenyl derivative 2.105b would behave similarly and that the 5-nitro product 2.106a would be obtained readily. It was envisaged that palladium catalysed reductive N- heteroannulation of nitro compound 2.106a would furnish 0-carboline 2.108a presumably via a nitrene. The next logical step would be to investigate a possible one-pot procedure to 13-carboline 2.108b via corresponding triflate 2.107a by the reductive cyclisation reaction in the presence of carbon monoxide and a suitable nucleophile such as benzylamine. Ideally this would give access to

87 the required 13- carboline 2.108b which would be suitably substitued to allow introduction of a wide range of substituents into C-4 of the (3-carboline via DoM-methodology.48

00 OAc OH NH 0

Ac20 Hydrolysis

2.104 2.105a 2.105b

Nitration

OR Reductive cyclisaton 02N Pd(0)/C0

2.108a R = OH 2.106b R = H 2.106a R = H 2.108b R = CONHCH2Ph 2.107a R = OTf 2.107b R = OTf

SCHEME 2.43: PROPOSED APPROACH TO (3-CARBOLINES

Thus, the treatment of N-oxide 2.104 with acetic anhydride proceeded smoothly to afford the acetate 2.105a in 53 % yield and was quantitatively hydrolysed to the corresponding 2-hydroxy derivative 2.105b. Its I li NMR spectrum, acquired in d6-DMSO, displayed three sets of one proton doublets and two multiplets (integrating for 2 and 3 protons, respectively) corresponding to the aromatic protons and a broad singlet resonating at 5 11.58 arising from the phenolic proton. However, subsequent reactions furnished 2.105a in variable yield, while nitration of 2.105b with HNO3-H2SO4, unlike in the case of 2-pyridone, gave a complex mixture of numerous products. These results indicate that the introduction of the phenyl ring at C-4, with its contributing resonance effects, significantly changes the electronic nature of the pyridone ring of 2.105b compared to the 88 simple model compound, 2-pyridone. Furthermore, reductive cyclisation of the model compound 2-nitrobiphenyl failed and produced only the reduced product, 2-aminobiphenyl. Various catalysts, temperatures, reaction times, solvents and carbon monoxide pressures were investigated to change the outcome of the reaction, but with no obvious success. Clearly, the nitrene intermediate in the Pd-catalysed reduction of the nitro group is not sufficiently reactive to attack the benzene ring.

Our next approach to the p-carbolines was based on an approach to streptonigrin developed in the RAU laboratories. 18 This involved the Suzuki cross-coupling of 4-chloro-3-nitropyridine (readily prepared from 4-hydroxypyridine via nitration and halogenation) to phenylboronic acid to yield the coupled product, 3-nitro-4-phenylpyridine in excellent yield. This was followed by N-oxidation and halogenation to the corresponding 2-chloro derivative which was finally coupled to 2- trimethylstannylquinoline to produce the streptonigrin skeleton 2.19 as previously shown in Scheme 2.9.

Application of this protocol to a suitable lavendamycin AB synthon would give access to the precursor 2.109 which on reductive cyclisation in the presence of triethylphosphite should furnish the pentacyclic lavendamycin skeleton 2.110 (Scheme 2.44). Coupling of the AB and CDE precursors would require the introduction of a halogen at C-2 of the right hand equivalent and, therefore, the preparation of a 2-trimethylstannyl derivative of the AB quinoline system from the corresponding bromides 2.87 or 2.88. A shorter option requiring fewer reactions would involve the use of 2-chloro-B-carboline prepared from tryptamine by the method of Bracher, 19b in stead of 3- nitro-4-phenylpyridine as coupling partner. Completion of the left hand side has already been discussed and shown to proceed by a sequence of nitration, reduction and oxidation while completion of the right hand side of the molecule would follow N-oxidation of the pyridine C-ring, halogenation , amino- or hydroxycarbonylation 49 and finally, introduction of a C-4 substituent via DoM-methodology. 48 N-oxidation of compound 2.110 would probably proceed selectively at the pyridine C ring because of steric crowding at the quinoline nitrogen and strong hydrogen bonding between the latter and the indolic proton. 89

Stille coupling I, Cl N Br SnMe3 OR OR OR 02N 0 N 2.87 R = allyl R = allyl or Me 2.88 R = Me (36 % over 6 steps from 4-hydroxypyridine) 2.109a R = allyl 2.109b R = Me

P(O Et)3

2.110a R = allyl R = allyl or Me 2.110b R = Me Ri = OH or NHCH2Ph

N-oxidation Base, E+ 1

COR 1 CI Pd(0), CO Halogenation OR -4— OR H2O or HN HN PhCH2NH2

R = allyl or Me R = allyl or Me R = allyl or Me R 1 = OH or NHCH2Ph

SCHEME 2.44: AN ALTERNATIVE PROPOSED ROUTE TO LAVENDAMYCIN

The palladium catalysed aryl halide carbonylation has become an important tool in synthetic organic chemistry in recent years. The effective conversion of simple dichloropyridines into their 90 corresponding diesterssoa and the existence of a facile preparation of pyridine-2-carboxamides via aminocarbonylation of 2-bromopyridine 5°bAd provided sufficient indication that we would be able to obtain the required C-3 substituted (3-carboline. If the carbonylation reaction is performed in the presence of water it would give access to the corresponding carboxylic acid which could then directly be lithiated by the recently reported method of Queguiner 5I and allow the introduction of an electrophile in the C-4 position of the 13-carboline system.

However, it was realised that this cross-coupling strategy would be very long, requiring four and six reactions, respectively to prepare the AB and CE synthons and after the creation of the B-C biaryl axis by the Stille reaction, it once again, requires three and five reactions for the completion of the left hand and right hand side, respectively. Although employing modern catalytic processes and proposing a convergent approach, this route was considered to be impractical. Therefore, a total rethink of our strategy regarding the formation of the B-C axis was required.

2.6 AN ALTERNATIVE APPROACH TO LAVENDAMYCIN VIA CARBONYLATION

We focused our new strategy around the availibility of 2-chloro and 2-bromo AB precursors 2.84 - 2.88. On closer scrutiny it was realised that if the above mentioned hydroxy carbonylation reaction of 2-chloro- or 2-bromopyridines would be applied to a suitably substituted 2-chloroquinoline derivative in water, it would lead to the corresponding quinaldic acid, which, if coupled to 13- methyltryptophan, would provide amide 2.111. The traditional Bischler-Napieralski cyclisation of the amide would then give access to the desired pentacyclic lavendamycin precursor 2.112 in a few steps. A further development was suggested by a report 52 on the synthesis of succinate containing dipeptide isosteres utilising a-amino acid derivatives as nucleophilic traps in the carbonylation of enol triflates. Therefore, in our new strategy it was envisaged that the preparation of the amide 2.111 via a one pot reaction would involve palladium catalysed carbonylation of an appropriate 2-

chloro- or 2-bromoquinoline derivative 2.84 - 2.88 with tryptophan as the nucleophile (Scheme 2.45). This will be followed by a Bischler-Napieralski cyclisation, hydrolysis to the corresponding acid and the introduction of the final substituent by employing DoM-methodology.

91

3 steps

CI (or Br) OH OR 2.84-2.88 Pd(0), CO NH2 CO21vb N H

CO2Me

R = Me or Bn or ally] 2.111a R= Me 2.111b R= Bn 2.111c R = ally!

COOH COOMe Hydrolysis

2.112a R= Me R = Me or Bn or ally! 2.112b R = Bn 2.111c R = ally'

SCHEME 2.45: A NEW APPROACH TO LAVENDAMYCIN VIA CARBONYLATION

The idea of carbonylation was first tested on the model compound, 2-chloroquinoline, with benzylamine as the nucleophile using the same reaction conditions as initially employed by Heck et. a1.53 Thus, 2-chloroquinoline, a slight excess of benzylamine, 25 mole % of PdC12(PPh3)2 and tri-n-butylamine was heated to 95 °C in the presence of carbon monoxide at atmospheric pressure. 92

After only one hour TLC indicated that the reaction was almost complete with the appearance of a new spot at lower Rf. After extractive work-up and chromatography, the product 2.113 was obtained (Scheme 2.46) as white crystals in a yield of 84 %. The 11-1 NMR of the compound displayed the expected signals including one broad NH singlet, two triplets and four doublets in the aromatic region (the quinoline ring protons), one multiplet integrating for five protons (phenyl ring protons) also in the aromatic region and a doublet from the two benzylic protons. In the 13 C NMR spectrum of the compound the characteristic amide signal resonated at 5 164.26 while the MS spectrum exhibited a molecular ion at the required m/z 262. Generally a tertiary amine was added to the reaction mixture to neutralise the hydrogen halide formed in the reaction. No significant change in the reaction was observed by using triethylamine instead of the higher boiling tri-n- butylamine. On changing the nucleophile to tryptophan methyl ester (as the HCI salt) the addition of a solvent such as DMF was necessary to dissolve all of the crystalline material. However, the aminocarbonylation to provide 2.114 proceeded in a very low yield (6.5 %) and could not be improved substantially by reacting the tryptophan methyl ester as the free base or by changing to another ligand, e.g. P(o-to1)3. Therefore, the reaction was performed at a higher carbon monoxide pressure (4 bar), but to our disappointment the yield improved to only 42 %. A closer look at the reaction mixture revealed the presence of a substantial amount of insoluble material on the side of the reaction vessel, presumably the tryptophan methyl ester as its HCI salt, which might explain the relatively low yield obtained at the higher CO pressure. The next investigation revolved around the improvement of the solubility of the tryptophan methyl ester HCI salt by changing the counterion to trifluoracetate. To our delight, the aminocarbonylation with 2-chloroquinoline and tryptophan methyl ester as the trifluoroacetate proceeded smoothly and was completed within 6 hours at 80 °C and 4 bar of CO pressure. Chromatography afforded 18 % unchanged starting material and 70 % of the required amide 2.114 with its 11-1 NMR, 13C NMR and MS data confirming the assigned structure. In a series of optimisation reactions it was shown that lower pressures resulted in slower reactions with the palladium metal being deposited within 4 hours at 1.2 bar. No improvement was observed with THE as solvent or performing the reaction at 100 °C. The reaction was also successfully repeated with tryptamine as nucleophile to furnish amide 2.115 (Scheme 2.46), characterised by II-1 NMR, "C NMR and MS spectroscopy. 93

H N

O O N H 2.113 2.114 R = CO2Me 2.115 R = H

SCHEME 2.46: PREPARATION OF AMIDES VIA CARBONYLATION

Since good results were achieved with the model compound, 2-chloroquinoline, the next step involved subjecting the previously prepared AB derivatives 2.84 - 2.86 to the same protocol. However, the allyl compound 2.84 yielded a plethora of compounds as indicated on TLC. This may be due, in part, to the competitive carbonylation of the allyl ether moiety, in particular since aryloxy groups are good leaving groups. This possibility was demonstrated by reacting 2-allyloxyquinoline 2.75 under the same reaction conditions which resulted in the isolation of a white crystalline product corresponding to the ester 2.116 (Scheme 2.47). The formation of this compound also requires a base catalysed double bond isomerisation to yield a conjugated system. The 11-1 NMR of the compound showed a three proton doublet at 5 1.88 of the methyl protons while the characteristic ester carbonyl could be found at 5 164.96 in the 13C NMR.

PdC I2(PPh3 )4 NEt3 CO (4 atm), N(Eti3 111•• CO2Me 0 0 oyo NH2 r 2.75 2.116

SCHEME 2.47: AN UNEXPECTED ALLYLIC CARBONYLATION

94

Both the benzyl and methyl turned out to be stable protecting groups when tested under comparable reaction conditions. Therefore, it was totally unexpected when the aminocarbonylation of 2.86 also failed, showing mainly starting material and very little of a new product on TLC. The addition of more catalyst made no difference. Even in presence of tryptamine the reaction proceeded poorly, indicating that the difficulties could possibly be inherent to the substrate 2.86. Thus, various options for improving this very important reaction in our strategy for lavendamycin synthesis were considered.

If the rate limiting step was the insertion of the palladium into the C-Cl bond, the rate of insertion could be improved by employing a better leaving group such as Br, I or OTf. However, no remarkable difference between the aminocarbonylation of 2-bromopyridine and 2-chloropyridine, respectively, was observed. A possibility to improve the situation was to introduce an iodide as the leaving group using an adaptation on our method for preparing C-2 bromides. Thus, the N-oxide 2.83 was treated with triflic anhydride in the presence of tetrabutylammonium iodide. However, this method furnished only 8 % of the desired 2-iodo derivative 2.117 while the major product was the dimer 2.118 (Scheme 2.35). As these attempted modifications were unsuccessful, the search for an improvement of the aminocarbonylation of 2.86 continued.

Tf20 TBAI OMe OMe 2.83 2.117 2.118

SCHEME 2.48: ATTEMTED IODONATION OF QUINOLINE N-OXIDES

Another consideration taken into account was the possibility that the electron rich methoxy group influences the insertion of the palladium into the C-Cl bond. If it retards the insertion step, it was 95 reasoned that an electron withdrawing protecting group such as a carbamate (Scheme 2.48) would have the opposite effect, thereby hopefully improving the reaction. Although the carbamoyl derivative 2.119, could be prepared and fully characterised, the reaction was sluggish and low- yielding. In addition, the subsequent N-oxidation failed and the required leaving group (Cl or Br) could, therefore, not be introduced into the C-2 position.

(iP02NCOCI Acetone DBU OH

2.119

SCHEME 2.48: THE ATTEMPTED USE OF A CARBAMATE AS A PROTECTING GROUP

Since the envisaged chemical transformations were unsuccessful, the last remaining option was to investigate what influence the physical conditions such as temperature and CO pressure have on the aminocarbonylation, especially the stability of the catalyst. One major problem was that the palladium tends to precipitate during the reaction. It was decided to initially study alkoxycarbonylation 54 rather than aminocarbonylation for practical reasons: the nucleophile will be used in excess (as the solvent) while the methyl ester product would be easy to handle as would be, interpretation of the spectra. Although the methoxycarbonylation of 2-chloroquinoline did not proceed at all under atmospheric pressure at 100 °C, the required product 2.120 (see Scheme 2.49) was obtained in 35 % yield under a CO pressure of 4 bar and at 80 °C. Under a pressure of 4 bar and 105 °C the reaction almost went to completion and under 5 bar and 105 °C a quantitative conversion could be observed on TLC when the reaction was complete. However, in the case of the electron rich system 2.86 under 5 bar CO pressure and 105 °C the palladium precipitated after one hour. Therefore, both the temperature and pressure were lowered to 4 bar and 90 °C, respectively. This resulted in isolation of the desired product 2.121 in a yield of 72 %. The 1 H NMR spectrum of 96 the compound confirmed the assigned structure showing the expected two methoxy singlets and the five aromatic protons resonating as four doublets and one triplet. The 13C NMR displayed the ester carbonyl carbon at 5 165.78 while the MS showed the expected M ± at m/z 217 with characteristic fragmentions occurring at m/z 202 (M + - CH3) and m/z 156 (1‘4 + - OCH3 - OCH2), respectively.

PdC13(PPh 3)4 , CO

1 N(Et)3, Me0H CO2Me R R = H 2.120 R = H 2.86 R = OMe 2.121 R = OMe

SCHEME 2.49: ALKOXYCARBONYLATION OF SELECTED SUBSTRATES

Since the previous aminocarbonylation of 2.86 with the tryptophan derivative under 4 bar and 80 °C was a failure, but the corresponding alkoxy carbonylation under 4 bar and 90 °C a success, it was decided to perform the aminocarbonylation at 4 bar and 105 °C. However, the palladium precipitated after two hours with the amide 2.111a obtained in a low yield of 40 %. By increasing the CO pressure to 5 bar and eventually to 10 bar, the yields of the amide 2.111a could be increased to 60 % and 89 %, respectively. The assigned structure of the amide 2.111a (Scheme 2.50) was confirmed by its complex 1H NMR5 featuring, amongst others, the characteristic two methoxy singlets, the doublet of triplets for the CH proton and the doublet of the methylene protons. The important amide signal was displayed in the 13C NMR at S 164.30 and the MS showed a m/z 404 corresponding to Ise + 1, typically from strong basic compounds in the EI-MS. The reaction was also successfully repeated with the benzyl derivative 2.85 and with the 2-bromo compound 2.88 (Scheme 2.50) showing a quantitative conversion on TLC to 2.111a and 2.111b, respectively. 97

NH3 OOCCF 3 co2Me

H

CO, PdC12(PPh3)4 N(Et)3 , DNF OR OR co2me H 2.85 R = Bn, X = CI 2.111a R = Me 2.86 R= Me, X = CI 2.111b R=Bn 2.88 R= Me, X = Br

SCHEME 2.50: AMINOCARBONYLATION OF 2-CHLORO- AND 2-BROMOQUINOLINE DERIVATIVES

It was reasoned that the difference experienced with the carbonylation of the electron rich systems compared to that of 2-chloroquinoline could not be ascribed to electronic effects alone. Therefore, another electron rich derivative, compound 2.123 was subjected to aminocarbonylation (Scheme 2.51). It is of interest to note that both the 2- and 4-chloroderivatives 2.123 and 2.124 (32 % vs. 56 %) were obtained in the halogenation reaction of N-oxide 2.122 with phosphorous oxychloride, even at low temperatures.

The aminocarbonylation of 2.123 proceeded smoothly at 4 bar and 105 °C with no palladium being deposited. The corresponding amide 2.125 was obtained in a yield of 82 % as illustrated in Scheme 2.51. The proposed structure was confirmed by the relevant 1 H NMR, 13C NMR and MS data. The inefficiency of the aminocarbonylation of 2.85 and 2.86 compared to that of 2.123 may be due to the ability of the former to act as bidentate ligands, thus deactivating the Pd towards the oxidative addition step of the aminocarbonylation sequence. 98

CI MeO MeO POCI3 MeO (CH2CO2 Np CI 09 2.122 2.123 (32 %) 2.124 (56 %)

CO, PdC12(PP113)4

N(Et)3, DrvF 9 e

NH300CCF3

CO2Me N

MeO

CO2Me H 2.125 (82 %)

SCHEME 2.51: ADDITIONAL EXAMPLE OF AMINOCARBONYLATION OF A 2- CHLOROQUINOLINE DERIVATIVE

Since the amides 2.111a. and 2.114 were prepared successfully, the next step was the Bischler- Napieralski cyclisation to furnish the pentacyclic lavendamycin precursor. The Bischler- Napieralski cyclisation has been previously employed by the groups of Kende ° and Rao8 on route to lavendamycin, reporting respective yields of 40 and 88 %. However, when using the experimental conditions which Rao reported (xylene, 4 h at reflux) for amide 2.114, the pentacyclic product 2.126 was isolated in the unexpectedly low yield of 23 % with lots of black tar remaining in the flask. Various attempts to improve the yields by the addition of a Lewis acid such as ZnCl2, lowering the reaction temperature (xylene, 90 °C) or by changing the reagent to triflic anhydride were unsuccessfit1. 8 The attempted cyclisation of the tryptamine derived amide 2.115 resulted only in degradation of the starting material while cyclisation of the lavendamycin precursor 2.111a. gave 2.112a in a yield of 18 %. The I H NMR spectrum of 2.112a was characterised by an expected 99 additional aromatic singlet not present in the starting material and a low field broad singlet at 5 12.50 of the indolic NH, hydrogen bonded to the quinoline nitrogen.

POC11 , R1 xylene reflux

2.115 R = H, R1 = H 2.23 R= H, R 1 = H (0 %) 2.126 R = H, R 1 = CO2Me (23 %) 2.114 R = H, R1 = CO2Me 2.111a R = OMe, R1 = CO2Me 2.112a R = OMe, R1 = CO2Me (18 %)

SCHEME 2.52: APPLICATION OF THE BISCHLER NAPIERALSKI CYCLISATION IN THE SYNTHESIS OF PENTACYCLIC SYSTEMS

No cyclisation products could be isolated when the Bischler-Napieralski reaction was carried out in degassed solvents indicating that oxygen is required for the process of aromatisation. Clearly unoxidised indolic intermediates are polymerised rapidly under the reaction conditions. Since it is known that indoles are acid-sensitive and that they may decompose under the reaction conditions, it was suggested that higher yields would be achieved over three to four successive steps by stopping the reaction at relatively low conversion and recycling the starting material. This has been achieved and yields as high as 60 % on 30 % conversion were obtained.

A French groups recently reported that they could not reproduce Rao's results and eventually opted for the Pictet-Spengler cyclisation utilising 2-formylquinolines. Clearly the problem encountered in the application of the Bischler-Napieralski reaction to indolic substrates, which are extremely sensitive to the harsh reaction conditions, has not been addressed and is presently the subject of investigation.55 It is also of interest to note that a comparison of the results of the attempted 100 cyclisation of 2.114 and 2.115 suggests that the aromatisation of the C-ring is facilitated by the presence of an enolisable hydrogen in its ring.

The results discussed in this chapter of the thesis established methodology for the synthesis of the pentacyclic system of lavendamycin. On optimisation of the yields of all steps in the sequence, completion of the synthesis will follow the methodology that has been proposed for the cross- coupling approach. This will entail a sequence of nitration, reduction and oxidation to complete the left hand side while hydrolysis, lithiation and introduction of a suitable electrophile will complete the right hand side. Although the synthetic approach for lavendamycin described in this thesis is not shorter than any of the published methods, it is unique in that the same methodology will allow access to a range of lavendamycin analogues required for the study of physiological structure- activity relationships.

2.7 REFERENCES

(a) K. Rao and W.P. Cullen, Anibiotica. Annu., 1959-1960, 950. (b) K.V. Rao, K. Biemann and R.B. Woodward, J. Am. Chem. Soc., 1963, 85, 2532. (c) Y.H. Chiu and W.N. Lipscomb, J. Am. Chem. Soc., 1975, 97, 2525. T.W. Doyle, D..M. Balitz, R.E. Grulich, D.E. Nettleton, S.J. Gould, C. Tann and A.E. Moews, Tetrahedron Lett., 1981, 22, 4595. A.J. Herlt, R.W. Rickards and J.P Wu, J. Antibiot., 1985, 38, 516. (a) M.A. Ciufolini and M.J. Bishop, J. Chem. Soc. Chem. Commun., 1993, 1463. (b) M.A. Ciufolini and N.E. Byrne, J. Chem. Soc. Chem. Commun., 1988, 1230. C. Barbier, A. Joissains, A. Commercon, J.-F. Riou and F. Huet, Heterocycles, 2000, 53, 37. (a) A.S. Kende and F.H. Ebetino, Tetrahedron Lett., 1984, 25, 923. (b) A.S. Kende, F.H. Ebetino, R. Battista, D.P. Lorah and E. Lodge, Heterocycles, 1984, 21, 91. A.V.R. Rao, Recent Progress in the Chemical Synthesis of Antibiotics, G. Lukacs and M. Ohno, Ed., Springer: Berlin, 1990, 497. 101

(a) A.V.R. Rao, S.P. Chavan and L. Sivadasan, Indian J. of Chem., 1984, 496. (b) A.V.R. Rao, S.P. Chavan and L. Sivadasan, Tetrahedron, 1986, 42, 5065. M. Behforouz, H. Zarrinmayeh, M.E. Ogle, T.E. Riehle and F. W. Bell, J. Heterocyclic Chem., 1988, 25, 1627. (a) M. Behforouz, Z. Gu, W. Cai, M.A. Horn and M. Ahmadian, J. Org. Chem., 1993, 58, 7089. (b) M. Behforouz, J. Haddad, W. Cai, M.B. Arnold, F. Mohammadi, A.C. Sousa and M.A. Horn, J. Org. Chem., 1996, 61, 6552. S. Hibino, M. Okazaki, K. Sato, I. Morita and M. Ichikawa, Heterocyles, 1983, 20, 1957. (a) S. Hibino, M. Okazaki, M. Ichikawa, K. Sato and T. Ishizu, Heterocyles, 1985, 23, 261. (b) S. Hibino, M. Okazaki, M. Ichikawa, K. Sato, A. Motoshima and H. Ueki, Chem. Pharm. Bull., 1986, 34, 1376. (a) P. Molina, F. Murcia and P.M. Fresneda, Tetrahedron Lett., 1994, 35, 1453. (b) P. Molina, P.M. Fresneda and M. Canovas, Tetrahedron Lett., 1992, 33, 2891. (a) D.L. Boger and J.S. Panek, Tetrahedron Lett., 1984, 25, 3175. (b) D.L. Boger, S.R. Duff, J.S. Panek and M. Yasuda, J. Org. Chem., 1985, 50, 5782. (c) D.L. Boger and M. Yasuda, Heterocycles, 1986, 24, 1067. (d) M. Yasuda and D.L. Boger, J. Heterocyclic Chem., 1987, 24, 1253. (e) D.L. Boger, M. Yasuda, L.A. Mitscher, S.D. Drake, P.A. Kitos and S.C. Thompson, J. Med. Chem., 1987, 30, 1918. (1) D.L. Boger, Strategies and Tactics in Organic Synthesis, T. Lindberg, Ed., Academic Press Inc.: San Diego, 1989, vol. 2, chapter 1. (a) A. Godard, P. Roccca, J.-M. Fourquez, J.-C. Rovera , F. Marsais and G. Queguiner, Tetrahedron Lett., 1993, 34, 7919. (b) P. Roccca, F. Marsais, A. Godard and G. Queguiner, Tetrahedron Lett., 1993, 34, 2937. A. Godard, P. Roccca,V. Pomel, L. Thomas-dit-Dumont, J.-C. Rovera , J.F. Thaburet, F. Marsais and G. Queguiner, J. Organomet. Chem., 1996, 517, 25. (a) C.L. Dwyer and C.W. Holzapfel, Tetrahedron, 1998, 54, 1993. (b) C.W. Holzapfel and C.L. Dwyer, Heterocycles, 1998, 48, 215. (c) C.W. Holzapfel and C.L. Dwyer, Heterocycles, 1998, 48, 1513. (d) R. Crous, C.L. Dwyer and C.W. Holzapfel, Heterocycles, 1999, 51, 721. C.L. Dwyer, Ph.D. dissertation: Approaches to the synthesis of streptonigrin, Rand Afrikaans University, 1999. 102

(a) B.E. Love, Organic Preparations and Procedures Int.,1996, 28, 3. (b) F. Bracher and D. Hildebrand, Ann., 1992, 1315. (c) F. Bracher and D. Hildebrand, Tetrahedron, 1994, 50, 12329. (a) P.A. Claret in Comprehensive Organic Chemistry, D.H. Barton and D. 011is, Eds., Pergamon Press: London, 1979, vol. 4, 155. (b) R. Acheson, An Introduction to the Chemistry of Heterocyclic Compounds, Interscience Publishers: New York, 1967, 261. V. Petrow and B.J. Sturgeon, J. Chem. Soc., 1954, 570. (a) M. Shibasaki, C.D.J. Boden and A. Kojima, Tetrahedron, 1997, 53, 7371. (b) A. de Meijere and F.E. Meyer, Angew. Chem. Int. Ed. Engl., 1994, 33, 2379. T. Kametani and K. Ogasawara, Yakugaku Zasshi, 1965, 8, 985. ( CA 1966, 64: 661 le) E.M. Hindmarsch, I. Knight and R. Robinson, J. Chem. Soc., 1917, 110, 943. (a) R.G. Pews, Z. Lysenko and P.C. Voscjpka, J. Org. Chem., 1997, 62, 8255. (b) M. Makosza and K. Sienkiewicz, J. Org. Chem., 1990, 55, 4979. F.D. Bellamy and K. Ou, Tetrahedron Lett., 1984, 25, 839. J. March, Advanced Organic Chemistry: Reactions, Mechanisms and Structure (3rd Ed.), John Wiley and Sons: New York, 1977, 616. J.R. Beck, Tetrahedron, 1978, 34, 2057 and cited references. T.M. Chapman and E.A. Freedman, Synthesis, 1971, 591. (a) G.W. Kabalka and R.J. Varma in Comprehensive Organic Synthesis, B.M. Trost and I. Fleming, Eds., Pergamon Press: Oxford, 1991, vol. 8, 367 - 373. (b) B.H. Han, D.H. Shin and S.Y. Cho, Tetrahedron Lett., 1985, 26, 6233. (c) D.G. Desai, S.S. Swami and S.B. Hapase, Synth. Commun., 1999, 29, 1033. (d) S.A. Eturi and S. Iyer, Synth. Commun., 1999, 29, 2431. (e) D.Nagaraja and M.A. Pasha, Tetrahedron Lett., 1999, 40, 7855. (1) P.S. Kumbhar, J. Sanchez-Valente and F. Figueras, Tetrahedron Lett., 1998, 39, 2573. (a) C.L. Perrin, J. Am. Chem. Soc., 1977, 99, 5516. (b) L. Eberson and F. Radner, Acc. Chem. Res., 1987, 20, 53. (a) Kirk-Othmer, Encyclopedia of Chemical Technology 4th Ed., Wiley-Interscience: New York, 1992, vol. 2, 586. (b) A. Zoran, 0. Khodzhaev and Y. Sasson, J. Chem. Soc. Chem. Commun., 1994, 2239. (a) H. Landscheidt, A. Klausener and H.U. Blank, Eur. Pat. Appl. EP 569,792 to Bayer A.- G., 18 Nov. 1993. (CA 1994, 120: 163724z). (b) ) H. Landscheidt, A. Klausener and H.U.

103

Blank, Ger. Offen, DE 4,023,056 to Bayer A.-G., 23 Jan 1992. (CA 1992, 116: 151302q). (c) Miura, K., T. Kurihara and K. Yanagida, Jpn. Kokai Tokkyo Koho JP 04,149,160 [92,149,160] to Koriyama Kasei Co., Ltd., 22 May 1992. (CA 1992, 117: 170969z) 34. T.B. Patrick, J.A. Schield and D.A. Kirchner, J. Org. Chem., 1974, 39, 1758. 35. (a) I.D. Entwistle, T. Gilkerson, R.A.W. Johnstone and R.T. Telford, Tetrahedron, 1978, 34, 213. (b) C.S. Rondestvedt Jr. and T.A. Johnson, Synthesis, 1977, 850. 36. (a) J.P Phillips, E.M. Barrall and R. Breese, Kentucky Acad. Sci., 1956, 17, 135. (b) G.R. Pettit, W.C. Fleming and K.D. Paull, J. Org. Chem., 1968, 33, 1089. 37. (a) Z.-X. Guo, A.N. Cammidge, A. McKillop and D. C. Horwell, Tetrahedron Lett., 1999, 40, 6999. (b) J. de Ruiter, A.N. Brubaker, W.L. Whitmer and J.L. Stein, J. Med. Chem., 1986, 29, 2024. 38. D.H. Bremner, K.R. Sturrock, G. Wishart, S.R. Mitchell, S.M. Nicoll and G. Jones, Synth. Commun., 1997, 27, 1535. 39. E. Pretsch, T. Clerc, J. Siebel and W. Simon, Tabellen zur Strukturaufkliirung Organischer Verbindungen, Springer-Verlag: Berlin, 1976. 40. A.R. Katritzky and J.M. Lagowski, Chemistry of the Heterocyclic N-Oxides, Academic: New York, 1971, 172. (b) E. Scriven in Comprehensive Heterocyclic Chemistry, A.R. Katritzky and C.W. Rees, Eds., Pergamon Press: Oxford, 1984, vol. 2, 225. (c) ibid. vol. 2, 227. 41. (a) T. Urbanski, CA 1958, 53: 375g. (b) P. Sutter and C.D. Weis, J. Heterocyclic Chem., 1986, 23, 29. 42. (a) W.M. Whaley and T.R. Govindachari, Org. React., 1951, 6, 151. (b) E.D. Cox and J.M. Cook, Chem. Rev., 1995, 95, 1797. (c) M. Cain, R.W. Weber, F. Guzman, J.M. Cook, S.A. Barker, K.C. Rice, J.N. Crawley, S.M. Paul and P. Skolnick, J. Med. Chem., 1982, 25, 1081. (d) W.M. Whaley and T.R. Govindachari, Org. React., 1951, 6, 74. (e) S. Nagubandi and G. Fodor, Heterocycles, 1981, 15, 165. 43. (a) P. Molina , P.M. Fresneda, S. Garcia-Zafra and P. Almendros, Tetrahedron Lett., 1994, 35, 8851. (b) P. Molina and P.M. Fresneda, J. Chem. Soc. Perkin Trans. I, 1988, 1819. (c) P. Molina , P.M. Fresneda, and P. Almendros, Tetrahedron, 1991, 47, 4175. 44. M. Dekhane and R.H. Dodd, Tetrahedron, 1994, 50, 6299. 45. P. Molina, P.M. Fresneda and P. Almendros, Tetrahedron Lett., 1992, 33, 4491. 104

(a) B.C. Soderberg and J.A. Shriver, I Org. Chem., 1997, 62, 5838. (b) M. Akazome, T. Kondo and Y. Watanabe, J. Org. Chem., 1994, 59, 3375. (a) E. Scriven in Comprehensive Heterocyclic Chemistry, A.R. Katritzky and C.W. Rees, Eds., Pergamon Press: Oxford, 1984, vol. 2, 189.(b) P.J. Brignell, A.R. Katritzky and H.O. Tarhan, J. Chem. Soc. (B), 1968, 1477. (c) G.R. Lappin, and F.B. Slezak, J. Amer. Chem. Soc., 1950, 72, 2806. (a) A. Batch and R.H. Dodd, Heterocycles, 1999, 50, 875. (b) V. Snieckus, Chem. Rev., 1990, 90, 879. (c) A. Mehta and R.H. Dodd, J. Org. Chem., 1993, 58, 7587. H.M. Colquhoun, D.J. Thompson and M.V. Twigg in Carbonylation: Direct synthesis of carbonyl compounds; Plenum Press: NewYork and London, 1991. (a) D. Najiba, J.-F. Carpentier, Y. Castanet, C. Biot, J. Brocarrd and A. Mortreux, Tetrahedron Lett., 1999, 40, 3719. (b) H. Horino, H. Sakaba and M. Arai, Synthesis, 1989, 715. (c) A. El-ghayoury and R. Ziessel, Tetrahedron Lett., 1998, 39, 4473. (d) G.G. Wu, Y. Wong and M. Poirier, Org. Lett., 1999, 1, 745. F. Mongin, F. Trecourt and G. Queguiner, Tetrahedron Lett., 1999, 40, 5483. J.N. Freskos, D.H. Rippin and M.L. Reilly, Tetrahedron Lett., 1993, 34, 255. A. Schoenberg and R.F. Heck, J Org. Chem., 1974, 39, 3327. M.A. Ciufolini, J.W. Mitchell and F. Roschangar, Tetrahedron Lett., 1996, 37, 8281. (a) J. Leonard, D. Appleton and S.P. Feamley, Tetrahedron Lett., 1994, 35, 1071. (b) M. R. Jorgensen, C.D. McCleland and B. Taljaard, J. Chem. Res. (5), 1997, 356. (c) S. Doi, N. Shirai and Y. Sato, J. Chem. Soc. Perkin Trans 1, 1997, 2217. CHAPTER 3

MELATONIN AS A SYNTHETIC TARGET

3.1 THE HISTORY OF MELATONIN

The compound, 5-methoxy-N-acetyltryptamine, better known as melatonin 3.1, is of significant importance as a neurohormone of the vertebrate pineal gland and regulates a variety of endocrinological, neurophysiological and behavioural functions in vertebrates.L 2 It was isolated in 1959 from the pineal gland in the brains of cattle on the basis of its skin lightening properties in amphibia. However, this action has not been found to extend to mammals, but some of its other properties have attracted widespread interest e.g. it is now recognised to be the regulator of circadian rhythm in humans and of seasonal breeding in different animal species. 3

MeO NHCOCH3

H 3.1 melatonin

SCHEME 3.1

Pharmacological administration of this unique hormone in humans has been shown to alleviate the symptoms of jet-lag, induce sleep and advance the sleep phase of subjects with delayed sleep phase syndrome while studies on elderly people and depressed patients suggest that it plays a role in ageing and seasonal depression. It has also been reported that the administration of melatonin 3.1, in combination with other methods, could also be useful for the treatment of AIDS as well as for tumoral prophylaxis and therapy. 4

All these potential applications of melatonin has not only stimulated interest in its synthesis, but to that of analogues such as azaindole 3 and benzimidazolic5 bioisosteres. 106

3.2 THE SYNTHESIS OF MELATONIN AND RELATED DERIVATIVES

Since other tryptamine derivatives such as serotonin 3.2 and 5-methoxytryptamine 3.3 are structurally related to melatonin 3.1 and readily transformed into the latter, the following discussion on the synthesis of melatonin will also include relevant approaches to 3.2 and 3.3. Treatment of 5-methoxytryptamine 3.3 with acetic acid in pyridine at room temperature forms the N,N-bis acetylated derivative, which can be converted to melatonin 3.1 in 80 % yield. The Hag6 and Nestle7 companies base their approach to melatonin 3.1 on the extraction, isolation and chemical derivatisation of the natural 5-hydroxytryptamides from coffee waxes, present in the outer layer of coffee beans (Scheme 3.2).

HO HO

NHCO(01-12)nR NH2 KOH Na2S2O4 H n= 16 or 20 or 22 R = CH3 or CH2OH (I) Ac20 (0) NaOH-Et0H MeO HO I NHAc NHAc Me2SO4

3.1

SCHEME 3.2:THE UTILISATION OF 5-HYYDROXYTRYPTAMIDES FROM COFFEE WAXES

3.2.1 SIDE-CHAIN ATTACHMENT STRATEGIES

An important group of synthetic strategies to melatonin employs various methods for side chain attachment onto the indole ring of a suitable starting material such as 5-methoxyindole. Thus, the gramine8, Knoevenagel 8 and oxalyl9 procedures, all depending on the fact that the indolic C- 3 position is nucleophilic in character, were successfully applied to give ready access to 5- 107 methoxytryptamine 3.3 and therefore, melatonin 3.1. The procedures followed are outlined in Scheme 3.3.

MeO MeO MeO CHO NO2 NH2 CH3NO2 LiAIH4 N NH40Ac Knoevenagel 111 3.3

MeO MeO MeO NMe2 CH2O, NHMe2 CH31 1 . Gramine KCN

Oxalyl chloride

Me0 0 CI MeO o NH31. NH2

H

SCHEME 3.3: SIDE CHAIN ATTACHMENT STRATEGIES TO MELATONIN

In another variant of this approach, Flaugh l° introduced the ethylacetamide chain to 5- methoxyindole by reaction with nitroethene (generated in situ by thermolysis of nitroethyl acetate) followed by hydrogenation of the nitro group and finally, acetylation to furnish melatonin 3.1 in 51 % overall yield (Scheme 3.4).

MeO MeO MeO NO2( NHCOCH3 Xylene, reflux .) H2, Pt02 (ii) Ac20/Py AGO NO2 H 3.1 (51% overall)

SCHEME 3.4: FLAUGH'S APPROACH TO MELATONIN 108

The introduction of an alkoxy or hydroxy group at C-5 of tryptamines has been pursued actively by Somei and coworkers i I via their unique method of preparing 1-hydroxyindoles followed by nucleophilic substitution and dehydroxylation. Thus, a four step process to melatonin 3.1 was established starting from tryptamine (Scheme 3.5).

NHAc NHAc EtiSiH) CF3CO2H

3.4 Na2W 04 1 H20 2 MeO NHAc NHAc BF3/Me0H

H OH 3.1 (85 %) 59 % (over 3 steps)

SCHEME 3.5: PREPARATION OF MELATONIN BY SOMEI

3.2.2 APPLICATION OF THE FISCHER INDOLE REACTION

The majority of reported synthesis of melatonin 3.1 was based on the versatile and widely studied Fischer indole 12 reaction and modifications thereof which proceeds via the formation of a phenylhydrazone followed by a [3,3]-sigmatropic rearrangement step and finally, cyclisation to the indole nucleus. Although melatonin 3.1 was used as an intermediate in the synthesis of carbolines as early as 1930, it was neither purified nor characterised.' The preparation of this crude melatonin proceeded via a modified Fischer indole reaction starting from 5-methoxy- phenylhydrazine and 4-aminobutanal diethylacetal followed by acetylation (Scheme 3.6). In another variation melatonin 3.1 was prepared in a low-yielding (26 %) one step process utilising N-acetyl-4-aminobutanal diethylacetal.

109

Et0 MeO NH2 MeO Ett0)..."-----%"---- NH2

ZnCl2, 170° C NHNH2

lAc20

MeO NHCOCH3

SCHEME 3.6: PREPARATION OF CRUDE MELATONIN

The method of choice of Abramovitch and Shapiro ° to form the required phenylhydrazone 3.5 was the base catalysed Japp-Klingemann' 4 coupling of diazotised 4-methoxyaniline and 2- oxopiperidine-3-carboxylic acid as shown in Scheme 3.7. This was followed by a Fischer indole cyclisation to yield tetrahydro-B-carboline 3.6, alkaline hydrolysis and finally, decarboxylation to furnish 5-methoxytryptamine 3.3 (41 % overall yield) which can be readily converted into melaton in 3.1.

MeO MeO MeO >

3.3 3.6

V MeO MeO +

3.5

SCHEME 3.7: ABRAMOVITCH AND SHAPIRO'S APPROACH TO TRYPTAMINE 3.3

110

The yield of the decarboxylation step is strongly influenced by the nature of the aromatic substituents. Thus, the decarboxylation of 5-substituted tryptamine-2-carboxylic acids decreases in the order OMe > Me > H > Cl > NO2 in acidic medium. However, a slight modification involving side chain N-acylation prior to a thermal decarboxylation step catalysed by copper- quinoline resulted in the decarboxylation less dependent on the nature of the aromatic substituent.

In their strategy Frashini and coworkers 4 employed a phthaloyl protecting group for the side chain amino group of tryptamine on route to melatonin 3.1 as shown in Scheme 3.8.

MeO MeO

73 % 1 4 steps 3.1 (46.5 %)

SCHEME 3.8: THE APPROACH BY FRASHINI TO MELATONIN

In the Grandberg is modification of the Fischer indole reaction 4-chlorobutanal was reacted with 4-methoxyphenylhydrazine to furnish the required 5-methoxytryptamine 3.3 in a yield of 45 %. By employing the N-benzylderivative of 4-methoxyphenylhydrazine the yield was increased to 70 %. In Scheme 3.9 it is shown that the ene-hydrazine 3.8 undergoes cyclisation to form a second ring product 3.9 prior to the (3,3]-sigmatropic shift occurring, to give 3.10. A tricyclic intermediate 3.11 is then formed which opens to give the tryptamine derivative whereby both the nitrogen atoms of the starting material have been utilised. This Grandberg modification which is atom efficient has been used for the synthesis of a wide variety of substituted tryptamines e.g. it has been applied in the preparation of the anti-migraine drug, sumatriptan 16 (see section 3.3.2). tit

MeO MeO MeO II N—N N—N r\ N—No CI FR R R H R=H or Bn 3.8 3.9

MeO MeO 3.3 NHR R 3.11 3.10

SCHEME 3.9: THE METHOD EMPLOYED BY GRANDBERG TO PREPARE TRYPTAMINES

Another reaction which was also successfully applied in the preparation of 5-substituted tryptamines is the Horner-Emmons" reaction of 3-oxo-2,3-dihydroindoles as outlined in Scheme 3.10.

Me0 MeO 0 (EtO)2P(0)CH2CN CN NaH, THE

(i) NaOH, Me01- (ii) Ra-Ni, H2

MeO 1 NH2

88 %

SCHEME 3.10: APPLICATION OF HORNER-EMMONS REACTION TO TRYPTAMINES

112

The patented Batcho-Leimbruber ls route to indoles involves the preparation of 5-methoxyindole which is subsequently converted into melatonin 3.1. The aminomethylation of the appropriately substituted nitrotoluene 3.12 with pyrrolidine acetal produces enamine 3.13 which on reduction to the corresponding aniline undergoes intramolecular cyclisation (Scheme 3.11). A wide range of benzene ring substituted indoles has been prepared by this method.

OMe Me0v( MeO Me H 0 MeO MeO H2, Pd/C

NO2

3.12 3.13 76%

SCHEME 3.11: THE BATCHO-LEIMGRUBER METHOD TO INDOLES

Most of these discussed syntheses to melatonin 3.1 were either based on the attachment of a side chain to a preformed indole ring or on the Fischer indole synthesis. The first approach involves quite a few steps while a major drawback of the Fischer indole reaction is that yields are often low with numerous byproducts being formed.

3.3 NEW APPROACHES TO MELATONIN AND DERIVATIVES

It is clear that new, simple and efficient ways for constructing indole rings are still in great demand with the development of methods to indoles substituted in both of the rings receiving particular attention. ° A number of new and potentially quite versatile methods for the synthesis and functionalisation of indoles employing modem transition metal assisted synthetic methodology have been developed. 20'21 Therefore, we initially considered a strategy to melatonin via a samarium(II)iodide 22 promoted cyclisation. However, this elegant approach (see Scheme 3.12) was abandoned due to the relatively high cost of the dichloroalkyne 3.14. 113

MeO MeO CI CI CI + \ NCO I I CI SmI2 NH N11—/ COR 3.14 COR

I NH3

MeO NH2 3.1 -4—

SCHEME 3.12: A POSSIBLE APPROACH TO MELATONIN UTILISING SAMARIUM

Our attempts to develop a synthesis for melatonin 3.1 had its origin in an approach by a commercial company operating in the import replacement field. This dictated the development of a route which had to be, not only simple and direct, but also commercially viable.

3.3.1 THE HECK REACTION IN A NEW APPROACH TO MELATONIN

A number of indole syntheses use low-valent transition metals 23 for the reduction of aromatic nitro groups to amines, which subsequently react with suitable electrophilic substituents in the ortho positions to form indoles. These typically involve o-nitroaryl starting materials such as 316 and reducing agents such as iron(0), titanium(III)chloride or nickel boride. The nitrostyrene 3.16 and related compounds are generally prepared by aldol-type condensation reactions. 24 However, we envisaged that 3.16 could be synthesised via a Heck reaction of the readily prepared o-nitro triflate 3.15 with nitroethylene under neutral conditions. 25 The use of epichlorohydrin as neutral proton scavenger in Heck reactions has been developed at RAU and applied successfully on route to streptonigrin and lavendamycin (see Chapter 2 of this thesis). It was expected that the tendency of nitroethylene to polymerise under basic conditions 26a would be reduced in the presence of epichlorohydrin. The subsequent introduction of the aminoalkyl

-chain would employ, once again, nitroethylene as reagent as described by Ranganathan et. al.26b (Scheme 3.13).

114

MeO OTf MeO N MeO Heck Reduction and reaction cyclisation NO2 NO2 3.15 3.16

MeO MeO Reduction and NO2 3.1 acetylation

SCHEME 3.13: AN APPROACH TO MELATONIN UTILISING A HECK REACTION (PREPARED BY NITROETHYLENE)

However, a Heck reaction of the model compound, iodobenzene with nitroethylene in the presence of 15 % Pd(PPh3)4 and epichlorohydrin at elevated temperature furnished only a small amount (< 10 %) of nitrostyrene as indicated by GC-MS analysis of the total reaction product. This analysis also indicated the presence of unreacted iodobenzene as well as the major product which presumably resulted from insertion of the palladium into the C-Cl bond of the epichlorohydrin (Scheme 3.14) followed by reaction with nitroethylene.

Pd(PPh3)4 NO2 NO2 O1 0

SCHEME 3.14: PALLADIUM CATALYSED REACTION OF EPICHLOROHYDRIN WITH NITROETHYLENE

3.3.2 A MODIFIED FISCHER INDOLE REACTION

Since the Heck reaction could not be improved by replacing the neutral proton scavenger epichlorohydrin with cyclohexene oxide, the search for an appropriate synthesis for melatonin 3.1 continued. A closer scrutiny of the literature revealed that despite its myriad complications, rearrangements and mechanistic mysteries, the Fischer indole 7 synthesis remains the epitome of

115 indole ring construction methods with recent years yielding several improvements and novel applications of this classical reaction. Since Fischer indole reactions are generally characterised by low yields much effort was directed at yield improvement by changing the nature of the catalyst e.g. replacing heterogeneous catalysts such as zinc chloride with homogeneous catalytic systems such as formic acid, dilute acetic acid or the use of stoichiometric amounts of PC13 in benzene or dichloromethane. The aldehydes employed in the Fischer indole reaction giving access to 2-unsubstituted indoles are prone to oxidation and aldol reactions and are, therefore, often protected as acetals. These are hydrolysed in situ to allow hydrazone formation. An attractive alternative is the protection of the aldehyde as the bisulfite addition product, as utilised in the synthesis of the anti-migraine drug, sumatriptan I6 (Scheme 3.15).

MeNHSO2CH2 C I MeNHSO2CH2 SO3Na NHNH2 OH H sumatriptan

SCHEME 3.15: THE FISCHER INDOLE REACTION TO SUMATR1PTAN

The aldehyde precursor can also be protected intramolecularly as the hemiaminal as shown in Scheme 3.16 in the synthesis of the alkaloid physostigmine. 27

MeO 4-OMeC6H5NHNH2 HCI OBn OBn Pyridine

Me0

SCHEME 3.16: PREPARATION OF THE ALKALOID, PHYSOSTIGMINE

116

These promising results reported in the literature led us to believe that the Fischer indole reaction still had the potential as a economic route to melatonin 3.1. The successful and extensive use of the Fischer reaction by SoII and his colleagues28 at Wyeth-Ayerst to prepare etodolac derivatives (Scheme 3.17) using dihydrofuran again appeared to owe its success to the use of a protected or masked aldehyde.

,HCI,THF 0 NHNH2 (ii) ZnCl2, (HOCH2)2, 170°C Br 48%

SCHEME 3.17: THE PREPARATION OF ETODOLAC DERIVATIVES

By analogy, it was reasoned that cyclic enamide 3.17 could act as synthetic equivalent of the amino aldehyde required for the melatonin synthesis as retrosynthetically shown in Scheme 3.18.

MeO NHCOCH3 MeO > NHNH2 H COCH3 3.1 3.17 3.18

SCHEME 3.18: THE POSSIBLE UTILISATION OF A CYCLIC ENAMIDE TO MELATONIN

The hydrazine component of the Fischer indole synthesis is usually generated from the corresponding aniline by formation of the diazonium salt with HC1/NaNO2 and subsequent reduction by a variety of reagents. Since the hydrazines are often unstable as their free bases, they are generally stored as HCI salts, and can often be used directly as such. The 5- methoxyphenylhydrazine 3.18 required for the synthesis of melatonin 3.1 could be readily prepared from the corresponding aniline which, in turn, is obtained by the selective reduction of nitrobenzene to the corresponding hydroxylamine followed by a Bamberger rearrangement in methanol, a process (Scheme 3.19) described in several patents.29 117

OMe OMe

NO2 NHOH NH2 NHNH2

SCHEME 3.19

It is of interest to note that endocyclic enamides have found considerable application as versatile synthetic intermediates for the synthesis of nitrogen-containing heterocycles of biological significance. 3° However, the methods available for the synthesis of cyclic enamides are rather limited, allowing neither the use of a wide variety of reactants nor reaction conditions. This has probably limited a more extensive use of these moderately reactive N-acyl enamines in organic synthesis.

Stille and co-workers 31 reported a novel, but expensive rhodium or ruthenium hydride mediated isomerisation of N-acyl-3-pyrrolines into the corresponding enamides. Unfortunately, a general catalyst could not be found, and it was necessary to match the transition-metal complex, temperature, and time with a particular substrate to reach optimum conditions. Thus, the isomerisation of N-acyl-3-pyrroline in refluxing xylene under an inert atmosphere provided the amide 3.17 in yields ranging from 60 to 92 %, depending on the catalyst (Scheme 3.20).

HRh(CO)(PPh3)3 0 xylene, 44 h 1■1 Ac Ac 3.17 (92 %)

SCHEME 3.20: STILLE'S APPROACH TO CYCLIC ENAMIDES

Stille32 also employed the above mentioned rhodium catalyst in the hydroformylation of N- allylacetamide which yielded a 54:46 mixture of the a- and 13-amido aldehydes 3.19a and 3.19b, respectively. According to Stille the linear aldehyde 3.19a cyclised readily under the mild reaction conditions to give the desired enamide 3.17 (Scheme 3.21). However, Ojima et. al.33 118 reported different selectivities and products under their reaction conditions (80 °C and 1200 psi), but not observing the formation of N-acetyl-2-pyrroline 3.17.

HRh(CO)(PPh3)3 CHO NHAc NHAc CHO) H2/CO (500 psi), NHAc 40°C, 30 h 3.19a 3.19b 1

Ac 3.17

SCHEME 3.21: HYDROFORMYLATION OF N-ALLYLACETAMIDE

Shono's34 protocol involved anodic oxidation of N-acyl cyclic amines in methanol to furnish the corresponding a-methoxylated amides and carbamates which could either be used as masked aldehydes to prepare I3-substituted indoles or transformed into the correponding cyclic enamides and enecarbamates through acid-catalysed elminination of methanol (Scheme 3.22). However, electrosynthesis has limited use and is inaccessible to many laboratories.

Om - e NH4CI Me0H OMe COR COR COR R = OCH3 (80 %) R = OCH3 (91 %) R= CH3 (45 %) R = CH3 (no results)

SCHEME 3.22: SHONO'S APPLICATION OF ELECTROSYNTHESIS TO CYCLIC ENAMIDES AND ENECARBAMATES

Kraus and Neuenschwander 35 suggested that acylation of 1-pyrroline with an acid chloride or alkyl chlorocarbonate would be an attractive route to 3.17, but 1-pyrroline is unstable and has been little studied. Although a French36 group prepared the unstabilised cyclic imine, 1- pyrroline 3.21 in gramscale by vacuum dehydrochlorination of N-chloropyrrolidine 3.20, readily

119 prepared from pyrrolidine and N-chlorosuccinimide (Scheme 3.23), two drawbacks of this protocol are that the N-chloroamine 3.20 is explosive in nature and has to be kept in solution all the time with the final product 3.21 also unstable.

NCS base O 10-2 torr H 61 3.20 3.21 (93 %)

SCHEME 3.23: A PREPARATION OF 1-PYRROLINE

However, the trimer of 1-pyrroline, 3.22 is readily available by oxidation of pyrrolidine with sodium peroxodisulfate and 0.5 % silver nitrate 37 and has been used as a synthetic equivalent of 1-pyrroline 3.21 as demonstrated in the synthesis of N-acyl-2-pyrrolines by Kraus and Neuenschwander35 (Scheme 3.24).

3.21 3.22 1 CICOR NEt3 -78°C

1 COR 3.17 R = CH3 (71 %) R = OCH2Ph (79 %) R = OCH3 (78 %)

SCHEME 3.24: THE APPROACH TO ENAMIDES VIA PYRROLIDINE TRIMER

Since this last procedure appeared to be safe, practical and economically feasible, it was decided to optimise this approach to endocyclic enamides and enecarbamates. The preparation of the trimer 3.22 ensued following the methodology described by Nomura and co-workers. 37 The

120

silver(I) catalysed oxidation of pyrrolidine with peroxodisulphate afforded the 1-pyrroline trimer 3.22 as a light orange oil in a yield of 56 %. Since the trimer decomposes readily, the solvent was removed under reduced pressure on an ice-bath. Nomura 37 reported that a mass spectroscopic analysis suggested that the trimer 3.22 decomposes under the measurement conditions into the corresponding monomer, 1-pyrroline 3.21. Therefore, the characterisation of the trimer 3.22 relied on the extensive use of NMR spectroscopy. Its 13C NMR spectrum indicated four peaks at 8 81.96, 45.84, 27.86 and 20.25 ppm, respectively which was in complete agreement with its apparent C3 symmetry and the reported NMR data.” The presence of a proton attached to a sp 2 carbon was precluded by its 1HNMR spectrum exhibiting complex multiplets at 8 1.75, 2.25 and 2.95 ppm, respectively.

The formation of 3.22 probably proceeds via the formation of aminyl radical intermediates 3.23. These aminyl radicals are generated in the oxidation of amines under alkaline conditions in analogy to the silver-(I)-catalysed oxidation of alcohols where a silver(II) ion abstracts an electron from the oxygen atom to give alkoxy-radicals. The combination of the aminyl radicals 3.23 gives the corresponding N,N-coupling products while hydrogen abstraction affords the cyclic imine 3.21 which trimerises to 3.22 (Scheme 3.25).

Ag+ + S2082- Ag++ + S042- + SO4

Ag++ ON—NO

3.23 -i-r1 (S04 1 + H --,.. HSOr )

polymers -IF r.11

3.21 3.22

SCHEME 3.25: THE FORMATION OF THE TRIMER

The relatively low yield of 3.22 may be ascribed to overoxidation and polymerisation. Overoxidation will also explain the substantially lower yields (8 - 30 %) with the use of bigger

121 amounts of peroxodisulphate (1.25 and 1.5 equivalents vs. 1 equivalent) and/or silver nitrate (5 % vs. 0.5 %).

The next step in the preparation of N-acy1-2-pyrrolines presented some difficulties. According to Kraus et. al.36 distillation of a 0.1 M tetrahydrofuran (THF) solution of freshly prepared trimer 3.22 into a precooled (-78 °C) flask followed by the addition of triethylamine and an acid chloride furnished the corresponding enamides in yields ranging from 39 % to 79 %, based on the trimer consumed (typically around 55 - 60 %). However, after several attempts in our laboratory very little of the desired enamide 3.25 could be isolated from the complex reaction mixtures. Although Kraus and co-workers reported that they had no success with the direct reaction of alkyl chloroformates with the trimer 3.22, it was decided to explore the possibility again since it has been reported 38 that heating of the trimer 3.22 and diethylphosphite (3 equivalents) together at 85 °C for 90 minutes under argon resulted in complete conversion to the phosphonate 3.24 (Scheme 3.26).

HP0(0E02 ■ 1 1,1 85°C, 90 min PO(OEt)2 LJ H 3.22 3.24

SCHEME 3.26: SYNTHESIS OF 2-PHOSPHONOPYRROLINE DERIVATIVES

Thus, the slow addition of trimer 3.22 to a refluxing solution of benzoyl chloride and pyridine in toluene under argon furnished the enamide 3.25 in a yield of 42 %. The NMR of 3.25 exhibited the characteristic olefinic protons as two triplets of doublets resonating at 8 5.12 and 6.42, respectively with the MS displaying the expected M + at m/z 173. The reaction was also successful when conducted at a lower temperature in refluxing benzene, but by employing Hiinig's non-nucleophilic base, N,N,N-ethyl-diisopropylamine in benzene the yield could be improved significantly (67 %). This one-pot protocol was also applied to the synthesis of N- acetyl-2-pyrroline 3.17 and permitted the preparation of multigram quantities of 3.17 and 3.25 (Scheme 3.27). 122

RCOCI, base 80°C, 30 min (—)Nj COR 3.22 3.17 R = CH3 (58 %) 3.25 R = C6H5 (67 %)

SCHEME 3.27: THE PREPARATION OF THE ENAMIDE VIA A ONE-POT PROCEDURE

The 1 14 and 13C NMR spectra of 3.17 are in correspondence with the reported data in the literature. 31'35 An important feature of the 111 NMR spectrum of 3.17 is its indication of the presence of the two rotamers A and B (Scheme 3.28), with form B the predominating species (B/A = 2.5) as e.g. evidenced by the two multiplets for H-2 resonating at 8 6.37 and 6.85, respectively. Apparently the ground-state energy of form A is higher than form B due to steric repulsion between the methyl group and the methylene group a to the nitrogen. 39

H H I■1 H C=0 C—C H3 CHI 0 A

SCHEME 3.28: THE ROTAMERS OF N-ACYL-2-PYRROLINE

Each of the desired enamides 3.17 and 3.25 were subsequently subjected to a Fischer indole reaction. Thus, phenylhydrazine, N-acyl-2-pyrroline 3.17 and ZnCl 2 in xylene were refluxed under argon (as described by Shone )) until TLC indicated consumption of the enamide. The indole 3.26 formed in the reaction was visualised on TLC as a purple spot by spraying with van Urk's reagent (4-dimethylaminobenzaldehyde-HC1). The product 3.26 was isolated by chromatography in a reasonable yield of 62 %. Its 11-1 NMR spectrum displayed the characteristic signals of a methoxy group, multiplets of CH 2 groups, two broad singlets of two NH protons with five aromatic protons resonating at 5 1.89, 2.95, 3.57, 5.58, 7.01, 7.11, 7.19, 7.35, 7.58 and 8.28. Extension of this methodology to the reaction of phenylhydrazine with N- benzoyl-2-pyrroline 3.25 provided the corresponding indole 3.27 in a 65 % yield with the relevant NMR and MS data in agreement with the proposed structure. 123

The next substrate, 4-methoxyphenylhydrazine is commercially available as the HCI salt since the free hydrazine is unstable. This required a change in reaction conditions. Thus, p- methoxyphenylhydrazine hydrochloride was reacted with enamide 3.17 in an acetic acid/water/ethyl acetate-mixture under reflux to afford the desired melatonin 3.1 in a isolated yield of ca. 75 %. This result represents a vast improvement on previously reported yields as discussed earlier on in this chapter. Reaction times were short (< 30 minutes) with TLC analysis indicating nearly quantitative conversion into the corresponding indoles. The methodology could also be applied to phenylhydrazine hydrochloride, but failed in the case of p- bromophenylhydrazine hydrochloride. In the latter case the reaction only proceed in a mixture of acetic anhydride and glacial acetic acid. This is in line with other observations that Fischer indole syntheses using hydrazines with electron-withdrawing substituents are often sluggish and low-yielding. I2 Purification of the products by column chromatography resulted in some decomposition and consequently lower indole yields. Nevertheless, significantly higher yields of indoles were obtained with the enamide substrates than with the corresponding free aldehydes or their acetals. Another advantage of this procedure is that it gives access to a variety of melatonin analogues.

N COR2 yR2 (i), (ii) or (iii) NHNH2 0 H 1.2 mole eq. ZnCl2, 3.26 Ri = H, R2 = CH3 (62 %) (method i) xylene 3.27 R1 = H, R2 = C6H5 (65 %) (method i) AcOH/EtOH/H20 3.1 R1 = OMe, R2 = CH3 (75 %) (method ii) (25:35:40) 3.28 R1 = OMe, R2 = C6H5 (85 %) (method ii) AcOH/Ac20 3.29 R1 = Br, R2 = C 6H5 (30 %) (method iii)

SCHEME 3.29: RESULTS OF THE FISCHER INDOLE REACTION OF ENAMIDES

It is suggested that the reaction proceeds via a reactive acyliminium ion intermediate 3.30 (Scheme 3.30).

124

N HO OHC'D CO R2 CO R2 CO R2 1LH9 +H -H20

V HN NHNFNIV NO NH— I CO R2 COR2 3.30

ti NH-- ;1 H COR2 H®

11 yR2 CSmin HiN 0 NW.; H COR2

1 -N H3

NHCOR2

NHCOR2 ra N NT) H' H Ho

SCHEME 3.30: A POSSIBLE MECHANISM OF THE REACTION BETWEEN PHENYLHYDRAZINES AND CYCLIC ENAMIDES AND ENECARBAMATES

The report of Martin et. a1.41 that the palladium catalysed reaction of N-protected phenylhydroxylamines 3.31 with vinylacetates affords 2,3-unsubstituted N-acylindoles via a hetero-Cope-rearrangement of the intermediate N-phenyl-O-vinylhydroxylamine derivatives (Scheme 3.31), raised the possibility of a Fischer-indole type reaction of N-protected phenylhydroxylamines 3.31 with cyclic enamides and enecarbamates 3.17 and 3.25. 125

Oak .=/ 0H Li2PdCt4 11 COR COR 3.31 1 (3,31

-H20 CHO NH COR

SCHEME 3.31: THE APPROACH OF MARTIN TO INDOLES

It was expected that such a reaction would proceed similarly to that shown in Scheme 3.30 to furnish 3-substituted indoles. However, although N-acetylphenylhydroxylamine could be readily prepared by the method of Lobo et. al.,42 the results of attempts to prepare indoles using the proposed method were not promising.

An attractive alternative to the above cyclic carbamates in the Fischer indole synthesis involves the use of the corresponding N-acyl-carbinolamines (Scheme 3.32) which could be easily prepared from readily available pyrrolidone 3.32. Protection of the nitrogen atom followed by selective reduction of the resulting N-benzyloxycarbonylderivative 3.33 with LiEt3BH in THE at -78 °C provided the 2-hydroxyl derivative 3.34 45'44 in a quantitative yield. The compound 3.34 occurs as two rotamers (2.5:1) exhibiting two H-2 resonances as singlets at 8 5.48 (major isomer) and 8 4.65. In the mass spectrum the required M + at iniz 221 was of low intensity while the base peak at m/z 203 resulting from the loss of water forms the molecular ion. It was anticipated that this compound would react in the same way as Shono's u-methoxylated cyclic amides to furnish indoles. 4° Thus, reaction of 3.34 with p-methoxyphenylhydrazine under acidic conditions did, indeed, furnish the indole 336 in high yield (>95%).

Attempts to dehydrate the alcohol 3.34 in the presence of acid gave only oligomeric products. Even improvement of the leaving properties of the hydroxyl group by either triflation or mesylation failed to provide the required enecarbamate 3.35 in useful yields. This is in agreement with the problems experienced by Leonard45 in the preparation of enecarbamates. However, the enecarbamate 3.35 126

could be synthesised from the 2-hydroxyl derivative 3.34 by heating with acetic anhydride in the presence of Hiinig's base. The published results46 was shortly afterwards confirmed by Correia et. al.47 who reported that the required 13-elimination could be promoted by trifluoroacetic anhydride in the presence of hindered nitrogenated bases such as 2,6-lutidine.

NaH/CBzCI LiEt3 BH, THE R '0 R OH CO2CH2C6H5 CO2CH2C6H5

3.32 R = H 3.33 R = H 3.34 R = H 3.37 R = COOH 3.39 R = CO2C(CH3)3 3.40 R = CO2C(CH3)3 3.38 R = CO2C(CH3)3

p- OMePhNHNH2

MeO ,C) NHCO2C H2C6H5 R CO2CH2C6H5

3.35 R = H 3.36 R = H 3.41 R = CO2C(CH3)3 3.42 R = CO2C(CH3)3

SCHEME 3.32: THE PREPARATION OF INDOLES FROM N-ACYL CARBINOLAMINES

Application of the methodology described above for the synthesis of cyclic carbinolamines to the N-protected tert-butyl ester 3.39 of L-pyroglutamic acid 3.37 (tert-butyl ester prepared 3.39 by the method of Taschner et a/47'48) lead to the quantitative formation of 3.40 which on acid catalysed reaction with p-methoxyphenylhydrazine afforded 3.42, the L-tryptophan analogue. Dehydration of 3.40 could also be effected by heating with acetic anhydride and base. However, the resultant product 3.41 was difficult to purify and could not be fully characterised.

We were able to expand this methodology for cyclic enamide synthesis successfully to the preparation of 6- and 7-membered cyclic enecarbamates from 8-valerolactam and c-caprolactam, respectively (Scheme 3.33), but the relevant experimental detail is not included in the experimental section. 127

(Cli2)1n (Cm in (CH )n LiEt3 c Ac20 1 T HFBV ) cNI 0 N OH Base C CO2CH2C6H5 CO2CH2C6H5 CO2CH2C6H5 n = 1 or 2 SCHEME 3.33

3.3 SUMMARY AND CONCLUSIONS

In this chapter of the thesis the development of methods for cyclic enamide synthesis was described. One of the methods for the preparation of cyclic enamides involved dehydration of the corresponding N-acyl-carbinolamines which was easily prepared from readily available pyrrolidone. The successful application of these compounds in the Fischer indole synthesis on route to the neurohormone, melatonin 3.1 and other derivatives, was demonstrated.

3.4 REFERENCES

H.M. Hugel and D.J. Kennaway, Org. Prep. Proc. Int., 1995, 27, 3. R.J. Reiter, Endocrinol. Rev., 1991, 12, 151. D. Mazeas, G. Guillaumet and M.-C. Viaud, Heterocycles, 1999, 50, 1065. F. Frashini, L. Di Bella and E. Diranto, Eur. Pat. Appl. EP 330,625, 30 Aug. 1989. P. Depreux, E. Fourmaintraux, D. Lesieur and P. Renard, Synth. Commun., 1994, 24, 2123. S. lyimen, G. Lehman, 0. Neuhoeffer and 0. Vitzum, Ger. Offen., 1973, DE 2156944 Chem. Abstr. 1973, 79, 35126d. R. Bertholet and P. Hirsbrunner, Eur. Pat. Appl. EP 130,303, 1985, Chem. Abstr. 1985, 103, 6154v. J. Smuszkovicz, W.C. Anthony and R.V Heinzelman, J. Org. Chem., 1960, 25, 857. J. Supniewski and S. Misztal, Bull. Acad. Pot Sci. Ser. Biol., 1960, 8, 479, Chem. Abstr. 1961, 55, 15458. M.E. Flaugh, T.A. Crowell, J.A. Clemens and B.D. Sawyer, .1 Med. Chem., 1979, 22, 63. 128

(a) M. Somei, Heterocycles, 1999, 50, 1157 and cited references. (b) M. Somei and N. Oshikiri, M. Hasegawa and F. Yamada, Heterocycles, 1999, 51, 1237. D.L. Hughes, Org. Prep. Proc. Int., 1993, 25, 609. (a) R.A. Abramovitch and D.J. Shapiro, J. Chem. Soc., 1956, 4589. (b) R.A. Abramovitch, J. Chem. Soc., 1956, 4593. (c) S. Misztal and J. Boksa, Pol..I. Pharmacol. Pharm.,1984, 36, 345. J. March, Advanced Organic Chemistry: Reactions, Mechanisms, and Structure, 2nd Ed., McGraw-Hill: London, 1977, 542. (a) I.I. Grandberg, T.I. Zuyanova, N.I. Afonina and T.A. Ivanova, Dokl. Akad. Nauk SSSR, 1967, 176, 583, Chem. Abstr. 1968, 68, 104883k. (b) I.I. Grandberg, N.I. Afonina and T.I. Zuyanova, Khim. GeterosikI Soedin., 1968, 1038., Chem. Abstr. 1969, 70, 68046. F.D. Albinson, D.L. Crookes, J.W.M. Mackinnon, Eur. Pat. Appl. EP 0 462, 837 A2 (1991). Chem Abstr. 1992, 116: P106088p. A. Buzas and C. Herrison, Synthesis, 1977, 129. (a) R.D. Clark and D.B. Repke, Heterocycles, 1984, 22, 195. (b) A.D. Batcho and W. Leimgruber, Org, Synth. Coll. Vol. VII, 1990, 34. N. Moskalev and M. Makosza, Tetrahedron Lett., 1999, 40, 5395. L.S. Hegedus, Angew. Chem. Int. Ed. Engl., 1988, 27, 1113. (a) C.W. Holzapfel and C. Dwyer, Heterocycles, 1998, 48, 1513. (b) B.C. S6derberg, J.A. Shriver, J. Org. Chem., 1997, 62, 5838. (c) M. Akazome, T.Kondo and Y. Watanabe, J. Org. Chem., 1994, 59, 3375. G.A. Molander and J.A. McKie, J. Org. Chem., 1995, 60, 872. (a) A. Sinhababu and R.T. Borchardt, J. Org. Chem., 1983, 48, 3347. (b) D.H. Lloyd and D.E. Nichols, Tetrahedron Lett., 1983, 24, 4561. (c) D.H. Lloyd and D.E. Nichols, J. Org. Chem., 1986, 51, 4294. J. Harley-Mason, .1 Chem. Soc., 1953, 200 and cited references. C.W. Holzapfel and C. Dwyer, Heterocycles, 1998, 48, 215. (a) G.D. Buckley and C.W. Scaife, .1 Chem. Soc., 1947, 1471. (b) D. Ranganathan, C.B. Rao, S. Ranganathan, A.K..Mehrortra and R. Iyengar, J. Org. Chem., 1980, 45, 1185. S. Takano, M. Moriya, Y. Iwabuchi and K. Ogasawara, Chem. Lett. 1990, 109. B. McKittrick, A. Failli, R.J. Steffan, R.M. Soll, P. Hughes, J. Schmidt, A.A. Asselin, C.C. Shaw, R. Noureldin and G. Gavin, J. Heterocycl. Chem., 1990, 27, 2151. 129

(a) Kirk-Othmer, Encyclopedia of Chemical Technology 4th Ed., Wiley-Interscience: New York, 1992, vol. 2, 586. (b) A. Zoran, 0. Khodzhaev and Y. Sasson, J Chem. Soc. Chem. Commun., 1994, 2239. (c) H. Landscheidt, A. Klausener and H.U. Blank, Eur. Pat. Appl. EP 569,792 to Bayer A.- G., 18 Nov. 1993. (CA 1994, 120: 163724z). (d) ) H. Landsheidt, A. Klausener and H.U. Blank, Ger. Offen, DE 4,023,056 to Bayer A.-G., 23 Jan 1992. (CA 1992, 116: 151302q). (e) Miura, K., T. Kurihara and K. Yanagida, Jpn. Kokai Tokkyo Koho JP 04,149,160 [92,149,160] to Koriyama Kasei Co., Ltd., 22 May 1992. (CA 1992, 117: 170969z) (a) G.R. Lenz, Synthesis, 1987, 489. (b) L. Ripa and A. Hallberg, J. Org. Chem., 1996, 61, 7147. (c) C.R.D. Correia, A.R. de Faria and E.S. Carvalho, Tetrahedron Lett., 1995, 36, 5109. (d) A.R. de Faria, C.R.R. Matos and C.R.D. Correia, Tetrahedron Lett., 1993, 34, 27. (e) M.A. Brodney and A. Pradwa, J. Org. Chem., 1999, 64, 556. J.K. Stille and Y. Becker, J. Org. Chem., 1980, 45, 2139. Y. Becker, A. Eistenstadt and J.K. Stine, J. Org. Chem., 1980, 45, 2145. (a) I. Ojima and Z. Zhang, J. Organomet. Chem., 1991, 417, 253. (b) I. Ojima and Z. Zhang, J. Org. Chem., 1988, 53, 4422. (c) G.D. Cuny and S.L. Buchwald, Synlett, 1995, 519. (d) G.D. Cuny and S.L. Buchwald, J. Am. Chem. Soc., 1993, 115, 2066. (a) Review: T. Shono, Tetrahedron, 1984, 40, 811. (b) T. Shono, Y. Matsumura, K. Tsubata, Y. Sugihara, S. Yamane, T. Kanazawa and T. Aoki, J. Am. Chem. Soc., 1982, 104, 6697. G.A. Kraus and K. Neuenschwander, I Org. Chem., 1981, 46, 4791. J.-C. Guilllemin, J.-M. Denis, M.-C. Lasne and J.-L. Ripoll, Tetrahedon, 1988, 44, 4447. (a) Y. Nomura, K. Ogawa, Y. Takeuchi and S. Tomoda, Chem. Lett. 1977, 693. (b) K. Ogawa, Y. Nomura, Y. Takeuchi and S. Tomuda, J. Chem. Soc. Perkin Trans I, 1982, 3031. E.W. Petrillo Jr. and E.R. Spitzmiller, Tetrahedron Lett., 1979, 51, 4929. F.A.L. Anet and A.J.R. Bourn, J. Am.. Chem. Soc., 1965, 87, 5350. T. Shono, Y. Matsumura and T. Kanazawa, Tetrahedon Lett., 1983, 24, 1259. P. Martin, Hely. Chim. Acta, 1984, 67, 1647. A.M. Lobo, M.M. Marques, S. Prahbakar and H.S. Rzepa, J. Org. Chem., 1987, 52, 2925. C. Pedregal and J. Ezquerra, Tetrahedron Lett., 1994, 35, 2053. M.V. Chiesa, L. Manzoni and C. Scolastico, Synlett, 1996, 441. 130

J. Leonard, S.P. Fearnley, M.R. Finlay, J.A. Knight and G. Wong, J. Chem. Soc. Perkin Trans. I, 1994, 2359. D.F. Oliviera, P.C.M.L. Miranda and C.R.D. Correia, J. Org. Chem., 1999, 64, 6646. W. Marais and C.W. Holzapfel, Synth. Commun., 1998, 28, 3681. (a) E. Taschner, C. Wasielewski and J.F. Biernat, Liebigs Ann. Chem., 1961, 646, 119. (b) E. Taschner, A. Chimiak, B. Bator and T. Sokolowska, Liebigs Ann. Chem., 1961, 646, 142. CHAPTER 4

THE ISOLATION AND CHARACTERISATION OF SECONDARY METABOLITES FROM SELECTED SOUTH AFRICAN MEDICINAL PLANTS

4.1 INTRODUCTION

Plants, unlike animals, accumulate in their cells a considerable range of low molecular weight substances, the majority of them belonging to the so-called secondary constituents which have no distinct function in primary metabolism. These compounds range from the toxic alkaloids and odoriferous essential oils to the cathartic anthraquinones, the coloured anthocyanins and flavones of flower petals which varies immensely both qualitatively and quantitatively in their distribution.' The study of these secondary metabolites has provided much of the impetus for the development of chemotaxonomy which has proved to be useful for unraveling biosynthetic pathways and has given a better understanding of phylogenetic relationships. Thus, as part of an ongoing collaboration with chemotaxonomists of the Department of Botany at RAU we focused on three projects, namely. (i) the Aloes of southern Africa, 2 (ii) the genus Alepidea (Apiaceae) and (iii) Siphonochilus aethiopicus which is commonly known as wild ginger and a member of the family Zingiberaceae. However, for the purpose of this thesis the discussion will be limited to those projects which resulted in the isolation and characterisation of new compounds. Therefore, the reader is referred to the published results 3 regarding Alepidea for the detail on the detection of various known kaurene derivatives in lipophilic extracts from the roots of 16 species of Alepidea utilising gas chromatography linked to mass spectrometry (GC-MS).

4.2 THE CHEMISTRY OF A FEW ALOE SPECIES

4.2.1 BACKGROUND

The Aloes comprise a large genus of over 400 species with additions being made regularly. Except for a few species which have become cosmopolitan through the action of man most Aloe species are confined to Africa and Arabia. The bitter leaf exudates of some Aloe species have been used since ancient times as both drugs and cosmetics, with recent applications including, 132

amongst others, use as additives in shampoos, shaving and skin care creams, as topical medication for the treatment of bums and also as bittering agent in alcoholic beverages 4,s

4.2.2 PLICATALOSIDE, AN 0,0-DIGLYCOSYLATED NAPTHALENE DERIVATIVE FROM ALOE PLICATILIS

Aloe plicatilis (L.) Mill. is a unique and striking tree aloe endemic to the western Cape mountains of South Africa. The stems are dichitomously branched and up to 5 m high in large specimens, with clusters of distichous strap-shaped leaves, each cluster resembling an open fan (hence the common name "fan aloe"). The anomalous morphology of A. plicatilis is reflected in the isolated position of the species in Reynold's classification system, where it was placed in the monotypic section Kumara Medic.6 It is one of only a few Aloe species found in 'fynbos' vegetation, where it grows in high rainfall areas. The unusual habitat may partly account for the exceptional morphology of the species, and it is generally accepted that A. plicatilis is an ancient palaeo-endemic species with unknown affinities. It is superficially similar to the equally anomalous A. haemanthifolia Berger and Marloth, but the latter is a small stemless plant with a totally different combination of morphological and chemical characteristics.

As part of a chemotaxonomic survey of the leaf and root chemistry of the genus Aloe, we found an unknown major compound in the leaf exudate of A. plicatilis, with an interesting UVNIS spectrum quite unlike those of the commonly encountered (with respect to Aloes) chromone-, phenylpyrone-, flavonoid- or anthraquinone derivatives (see Section 4.2.3). 6 A literature search failed to reveal any information on the chemistry of A. plicatilis. In a TLC survey of several Aloe species, Reynolds4a detected two, as yet, unidentified "purple-staining zones" in the leaf exudate of this species. The unusual leaf exudate reflected the taxonomic isolation of A. plicatilis, which suggested that the chemical structure may provide clues about the natural affinities of the species.

Analytical high pressure liquid chromatography (HPLC) and TLC of the exudate of Aloe plicatilis revealed the presence of only one major compound. The presence of a naphthalene type chromophore was suggested by the absorption maxima and fine structure of the UV-VIS spectrum7 of the compound 4.1. Purification of the major component, plicataloside 4.1, was effected by low temperature-flash chromatography over silica gel in water-methanol-ethyl 133 acetate mixture with the pure compound 4.1 crystallising from the eluate as white feathery crystals representing 31 % of the total dried leave exudate. A strong peak at m/z 515, equivalent to protonated 4.1 and little fragmentation, was evident in the fast atom bombardment (FAB) mass spectrum of plicataloside 4.1. The electron impact (EI) mass spectrum of the compound showed no molecular ion but a prominent fragment ion at m/z 190, corresponding to the aglycone moiety resulting from the loss of two glycosyl units.

OR OH

7

6

or 6" CH2OR1 R HO ==10 0 RIO- Tor 3" ' OR ' or 1"

4.1 R1 = H, plicataloside

SCHEME 4.1

The structure of plicataloside 4.1, C23H30013, in (CD3)2S0 was determined from one- dimensional 1 H and 13C NMR spectra and using the two-dimensional NMR techniques COSY, 8 TOCSY,9 ROESY, I° HMQC II and HMBC. I2 The 500MHz I H and 125MHz 13C NMR data of plicataloside 4.1 are given in Table 4.1 and could be readily interpreted in terms of a diglycosylated methylnaphthalene-triol. The I H NMR spectrum exhibits four proton signals in the aromatic region, an AMX system and a singlet. The coupling pattern in the AMX system indicated a tri-substituted aromatic moiety with three neighbouring protons while the singlet must arise from a single proton on an aromatic ring. The compound has one methyl group which gave rise to a characteristic three proton singlet at 81.1= 2.391. The nine exchangeable protons, as determined by exchange with D20, appeared as a broad proton singlet at SH = 9.3, a broad two proton hump at S H = 5.5, four doublets and two triplets (see Table 4.1). The remaining signals from the 14 protons appeared as two one proton doublets at 8H = 4.818 and SH = 4.985 and a number of overlapping signals between 8E, = 3 to 4. The chemical shifts of these resonances 134 clearly indicated two glycoside units.

A COSY experiment revealed the correlation of the hydroxy protons with the corresponding glycoside protons. Especially the resonance positions of the C-6 methylene protons were extablished from the cross-peaks to the two triplets (which disappear on addition of D20) at 8H = 4.305 and 8H = 4.646. The doublets at 4.818 ppm and 4.985 ppm must arise from C-1 glycoside protons as these are the only protons which can appear as doublets. Correlations with these protons gave the resonance positions of the C-2 glycoside protons. The C-2 glycoside protons show no correlation with hydroxy protons. Fast exchange of these C-2 glycoside hydroxy protons probably occur resulting in no observable coupling and the observance as a broad peak at 8H = 5.5. The resonance positions of the C-3 and C-4 glycoside protons have been deduced from their correlation with the remaining hydroxy proton signals.

The resonances belonging to a specific glycoside unit were determined from a TOCSY experiment. Spectral analysis of the 11.1 NMR spectrum of the individual glycoside moieties using the PERCH integrated software package 13 gave the chemical shifts and coupling constants collected in Table 4.1. The vicinal coupling constants between and the chemical shifts of the glycoside protons prove that both units are B-glueosides with C-1-0-glucoside bonds. In the COSY and TOCSY experiments cross peaks were observed between the methyl proton signal and the aromatic proton singlet, indicating that the methyl group is next to and on the same aromatic ring than the proton resonating at 8H = 7.158.

The connectivity of the identified moieties was based on the extensive use of DC NMR data as given in Table 4.1. The proton bearing carbon-13 resonances were assigned with a HMQC experiment. A HMBC experiment was used to assign the quaternary carbon atom signals and to determine the structure of plicataloside 4.1 unambiguously. The methyl protons correlate with C-4 and the carbon-13 resonances at Sc = 133.44 and Sc = 139.63. In the proton coupled DC NMR spectrum the signal at 5c = 133.44 appeared as a quartet with >1.I(C,H) = 6.0 Hz assigning this resonance to a carbon atom bearing a methyl group. The resonance at Sc = 139.63 therefore must arise from another carbon atom (C-2) three bonds away from the methyl protons. In aromatic systems three bond (C,H) couplings (ca 8 Hz) are normally bigger than two bond (C,H) couplings (ca 1-3 Hz),I4 135

1 11 13 C it OH' J Ocb >1J (PPrn) (Hz) (ppm) (Hz) (Hz) 1 9.3 (OH) 114.14 S - 2 - 139.63 Sm - 3 - - 133.44 Sq 6.0 4 7.158 s - 118.79 Dm 160.5 5 7.385 dd 8.29 (5,6) 121.89 Dm 161.9 - 6 7.254 dd 0.89 (5,7) 125.21 D 162.6 - 7 7.199 dd 7.71 (6,7) 109.74 Dd 161.7 9.5 8 154.02 Sd 7.8 9 - - 114.85 Sm - 10 - - 132.76 Sdm - 8.4 11 2.391 s 0.55 (4,11) 17.73 Qd 127.6 5.3

4.818 d 7.58 (1%2') 104.19 Dm 166.3 - 3.309 dd 8.75 (2',3') 74.23 Dm 144.3 - 2'-OH 5.5d 4.84 (3',OH) - - - 3.258 td 8.75 (3',4') 76.35 Dm 143.9 3'-OH 5.019 d 5.28 (4', OH) - 3.159 td 9.62 (4',5') 69.91 Dm 144.9 - 4'-OH 4.910 d 2.07 (5', 6'a) - 3.087 dddd 5.70 (5',6'b) 77.08 Dm 141.2 6'(a) 3.607 ddd -11.68 (6'a,6'b) 60.93 Tm 139.9 - 6'(b) 3.426 dt 5.91 (6'a,OH) 6'-OH 4.305 t 5.91 (6'b,OH) - - 4.985 d 7.69 (1",2") 102.69 Dm 165.0 - 3.388 dd 8.72 (2",3") 73.61 Dm 142.8 2"-OH 5.5d 5.09 (3",OH) - - 3.330 td 8.72 (3",4") 76.20 Dm 145.6 - 3"-OH 5.107 d 5.47 (4",OH) 3.206 td 9.59 (4",5") 69.85 Dm 144.9 - 4"-OH 5.061 d 1.93 (5",6"a) 3.407 dddd 6.03 (5",6"b) 77.60 Dm 140.9 - 6"(a) 3.756 ddd -11.92 (6"a,6"b) 60.74 Tm 139.6 6"(b) 3.514 dt 5.82 (6"a,OH) - 6"-OH 4.646 t 5.82 (6"b, OH) - - -

TABLE 4.1: 11-1 (500.13 MHz) AND 13C (125.76 MHz) NMR DATA FOR PLICATALOSIDE 4.1 IN DMS0-4 AS SOLVENT

With a mixing delay of 50ms in the HMBC experiment the intensities of cross peaks resulting from vicinal coupled carbon-13 and proton nuclei are more intense than those between nuclei 136 with smaller couplings." Keeping this criterion in mind we observed that H-4 correlates with C- 2, C-5 and the signal at Sc = 114.85. The correlation with C-5 proves that 4.1 has a naphthalene moiety and, therefore, the resonance at Sc = 114.89 must be from the ring juncture carbon atom, C-9. Cross peaks were observed from H-6 to two quaternary carbon signals at Sc = 154.02 and Sc = 132.76 which must arise from carbon atoms (C-8 and C-10) three bonds from H-6. Intense cross peaks were noted between H-7 and C-9 and C-5 and a correlation with a lower intensity with the resonance at Sc = 154.02 thereby assigning this signal to C-8 and the resonance at 6c = 132.76 to C-10. The remaining quaternary carbon signal at 8 c = 144.14 is assigned to C-1. The chemical shift of C-8 is characteristic for an oxygen bearing carbon atom while the upfield shifts of C-1 and C-2 can be attributed to the influence of the two ortho oxygen atoms.

With all the quaternary carbon resonances assigned the substitution positions of the two glycoside units follow directly from the HMBC results. Correlations were observed between H-1' and C-2 and between H-1 1. and C-8. The remaining hydroxy group must be connected to C-1. Further evidence for the structure and an indication of the stereochemistry were obtained from a ROESY experiment e.g. correlation between H-4 and H-6 substantiated the existence of the naphthalene moiety. A correlation between H-1" and H-7 shows that the movement of the glucoside unit on C-8 is restricted. Correlations between H-1' and H-5', H-1' and H-3', H-5' and the C-4' hydroxy proton, H-1" and H-5", H-1" and H-3" and H-5" and the C-4" hydroxy proton provided additional evidence that both of the glycosides are glucoside units.

Acetylation of plicataloside 4.1 (Scheme 4.2) under a variety of conditions furnished two acetylated products, 4.2 and 4.3, which were isolated by reverse phase HPLC. The 'H and "C NMR data of these two products are collected in Table 4.2. It showed that 4.2 and 4.3 resulted from the acetylation of eight and nine hydroxy groups, respectively. 137

4.2 4.3 1 H "C 1 11 kb J I J >1.7 8141 thia J (Plmn) (Hz) (1)Pm) (Hz) (Hz) (PPIn) (Hz) 1(OH) 8.233 s (OH) - 144.12 S - - - 2 - - 139.29 Sm - - - - 3 - 135.04 Sq - 6.1 - - 4 7.080 s - 119.29 Dm 160.5 - 7.460 - 5 7.371 dd 7.91 (5,6) 123.43 Dm 159.6 - 7.421 8.00 (5,6) 6 7.184 t 0.55 (5,7) 124.49 D 160.5 - 7.268 7.70 (6.7) 7 6.828 dd 7.66 (6,7) 107.76 Dd 160.2 8.3 6.96 - 8 - - 152.72 Sd - 11.0 - - 9 - 114.78 Sm - - - - 10 - - 133.51 Sd - 8.6 - - 11 2.362 s 0.56 (4,11) 17.58 Qd 128.4 5.0 2.404 - 5.264d 7.60 (1',2') 101.12 Dm 168.2 - 5.32 - 5.330 m 8.01 (21,3') 71.81 Dm 153.8 6.1 - - 5.304 m 9.14 (3',4') 72.98 Dm 150.0 - 5.12 - 5.146 tm 10.04 (41,5) 68.75 Dm 150.5 - - - 3.606 ddd 4.63 (5',6'a) 71.53 Dm 144.6 - 3.634 - 3.188 dd 2.60 (5',6'b) 61.77 Tm 147.9 - 4.200 - 4.007 dd -12.20 (6'a,6'6) - - - 4.007 - 5.394 d 7.60 (1",2") 98.93 Dm 164.6 - 5.32 5.402 m 7.46 (2",3") 70.91 Dd 154.3 5.6 - 5.357 m 8.77 (3",4") 72.38 Dm 154.0 - 5.12 5.228 dd 9.96 (4',5') 68.20 Dm 153.6 - - 3.923 ddd 4.66 (5",6"a) 72.35 Dm 144.6 - 3.823 4.246 dd 3.04 (5",6"b) 61.48 Tm 148.5 - 4.244 4.226 dd -12.43 (6"a,6"b) - - - 4.181 OCOCH3 2.165 s - 20.80 Q - - 2.099 s - 2.056 s - 20.62 Q - 2.025 s - 2.044 s - 20.58 Q - 2.010 s - 2.038 s - 20.54 Q (3) - - 2.003 s - 2.015 s (2) - 20.42 Q (2) - - 1.983 s (2) - 1.994 s - - - - 1.950 s - 1.958 s - - - 1.930 s ------2.424 s OCOCH3 - - 170.53 s - - - - 170.41 s ------170.15s - - - - 170.06 s 170.02 s 169.50 s - - 169.31 s - - - -

TABLE 4.2: (500.13 MHz) AND "C (125.76 MHz) NMR DATA FOR ACETYLATED PRODUCTS 4.2 AND 43 IN CDC1 3 AS SOLVENT 138

OR 0R2 OR 7

6 Me

6' or 6" CH2OR 1 R 1 0 R = 0 R 1 0 3' or 3" \ OR 1' or 1"

4.2 R1 = Ac, R2 = H 4.3 R1 = R2 = Ac

SCHEME 4.2: ACETYLATED PRODUCTS 4.2 AND 4.3

Interestingly 4.2 exhibited a sharp one proton singlet at 8 n = 8.223, which disappear on addition of D20, indicating the presence of a phenolic hydroxy group. The 'H NMR spectrum of compound 4.3 exhibited broad signals, possibly attributed to restricted motion caused by steric congestion of the bulky substituents.

Several attempts to effect acid-catalysed hydrolysis of plicataloside 4.1 in an inert atmosphere resulted in rapid blackening of the solution. Chromatography of the reaction mixture failed to furnish any of the expected aglycone. This may be due to the known oxidation sensitivity of naphthalene polyols. 15,16

Structure analysis of the aglycone moieties of the secondary metabolites that normally occur in the Aloe leaf exudate apparently derive from biogenetic intermediate C-14 or C-16 polyketides, formed by the condensation of acetic acid units, modified by cyclisation (intramolecular aldol reactions), redox reactions or a loss of a terminal carbon atom as carbon dioxide. However, the aglycone of plicataloside appears to be derived from a C-12 polyketide which suffered the loss of a terminal carboxylic acid group. This biogenetic difference is in line with the phylogenetic isolation of A. plicatilis. 139

4.2.3 CHROMONE AND ALOIN DERIVATIVES FROM, ALOE AFRICANA, A. SPECIOSA AND A. BROOMIL

During the earlier mentioned chemotaxonomic screening of Aloe leaf exudates, some unknown major compounds were also detected by HPLC in three other South African endemic aloes. The species were Aloe africana Mill. from the section Pachydendron, A. speciosa Bak. from the monotypic series Principales Berger and A. broomii Schonl. from the artificial series Longistylae Berger. ° The infrageneric positions of these species were based on superficial characterisations and it was felt that the leaf exudate chemistry may provide clues about natural affinities. Since the composition of Aloe leaf exudates have been investigated extensively, 4b the identified compounds can generally be classified into two main groups, namely chromone derivatives and anthraquinone derivatives. In some cases both types are present and in other cases only one. Furthermore, many of the major constituents such as aloesin 4.4 and the steroisomers, aloin A and B 4.5 occur in chemotaxonomically distinct species. We have started a detailed screening of Aloe using a new approach, viz. HPLC coupled to electrospray mass spectrometry (ES-MS) with the aim of identifying unique constituents which may be used as markers with possible chemotaxonomic significance.

4.2.3.1 THE IDENTIFICATION OF A. AFRICANA CONSTITUENTS

HPLC analysis and TLC of the exudate of A. africana revealed the presence of aloesin 4.4, aloin A and B 4.5 (see Scheme 4.6), aloinoside A and B and an unidentified compound later characterised as 4.6 (Scheme 4.3).4 The purification of 4.6 was effected by low temperature flash chromatography over silica gel in a methanol-chloroform mixture. The compound 4.6 exhibited absorption maxima in its UV-VIS spectrum characteristic of C-glucosylated 5- methylchromones. I7 The IH and I3C NMR (see Table 4.3) spectra have characteristics similar to those of aloeresin A 4.7 18.19 except for (i) the simple AA'BB' pattern for the protons of the coumaric ester were replaced by an ABX pattern for 3 protons thus indicating the presence of a ferulic ester and (ii) the spectrum contains a 3-proton singlet of an OMe-group. This compound

4.6 corresponds to the 7-methyl ether of 2"-feruloylaloesin isolated by Makino et 01.20 140

R 1 0

C H3 0

; R2 = H or A

0 A

RI R2 R3 R4 4.4 aloesin H H 4.6 new compound Me A OMe OH (A. Africana) feruloyl

4.7 aloeresin A H A H OH coumaroyl

SCHEME 4.3: A NEW CONSTITUENT (ISOLATED FROM A. AFRICANA) AND OTHER RELATED CHROMONES

The difference in chemical shifts observed between the aromatic carbons of 4.6 and aloesin 4.4 was consistent with the methylation of the 7-hydroxyl group. The ester group of 4.6 is attached to C-2' of the carbohydrate moiety. The corresponding H-2' resonates at 814 5.47. The assignment of all I Hand 13C signals of this and all other compounds described herein were confirmed using COSY,8 TOCSY,9 ROESY, I9 HMQCI I and HMBC.I2 141

C 4.6a C 4.6a 2 160.6 3' 75.8 3 112.5 4' 70.6 4 178.6 81.9 4a 115.6 61.5 5 141.8 1" 165.4 6 111.5 2" 114.2 7 159.7 3" 144.6 8 110.8 125.5 la 157.4 111.1 9 47.9 6" 147.9 10 202.1 7" 149.2 11 29.6 115.5 5-Me 22.7 122.9 70.4 7-OMe 55.5b 72.2 6"-OMe 56.34

Solvent: DMSO-d6

Signals interchangeable TABLE 4.3: 13C NMR (125.76 MHz) CHE- MICAL SHIFTS OF 4.6 (A. AFRICANA)

Strong peaks with m/e 583 *if and m/e 407 [M-177] - were observed in the negative mode ES-MS spectrum. The loss of a fragment with mass 177 could be interpreted as the loss of a COCH=CHC6H3(OCH3)(OH) fragment from a ferulic acid derivative. This fragmentation is followed by the loss of water from m/e 407. The spectrum showed no peaks corresponding to the subsequent loss of the carbohydrate moiety after the loss of the acyl group. In contrast, the negative ES-MS of aloesin 4.4 showed a strong [M-1] - ion and a base peak at m/e 272 corresponding to the loss of the carbohydrate moiety. This information allows formulation of compound 4.6 as (E)-2-acetony1-8-(2'-0-feruloy1)-0-0-glucopyranosyl-7-methoxy-5-methyl- chromone.

4.2.3.2 THE IDENTIFICATION OF A. SPECIOSA CONSTITUENTS

In addition to homonataloin (Scheme 4.4) 4b HPLC analysis of the leaf exudate of A. speciosa indicated the presence of a previously unidentified compound 4.8. 142

OMe 0 OH HO

*SO Glc H Homonataloin

SCHEME 4.4

The unknown compound 4.8 (Scheme 4.5) was also isolated by low temperature flash chromatography. The UV-VIS spectrum of 4.8 was closely related to that of aloeresin C 4.9 while the negative mode ES-MS showed a strong [M-11 peak at m/e 685. Other strong peaks in the high mass range resulted from the consecutive loss of a fragment of mass 147 (equivalent to COCH=CHC6H4OH) and water.

CH3 0

; R2 and R3 = H or A

0 A

R5 R4

Ri R2 R3 R4 R5 4.8 A. speciosa A A H OIj (new compound) coumaroyl 4.9 Aloeresin C [3-o-Glc A H H OH coumaroyl SCHEME 4.5 143

The 'H and 13C NMR data (see Table 4.4) of 4.8 were in close agreement with those reported for the chromone derivatives of the aloeresin series. 18 The presence of 4 sets of doublets in the aromatic region of the 'H spectrum suggested the presence of two coumaric acid ester residues. The 13C NMR showed the expected two sets of signals for two coumaroyl residues (See Table 4.4). The remaining ambiguity, i.e. the positions of attachment of the coumaroyl groups, was resolved on the basis of the low field positions of H-6'a and H-6 11) at SH 4.52 and 4.06 respectively, and H-2' at 8H 5.49 and confirmed using COSY, 8 TOCSY,9 ROESY, 19 HMQC 11 and HMBC. 12

C 4.8a C 4.8a 2 160.1 1" 165.4 3 112.5 2" 114.0 4 178.4 144.4 4a 114.7 125.1 5 140.9 5" 130.1 6 114.8 6" 115.7 7 159.8 7" 158.3 8 108.7 115.7 la 159.8 130.1 9 48.1 166.7 10 202.3 114.0 11 29.2 144.9 5-Me 22.5 125.0 70.7 51" & 9"' 130.3 72.0 6"' & 8"' 115.7 75.6 159.1 70.4 78.3 64.4

a: Solvent: DMSO-d6

TABLE 4.4: 13C NMR (125.76 MHz) DATA OF 4.8 (A. SPECIOSA)

4.2.3.3 IDENTIFICATION OF A. BROOMII CONSTITUENTS

Preparative HPLC (reverse phase) was used to isolate the four compounds 4.10, 4.11, 4.12 and 4.13 from the leaf exudate of A. broomii. The major component was identified as 5- hydroxyaloin A 4.13, previously described as a constituent of this Aloe species.21 However, the evidence for the structure was incomplete, particularly with regard to NMR data. 22-25 The 144 structure was therefore confirmed on the basis of a complete analysis of 1H and 13C NMR data (see Table 4.5) of this molecule 4.13 including the use of COSY, 8 HMQC 11 and HMBC.' 2

OH 0 OH

R20 HO R = 0 HO

4.5 aloin R1 = R = H 4.10 A. broomii R1 = OH, R = caffeoyl

4.13 5-hydroxyaloin RI = OH, R = H

SCHEME 4.6: ALOIN DERIVATIVES 4.10 AND 4.13 FROM A. BROOMII

C 4.10 4.13 C 4.10 4.13 1 161.3 161.3 2' 78.9 77.2 1 a 116.5 116.8 3' 71.2 71.4 2 113.0 112.8 4" 71.4 70.1 3 151.3 151.0 5' 77.1 77.8 4 116.1 115.8 6' 62.6 63.6 4a 144.5 144.9 1" 166.9 5 145.9 145.7 114.2 5a 124.9 124.9 144.6 6 124.1 123.8 4" 126.4 7 116.4 116.2 5" 114.8 8 155.4 155.3 147.7 8a 117.6 117.8 145.1 9 193.8 193.8 115.6 11 63.4 63.3 121.3 1' 83.8 83.8

TABLE 4.5: 13C NMR (125.76 MHz) CHEMICAL SHIFTS IN 5 % DMS0-4 AND CDC13 OF 4.10 AND 4.13 (A. BROOMII) 145

It is of interest to note that the negative mode ES-MS spectrum showed a strong [M-lf parent ion at m/e 433 while the base peak at m/e 271 resulted from the loss of the sugar moiety, C6H1105, from the parent molecule. A moderately abundant ion at m/e 313 probably results from the loss of a C4H504-fragment from the [M-11 - parent ion. The loss of this fragment from the [M-1 parent ions of aloin A and B 4.5 and homonataloin (Scheme 4.4) give rise to the base peaks of the negative mode ES-MS spectra of these molecules. In the case of aloin A and B 4.5 this fragmentation is tentatively rationalised in Scheme 4.7.

OH 0 OH OH 0 OH

(-C41-1804) oe CH2OH 0 OH HOCH2 OH

4.4 R = H m/z = 417 m/z = 297 (base)

SCHEME 4.7

The FAB-spectrum of 4.13 showed the corresponding positive ions at m/z 435 and 272, respectively, together with a fragment with m/z 255, probably resulting from the loss of OH.

The remaining three compounds of A. broomii 4.10, 4.11, and 4.12 have not been described previously. The FAB-spectrum of compound 4.10 showed a strong [M+11 + parent ion at m/z 597 corresponding to the caffeoyl ester of 5-hydroxyaloin. Fragment ions at m/z 433 (weak) and 271 (strong) corresponded to the loss of caffeoyl and caffeoyl plus carbohydrate moieties from the parent molecule. A further fragment ion with m/z 314 corresponds to the loss of the ester group together with a four carbon fragment (vida supra) of the carbohydrate moiety. The negative ES-MS spectrum showed a strong [M-11" parent ion at m/e 595 and a significant fragment ion at m/e 271. The deductions on the basis of the mass spectra were confirmed by the and "C NMR (see Table 4.5) spectra, e.g. the position of attachment of the ester moiety followed from the appearance of the low field positions of H-6'a and H-6' at 5E, 4.20 and SH 3.78, respectively as well as extensive decoupling experiments. 146

CH3 0

HO HO R = 0 HO OR

0

R5 R4

RI R2 R4 R5 4.11} Me CH 2COCH3 OH OH A. broomii 4.12 Me CH 2COCH3 H H 4.14 rabaichromone Me CH2CH(OH)CH3 OH OH

SCHEME 4.8

The negative ES-MS spectrum of the second new compound 4.11 from A. broomii showed a strong [M-If parent ion at m/e 569 which was tentatively interpreted as that of a caffeoyl ester of aloesin 4.4. The compound was formulated as (E)-2-acetony1-8-(2 1-0-caffeoy1)-13-D- glucopyranosy1-7-methoxy-5-methylchromone 4.11 on the basis of its 1H and 13C NMR spectra (see Table 4.6). The corresponding dihydrocompound, rabaichromone 4.14 is a known constituent of A. rabaiensis.26 147

C 4.11 4.12 C 4.11 4.12 2 160.6 159.8 3' 75.7 76.5 3 112.5 112.9 4' 70.5 70.8 4 178.5 178.6 81.8 80.1 4a 115.6 116.2 61.5 62.1 5 141.7 143.1 1" 165.3 163.9 6 111.5 111.1 2" 113.7 117.4 7 159.6 155.5 3" 144.6 144.4 8 110.8 110.3 125.3 134.0 1 a 157.4 156.0 114.6 127.7 9 47.9 48.7 6" 145.5 128.7 10 202.1 201.8 7" 148.3 130.1 11 29.5 29.4 115.7 128.7 5-Me 22.7 23.2 121.1 127.7 70.4 71.0 7-OMe 56.4 56.0 72.2 72.3

TABLE 4.6: 13C NMR (125.76 MHz) DATA OF 4.11 (IN 5 % DMSO-d6) AND 4.12 (IN 5 % DMS046 AND CDC13) FROM A. BROOMH

The negative ES-MS of 7 showed a strong [M-If parent ion at m/z 537 and fragments at m/z 407 and 389 corresponding to the consecutive loss of the cinnamoyl group and water. From the low field signal at 8H 5.42 it followed that the ester group is attached to C-2 of the carbohydrate moiety.

The molecular mass of this and all other compounds described herein was confirmed by positive mode ES-MS. However, these spectra showed more extensive and complex fragmentation patterns than the corresponding negative mode ES-MS.

The types of compound isolated in this study can be regarded as typical of the genus Aloe and may prove to have some chemotaxonomical significance. Evidence is provided that ES-MS is a valuable tool for the identification and structural elucidation of Aloe exudate components.

4 3 THE STUDY OF WILD GINGER

Siphonochilus aethiopicus, commonly known as wild ginger, is a member of the family Zingiberaceae and is restricted to southern Africa, including South Africa, Zimbabwe, Malawi and Zambia. This species is the only indigenous member of the family in South Africa and has a unique and distinctive morphology. 29 The rhizomes and roots are much used for colds, coughs 148 and influenza, but also for hysteria, pain and several other traditional and cultural practises. 22-29 Pressure on wild populations has led to local extinctions, notably KwaZulu-Natal. 3°

A chemotaxonomic survey of plants and plant parts from various provenances invariably yielded substantial quantities of a pale yellow semi-crystalline mixture. This mixture, obtained by steam distillation of fleshy crushed roots of several plants of S. aethiopicus, contained a low content of monoterpenenoids, but substantial amounts (up to 0.2% of wet weight of plant) of an unknown major constituent, 4.15 accompanied by a minor compound 4.16 Extraction of the roots with cold ethanol for five days followed by evaporation of the solvent furnished a residue which, again, contained mainly 4.15.

4.15 R = H 4.16 R=OH 4.17 R = OCOCH3

SCHEME 4.9

The purification of 4.15 and 4.16 was effected by flash chromatography over silica gel deactivated with 2% triethylamine in toluene. The new furanoterpenoid 4.15 was assigned the composition, CI5H1802, on the basis of mass spectrometry and accurate mass determinations of a molecular ion at m/z 230. The NMR data (see Table 4.7), especially the correlations observed in the COSY8 and HMBC 12 experiments, gave unambiguous proof of the structure. The fifteen signals in the 13C NMR spectrum, one a ketone carbon resonating at 5 204.0 as well as five other low field signals suggested that 4.15 is an oxygenated sesquiterpene. The 1H NMR spectrum exhibited, amongst others, a furan proton signal at 8 7.02, two sets of doublets at 5 5.91 and 6.66, respectively, one AB system and three methyl signals, of which one is secondary. Irradiation of the methyl signal at 5 1.21 allowed the assignment of H-5 at 8 2.41. This proton represents the X part of an ABX system. The proton-proton coupling constants associated with that system allows assignment of the relative stereochemistry of 4.15. Thus, proton H-4a exhibiting a 149 threefold doublet at 8 1.81 with two 10 Hz proton-proton couplings indicated that H-4a is trans to H-5 and that the structure is fairly rigid. Confirmation that the methyl substitution occurred on position 3 and not C-2 is given by the observed directly bonded (C,H) coupling of 198.9 Hz measured for C-2 which complies with reported data .31 '32 Therefore, the compound 4.15 was formulated as (4aa,51:1,8a(3)-3,5,8a-trimethyl-4,4a,9-tetrahydronaphtho[2,3b]-furan-8(5H)-one.

4.15 4.16 4.17 H 5H J (Hz) 8„ J (Hz) 5H J (Hz) 2 7.02 br m - - - - 4A 2.68 ddd 15.7; 5.4: 1.6 2.80 dd 13.5; 3.6 2.87 dd 13.7; 3.7 4B 2.12 dddd 15.7; 10.8; 3.0; 1.4 2.42 ddq 132; 13.2; 1.5 2.21 ddq 13.7; 12.7; 1.4 4a 1.81 ddd 10.8; 10.2; 5.4 1.55 ddd 13.2; 9.9; 3.6 1.57 ddd 12.7; 9.8; 3.7 5 2.40 dqdd 10.2; 7.1; 2.7; 2.1 2.47 dqdd 10.2; 7.2; 2.4; 2.4 2.44 dqdd 9.8; 7.2; 2.8; 2.0 6 6.66 dd 10.1; 2.1 6.63 dd 10.2; 2.1 6.62 ddd 10.1; 2.0; 0.4 7 5.91 dd 10.1; 2.7 5.84 dd 10.2; 3.0 5.84 dd 10.1; 2.8 9A 2.73 dd 16.7; 1.4 1.63 d 14.4 3.04 d 14.9 9B 2.64 br d 16.7 2.64 d 14.4 1.53 d 14.9 2-OH - - 4.51 br s - - - 2-000OCH3 - - - - 2.01 s - 3-Me 1.90 d 1.3 (H-2) 1.78 d 1.5 (H-4B) 1.83 d 1.4 (H-4B) 5-Me 121 d 7.1 1.21 d 72 122 d 7.2 8a-Me 1.02 s - 1.33 s 124 s -

TABLE 4.7: 1 1-1 (500.13 MHz) NMR DATA FOR 4.15, 4.16 AND 4.17 IN CDC13 AS SOLVENT

A strong peak at m/z 247, equivalent to protonated 4.16 and little fragmentation, was evident in the FAB-mass spectrum of 4.16. The structure of 4.16, C15111803, followed from the 11-1 NMR data as shown in Table 4.7. While most signals were similar to those of 4.15, the biggest differences in the 'H spectrum of 4.16 were the disappearance of the furan signal, the presence of an additional broad singlet at 8 4.51, the upfield shift of H-9 to 8 1.63 as well as a slight downfield shift of the 8a-methyl group. A big downfield shift of C-2 from 137.6 to 172.2 ppm in the 13C NMR spectrum (Table 4.8) was also observed. Hydroxylation in the 2 position has a big influence on the electron distribution in the furane ring as shown by the change in the chemical shifts of C-3 (-15.5), C-3a (+7.7) and C-9a (+11.7) when compared with the corresponding values in 4.15. It effects C-9 (+10.8 ppm and J9A,9B = 14.4 Hz vs 16.9 Hz in 4.15) and it also has a subtle influence on the preferred conformation of 4.16, because the substantial long range proton-proton coupling constants observed between C-4 and C-9 protons in 4.15 are no longer detected. A change in the conformation of the cyclohexenone ring due to the buttressing effect of substituents may result in orientation of the C-8 carbonyl group parallel with the C-9 equatorial proton, thus explaining the big upfield shift of the C-9 equatorial proton in 4.16. 150

4.15 4.16 4.17 C 8c 8c 8c 2 137.5 Dq 172.2 167.9 Sm 3 119.0 Sm 103.5 103.7 Sm 3a 114.6 Sm 122.3 124.1 Sm 4 22.5 T 24.1 24.1 Tm 4a 45.0 Dm 49.9 49.7 Dm 5 34.2 Dm 33.8 33.7 Dm 6 154.2 Dqn 153.7 153.1 Dm 7 126.6 Dd 126.1 126 Dd 8 204.0 Sm 203.0 201.6 Sm 8a 44.9 Sm 44.8 44.5 Sm 9 31.9 Tm 43.6 42.8 DDm 9a 149.3 Sm 158.5 155.6 Sm 3-Me 8.1 Q 8.3 8.3 Q 5-Me 18.7 Qm 18.0 18.0 Qm 8a-Me 16.6 Qm 16.6 16.3 Qm COOCH3 - - - 170.8 Sm COOCH3 _ _ - 21.9 Q

TABLE 4.8: 13C (125.76 MHz) NMR DATA FOR 4.15, 4.16 AND 4.17 IN CDC13 AS SOLVENT

Reaction of 4.16 with acetic anhydride and triethylamine resulted in O-acetylation on the 2- position and are in complete agreement with the data obtained from the 1 H and 13C NMR and COSY experiments. Once again, similar changes in chemical shifts were observed in the 13C NMR spectrum of 4.17. This information allows formulation of the new compound 4.16 as the 2-hydroxy derivative of 4.15, namely (4aa,5f3,84)-2-hydroxy-3,5,8a-trimethy1-4,4a,9- tetrahydronaphtho [2,3 b]furan-8(5H)-one.

The two new compounds 4.15 and 4.16 are members of the eudesmane family of sesquiterpenoids. These classes of compounds are produced by other plants and marine organisms and display a significant variance in stereochemistry33 -35 e.g. tuberiferine• 36 4.18 from Sonchus Tuberifer Svent and tubipofuran37 4.19 from Tubipora musica Linnaeus, respectively, as illustrated in Scheme 4.10. 151

C H3

4.18 tuberiferine 4.19 tubipofuran

SCHEME 4.10: OTHER EXAMPLES OF EUDESMANE TYPE NATURAL PRODUCTS

Interestingly, these eudesmane type compounds such as 4.15 and 4.16 have not been found to be present in ginger oil from Zingiber officinale Roscoe (Zingiberaceae). The latter is well studied38 and reported constituents include, amongst others, monoterpene hydrocarbons such as camphene and fl-phellandrene, oxygenated monoterpenes, sesquiterpene hydrocarbons (SQHC) and oxygenated sesquiterpenes. The most important group of compounds, quantitatively, are the sesquiterpene hydrocarbons with a-zingiberene (Scheme 4.11) generally occurring as 10 - 30 % of the oil while the oxygenated sesquiterpenes are normally present in small amounts (0.4 - 2 %). The obvious chemical and thus, biogenetic differences in the structure of the terpenoid constituents of S. aethiopicus and Zingiber officinale Roscoe is probably a reflection of morphological differences.

a-zingiberene

SCHEME 4.11

4.4 SUMMARY AND CONCLUSIONS

This chemotaxonomic survey resulted not only in the isolation of several new compounds, one of them the unique compound plicatoloside 4.1, but it also successfully demonstrated the 152 application of HPLC coupled to ES-MS in identifying such compounds. In addition, the major constituent, 4.16 of the wild ginger, S. aethiopicus, was also isolated and characterised.

4.5 REFERENCES

J.B. Harborne in Chemotaxonomy and Serotaxonomy; J.G. Hawkes, Ed.; Academic Press: London, 1968; Chapter 16. (a) P.L. Wessels, C.W. Holzapfel, B.-E. van Wyk and W. Marais, Phytochem., 1996, 41, 1547. (b) P.L. Wessels, C.W. Holzapfel, B.-E. van Wyk, W. Marais and M. Portwig, Phytochem., 1997, 45, 97. C.W. Holzapfel, B.-E. van Wyk, A. de Castro,W. Marais and M. Herbst, Biochem. Syst. Ecot, 1995, 23, 799. (a) T. Reynolds, Bot. .1 Linn. Soc., 1985, 90, 179. (b) T. Reynolds, Bot. J. Linn. Soc., 1985, 90, 157 and cited references. E. Dagne, Bull. Chem. Soc. Ethiop., 1996, 10, 89 and cited references. G.W. Reynolds, in The Aloes of Southern Africa; Aloes of South Africa Book Fund: Johannesburg, 1950,pp. 162, 422, 456, 502. H.H. Jaffe and M. Orchin in Theory and Applications of Ultra Violet Spectroscopy; John Wiley and Sons, Inc.: New York, 1965, 306. W.P. Aue, E. Bartholdi, and R.R. Ernst, J Chem. Phys., 1976, 64, 2229. L. Braunschweiler and R.R. Ernst, J. Magn. Reson., 1983, 53, 521. A. Bax and D.G. Davis, J. Magn. Reson., 1985, 63, 207. A. Bax, R.H. Griffey and B.L Hawkins,. J. Am. Chem. Soc., 1983, 105, 7188. A. Bax and M.F Summers, J. Am. Chem. Soc., 1986, 108, 2093. R. Laatikainen and M. Niementz, PERCH, An integrated software for analysis of NMR spectra on PC Version 1-95, Perch Project University of Kuopio, Kuopio. L. Ernst, V. Wray, V.A. Chetkov and M. Sergeyef, J Magn. Reson., 1977, 25, 123. G. Bringman, Angew. Chem. Int. Ed. English, 1982, 21, 200. S. Mongkolsuk and C. Sdarwonvivat, J. Chem. Soc., 1965, 1533. K. Sen and P. Bagchi, J. Org. Chem., 1959, 24, 316. G. Speranza, P. Gramatica, G. Dada and P Manitto, Phytochem., 1985, 24, 1571. G. Speranza, G. Dada, L. Lunazzi, P. Gramatica and P. Manitto, Phytochem., 1986, 25, 2219. 153

K. Makino, A. Yagi and I. Nishioka, Chem. Pharm. Bull., 1985, 22, 1565. H.W. Rauwald and A. Beil, Z. Naturforsch., 1993, 48c, 1. H.W. Rauwald, Pharm. Weekbl. Sci. Ed. Pharmaceutisch, 1987, 9, 215. H.W. Rauwald and A. Beil, J. Chromatogr., 1993, 639, 359. H.-D. Hiiltje and K. Stahl, Arch. Pharm. (Weinheim), 1991, 324, 859. H.W. Rauwald, Pharm. Ztg. Wiss., 1990, 3/135, 169. J. M. Conner, A.I. Gray, T. Reynolds and P. G. Watermann, Phytochem., 1989, 28, 3551. J.M. Watt, and M.G. Breyer-Brandwijk, The Medicinal and Poisonous Plants of Southern and Eastern Africa; 2nd edn Livingstone: London, 1962, 1063. A. Hutchings, Zulu Medicinal Plants; Natal University Press, 1962, 64. B.-E. Van Wyk, B. Van Oudtshoorn and N. Gericke, Medicinal Plants of South Africa; Briza Publications: Pretoria, 1997, p. 240 -241. A.B. Cunningham, An investigation of the herbal medicine trade in Natal/Kwazulu. Investigational report no 29, Institute of Natural Resources, University of Natal, 1988. A. Ojida, F. Tanoue and K. Kanematsu, J. Org. Chem., 1994, 59, 5970. E. Pretsch, T.C. Clerc, J. Seibl and W.Simon Tables of Spectral Data for Structure Determination of Organic Compounds; Springer-Verlag: Berlin, 1989. R.B. Miller and E.S. Behare,. J Am. Chem. Soc., 1974, 96, 8102. M.S. Al-Said, 5.1 Khalifa, F.S. El-Feraly and C.D. Hufford, Phytochem., 1989, 28, 107. K. Yamakawa, K. Nishitani and T. Tominaga, Phytochem., 1975, 33, 2829. J.B. Barrera, J.L. Breton and A.G. Gonzalez, Tetrahedron Lett., 1967, 36, 375. K. Iguchi, K. Mori, M. Suzuki, H. Takahashi and Y. Yamada, Chem. Lett., 1986, 1789. T.A. Van Beek in Modern Methods of Plant Analysis New Series Vol. 12 - Esssential Oils and Waxes; H.F. Linskens and J.F. Jackson, Eds., Springer-Verlag: Berlin, Heidelberg and New York, 1991, 79. CHAPTER 5

EXPERIMENTAL DATA

5.1 GENERAL

All reactions were performed under positive nitrogen or argon pressure with dry solvents, in flamed-out glass apparatus, unless otherwise specified. The carbonylation reactions at high pressures (2 - 10 bar) were carried out in thick-walled pressure vessels of glass or stainless steel. The necessary safety precautions were taken.

All reagents were obtained from commercial suppliers and used without purification unless otherwise indicated. Solvents were purified and distilled prior to use, according to standard procedures. 1'2 Palladium chloride and palladium acetate were purchased from the relevant suppliers in analytically pure form. Tetrakis(triphenylphosphine)palladium(0) and dichlorobis- (triphenylphophine)palladium(II) were prepared as per literature procedures 3 and stored under argon at -20°C.

Analytical thin-layer chromatography (TLC) was performed on Merck silica gel (60 F254) plates precoated (0.25 mm) with fluorescent indicator. Thin-layer chromatograms were developed with solvents as indicated for column chromatography. Detection of the developed plate was accomplished using an ultraviolet lamp (254 nm and/or 365 nm) followed by spraying with chromic acid and heating over an open flame. Indoles were detected using Van Urk's spray reagent (4-dimethylaminobenzaldehyde-HC1) followed by heating over an open flame. "Chromatography" refers to column chromatography on Merck Kieselgel 60 (70-230 mesh) and "flash chromatography" 4 to column chromatography on Merck Kieselgel 60 (230-400 mesh) using v/v mixtures of the indicated eluents under a positive nitrogen pressure.

NMR spectra ( 1H (300.060 MHz) and 13C (75.459 MHz) were recorded on a Varian Gemini-300 spectrometer. In the case of the structure elucidation of the natural products NMR spectra were recorded on a Bruker AMR-500 NMR spectrometer operating at 500.13 MHz and 125.76 MHz for 1 H and 13C, respectively. Unless otherwise specified, all NMR spectra were obtained as 155

solutions in deuterochloroform (CDC13) and are reported in parts per million (ppm). In all cases where the detection of the nitro-bearing carbon was difficult, the addition of 10 mg of Cr(acac)3 to the NMR tube resulted in a sufficient increase in signal intensity to allow assignment of the peak. Where required, nuclear Overhauser effect (nOe) spectroscopy, HETCOR and decouplings were employed to determine the position of substituents. The following abbreviations are used for the multiplicity of the signals:

s singlet br s broad singlet d doublet dd doublet of doublets t triplet td triplet of doublets dt doublet of triplets q quartet m multiplet

Electron-impact mass spectra were recorded on a Finnigan-MAT 8200 spectrometer at 70 eV. Major peaks are reported with intensities as percentages of the base peak. Accurate masses of key compounds were recorded on a VG70-70E double focusing magnetic sector mass spectrometer using a VG 11-250J data system. Melting points were recorded on a Reichert Koffler hot-stage apparatus, and are uncorrected. Optical rotations were obtained on a Jasco DIP 370 digital polarimeter.

5.2 NAMING AND NUMBERING OF COMPOUNDS

General classes of compounds frequently encountered in this thesis include methyl cinnamates, quinolines, indoles and pyrrolidones. The standard numbering of these systems is illustrated in Scheme 5.1. The naming and numbering follows IUPAC rules, 5 although commonly used trivial names have been employed in the text of the previous chapters (e.g. guaiacol, melatonin, tryptophan etc.). 156

CINNAMATES QUINOLINES 6' 3 5 CO2Me 4 4 6 3 7 2 Ba Nr 1

INDOLES PYRROLIDONES 4 4 3 6 2 7 7a H

SCHEME 5.1

5.3 PREPARATION OF LAVENDAMYCIN PRECURSORS

5.3.1 PREPARATION OF 2-HYDROXYQUINOLINES

5-Methoxy-2,3-dinitrophenol (2.60b)

A well stirred solution of 3-methoxyphenol (660 mg, 5.32 mmol) in DME was cooled to -78 °C. Nitronium tetrafluoroborate (1.94 g, 11.7 mmol) was added to the solution in one portion and the reaction was monitored by TLC analysis until consumption of the starting material was indicated. The entire reaction mixture was filtered through a silica column, the solvent was removed in vacuo, and the crude product purified via column chromatography on silica gel using ethyl acetate-hexane (1:1) as the solvent to yield the compound as yellow crystals (746 mg, 66 %).. Mp: 153 -155 °C (lit. 6 157 - 159 °C)

1 H 8: 4.03 (3H, s, OCH3), 6.71 (1H, s, H-6), 8.79 (1H, s, H-3), 11.10 (1H, br s, OH), "C 8: 57.5 (-OCH3), 102.5 (C-3), 124.9 (C-6), 125.9 (ipso, C-2), 32.5 (ipso, C-4), 159.6 (ipso, C-1), 160.1 (ipso, C-5) MS m/z: 214(M+, 100 %) 157

2,4-Dinitrophenyltriflate (2.54b)

OTf

0 2N NO2

To a stirred solution of 2,4-dinitrophenol (502 mg, 2.73 mmol) and 2,4,6-collidine (0.43 ml, 3.27 mmol) in dichloromethane (25 ml) was added triflic anhydride ( 0.55m1, 3.27 mmol). The reaction mixture was stirred at room temperature until TLC indicated total conversion of starting material. Removal of the solvent in vacuo, followed by flash chromatography (9:1 hexane- EtOAc) afforded the product as light orange viscous oil (860 mg, 99 %).

I li 8: 7.70 (1H, d J6,3 = 9.0 Hz, 14-6), 8.61 (1H, dd J5,3 = 2.7 Hz J5,6 = 9.0 Hz , H-5), 9.09 (1H, d13,3= 2.7 Hz, H-3)

13C 8: 116.31 and 120.56 (ipso, C-2 and C-4), 122.44 (C-6), 125.66 (C-5), 129.51 (C-3), 145.06 (ipso, C-1) MS m/z: 316 (Mt 23 %)

5-Methoxy-2,4-dinitrophenyltriflate (2.60c)

Me0 011

02N NO2

A stirred solution of 2.60b (259 mg, 1.21 mmol) in dichloromethane (6 ml) was cooled to 0 °C under argon. 2,4,6-Collidine (0.19 ml, 1.45 mmol) was added and the reaction mixture stirred for 1 minute. Triflic anhydride (0.24 ml, 1.45 mmol) was then added and the reaction mixture stirred at room temperature until TLC indicated total conversion of starting material. Removal of the solvent in vacuo was followed by flash chromatography (9:1 hexane-EtOAc) which afforded the product as light orange crystals (356 mg, 85 %).

Mp: 49 - 52 °C

1 1-1 8: 4.11 (3H, s, OCI13), 7.08 (1H, s, H-6), 8.79 (1H, s , H-3)

13 C 8: 58.17 (-0CH3), 110.39 (C-6), 116.35 and 120.61 (ipso, C-2 and C-4), 124.99 (C- 3), 145.06 (ipso, C-1), 157.31 (ipso, C-5) MS m/z: 346 (M+, 57 %), 316 (M+-NO, 4 %) 158

Methyl trans-2'-nitrocinnamate (2.55a)

CO2Me

A mixture of 2-bromonitrobenzene 2.54a (808 mg, 4.0 mmol), methyl acrylate (0.72 ml, 8.0 mmol), triethylamine (0.70 ml, 5.0 mmol), palladium(II)acetate (46 mg, 0.2 mmol) and triphenylphosphine (106 mg, 0.4 mmol) was heated to 95 °C in a sealed flask. The reaction mixture was stirred at this temperature for 18 hours, filtered through celite and followed by evaporation in vacuo. Flash chromatography (4:1 hexane: EtOAc) afforded 2.55a as a pale yellow liquid (811 mg, 98%).

1 11 5: 3.71 (3H, s, -COOCH3), 6.27 (1H, d J2,3 = 15.9 Hz, H-2), 7.46 (1H, t J5c6i = J5,4 =

7.5 Hz , H-5'), 7.55 (2H, m, H-3' and H-4 1), 7.91 (1H, d J6 , 5 = 7.8 Hz , H-6'), 7.98 (1H, d J3,2 = 15.6 Hz , H-3)

I3 C 5: 51.65 (-OCH3), 122.48 (C-2), 124.63 (C-3'), 128.88 (C-6'), 130.10 (ipso, C-1'), 130.21 (C-4'), 133.44 (C-5'), 139.86 (C-3), 148.02 (ipso, C-2'), 166.01 (CO, C-1) MS m/z: 207 (Mt, 15 %), 176 (Mt OCH 3, 38 %), 161 (Mt NO2), 130 (Mt OCH3 - NO2, 63 %)

Methyl trans-2',4'-dinitrocinnamate (2.55b)

CO2Me

A mixture of 2.54b (330 mg, 1.04 mmol), methyl acrylate (0.36 ml, 4.2 mmol), epichlorohydrin (0.16 ml, 2.1 mmol) and Pd(PPh3)4 (180 mg, 0.16 mmol) was heated to 95 °C in a sealed flask. The reaction mixture was stirred at this temperature for 18 hours, filtered through celite which was followed by evaporation in vacuo. Flash chromatography (3:1 hexane: EtOAc) of the residue afforded 2.55b as an orange oil (223 mg, 85 %).

1 H 5: 3.83 (3H, s, -COOCH3), 6.46 (1H, d .12,3 = 15.9 Hz, H-2), 7.82 (1H, d = 8.4

Hz, H-61), 8.09 (1H, d J3,2 = 15.6 Hz, H-3) 8.47 (1H, d J5.,6 = 8.4 Hz ./5',3' = 2.4 Hz H-5'), 8.87 (1H, d = 2.4 Hz, H-3') 159

13C 8: 52.35 (-OCH3), 120.51 (C-2), 126.01, 127.57 and 130.53, (C-3', C-5' and C-6'), 136.27 (C-1 1), 137.79 (C-3), 147.94 (ipso, C-2' and C-4'), 165.24 (CO, C-1) MS m/z: 252 (Mt, 5 %), 221 (Mt OCH3, 85 %)

Methyl trans-5'-methoxy-2',4'-dinitrocinnamate (2.60d)

MeO CO2Me

02N NO2

A mixture of 2.60c (180 mg, 0.52 mmol), methyl acrylate (0.19 ml, 2.1 mmol), epichlorohydrin (0.08 ml, 1.04 mmol) and Pd(PPh3) 4 (90 mg, 0.077 mmol) was heated to 95 °C in a sealed flask. The reaction mixture was stirred at this temperature for 18 hours, filtered through celite and followed by evaporation in vacuo Flash chromatography (3:1 hexane:EtOAc) afforded 2.60d as a viscous orange oil (136 mg, 85 %).

I ll 8: 3.84 (3H, s, -000Q13), 4.09 (3H, s, -0013), 6.35 (1H, d J2,3 = 15.6 Hz, H-2), 7.16 (1H, s , H-6'), 8.19 (1H, d J3,2. = 15.6 Hz , H-3), 8.71 (1H, s, H-3')

13 C 8: 52.35 (-COOCH3), 57.54 (OCH3), 113.65 (C-6'), 123.79 (C-2), 125.16 (C-3'), 137.58 (C-1 1), 138.05 and 139.42 (C-2' and C-4'), 139.56 (C-3), 155.99 (C-5'), 165.35 (CO, C-1) MS m/z: 282 (Mt, 8 %), 251 (Mt OCH3, 12 %), 236 (Mt NO2, 100%), 205 (Mt OCH3 - NO2, 24 %)

2-Hydroxyquinoline (2.57a)

N OH

To a stirred solution of methyl trans-2'-nitrocinnamate 2.55a (250 mg, 1.21 mmol) in absolute ethanol (3 ml) was added SnC12.2H20 (1.63g, 7.24 mmol). The reaction mixture was heated at 70 °C until TLC indicated total consumption of the starting material. The solvent was removed in vacuo followed by the addition of 3M HC1 and boiling of the resultant reaction mixture for 24 hours. The reaction mixture was cooled down to room temperature, poured onto ice, neutralised with a saturated sodium bicarbonate solution and extracted with dichloromethane (3 x 20 ml portions). The combined dichloromethane fractions were dried over MgSO4. Removal of the

160

solvent in vacuo followed by flash chromatography (1:1 hexane:EtOAc) afforded 2.57a as white crystals (105mg, 60%).

Mp: 196 -198 °C (lit. 7 198 -199 °C)

1 1-1 8: 6.71 (1H, d J3,4 = 9.5 Hz, H-3), 7.19 (1H, t, J6,5 = J6,7 = 7.0 Hz, I-1-6), 7.43 - 7.56 (3H, m, H-5, H-7 and H-8), 7.81 (1H, dJ4,3 = 9.5 Hz, H-4)

13 C 8: 116.26 (C-3), 119.90 (C-4a), 121.33 (C-6), 122.64 (C-5), 127.69 (C-8), 130.62 (C-7), 138.55 (C-8a), 141.01 (C-4), 164.72 (ipso, C-2) MS m/z: 145 (M+, 100 %), 117 (Mt CO, 73 %), 90 (Mt HCN)

7-Amino-2-hydroxyquinoline (2.57b)

H2N N OH

To a stirred solution of methyl trans-TA-dinitrocinnamate 2.55b (103 mg, 0.408 mmol) in absolute ethanol (2 ml) was added SnC1 2.2H20 (920 mg, 4.08 mmol). The reaction mixture was heated at 70 °C for 2 hours. The solvent was removed in vacuo followed by the addition of 3M HCl and boiling of the resultant reaction mixture for 24 hours. The reaction mixture was cooled down to room temperature, poured onto ice, neutralised with a saturated sodium bicarbonate solution and extracted with dichloromethane (3 x 25 ml portions). The combined dichloromethane fractions were dried over MgSO4. Removal of the solvent in vacuo followed by flash chromatography (1:1 hexane:EtOAc) afforded 2.57b (65mg, 100%).

1 H 8: (in DMSO-d6) 5.78 (2H, br s, -Nth), 5.99 (1H, d .J3,4 = 9.3 Hz, H-3), 6.33 (1H, d, 46 = 2.1 Hz, H-8), 6.40 (1H, dd J6,5 = 8.1 Hz, J6,8 =2.1 Hz, H-6), 7.22 (1H, d .15,6 = 8.4 Hz, H-5), 7.57 (1H, d J4,3 = 9.3 Hz, H-4), 11.26 (1H, br s, -OH)

13C 8: (in DMSO-d6) 96.32 (C-4a), 110.03 (C-8), 110.54 (C-3), 114.57 (C-6), 128.68 (C- 5), 140.16 (C-4), 141.02 (C-8a), 151.17 (ipso, C-7), 162.55 (ipso, C-2) MS m/z: 145 (M+, 100 %), 132 (Mt CO, 83 %) HREI-MS ink:160.063797 (C9H8N201 requires 160.063663) 161

7-Amino-2-hydroxy-6-methoxyquinoline (2.62)

Me0

H2N N 01-1

To a stirred solution of methyl trans-5'-methoxy-2',4'-dinitrocinnamate 2.60d (90 mg, 0.319 mmol) in absolute ethanol (2 ml) was added SnC12.2H20 (803 mg, 3.57 mmol). The reaction mixture was heated at 70 °C until TLC indicated consumption of the starting material. The solvent was removed in vacuo followed by the addition of 3M HC1 (25 ml) and boiling of the resultant reaction mixture for 24 hours. The reaction mixture was cooled down to room temperature, poured onto ice, neutralised with a saturated sodium bicarbonate solution and extracted with dichloromethane (3 x 25 ml portions). The combined dichloromethane fractions were dried over MgSO4. Removal of the solvent in vacuo followed by flash chromatography (EtOAc) afforded 2.62 (60 mg, 100%) as unstable white needles.

8: (in DMSO-d6) 3.77 (3H, s, 0013), 5.57 (2H, br s, -NI12), 6.04 (1H, d J3,4 = 9.3

Hz, H-3), 6.47 (1H, s, H-8), 6.93 (1H, s, H-5), 7.58 (1H, d J4 , 3 = 9.0Hz, H-4), 11.28 (1H, br s, -OH) 0C 8: (in DMSO-c/6) 55.49 (-OCH3), 96.89 (C-5), 107.03 (C-8), 110.54 (C-3), 111.65 (C-4a), 114.92 (C-3), 136.68 (C-8a), 139.81 (C-4), 141.86 and 142.75 (ipso, C-6 and C-7), 162.12 (ipso, C-2) MS m/z: 190 (Mt, 27 %), 175 (Mt CH3, 38 %) HREI-MS m/z :190.074131 (C1othoN202 requires 190.074228)

53.2 PREPARATION OF A LAVENDAMYCIN AB SYNTHON FROM GUIACOL

2-Methoxy-4,6-dinitrophenol (2.51)

ON OH

02N NO2

Guiacol 2.42 (12g, 0.097 mol) was dissolved in dichloromethane (150 ml) and the reaction mixture cooled to -10 °C. A 1:1 mixture of nitric acid: acetic acid ( 25 ml, 0.197 mol HNO3) was added to the vigorously stirred reaction mixture over 5 minutes. Upon completion of the 162

reaction, the reaction mixture was neutralised with a sodium bicarbonate solution and extracted with 3 x 100 ml portions of dichloromethane. The dichloromethane fractions were collected, dried over MgSO4 and the solvent was removed in vacuo. The crude product was purified by flash chromatography (1:1 hexane:EtOAc and 5% NEt3) and the product washed from the column with methanol to furnish 2.51 as an orange paste ( 14.1 g, 68%).

Mp: 85 -88 °C

1 E1 8: 3.99 (3H, s, -OCH3), 7.92 (1H, d J3,5 = 3.0 Hz, H-3), 8.35 (1H, d J5,3 = 2.7 Hz, H- 5)

13 C 8: 57.24 (-OCH3), 109.31 and 113.4 (C-3 and C-5), 136.2 and 137.56 (C-4 and C-6), 148.82 and 149.93 (C-1 and C-2) MS ink: 214 (Mt 100 %)

2-Methoxy-4,6-dinitrophenyltriflate (2.52)

02N

A stirred solution of 2.51 (205 mg; 0.958 mmol) in dichloromethane (14 ml) was cooled to 0 °C under argon. 2,4,6-Collidine (0.15 ml; 1.15 mmol) was added and the reaction mixture stirred for I minute. Triflic anhydride ( 0.19 ml; 1.15 mmol) was then added and the reaction mixture stirred at room temperature until TLC indicated total conversion of starting material. Removal of the solvent in vacuo was followed by flash chromatography (9:1 hexane-EtOAc) which afforded the product as a colourless oil (248 mg, 75 %).

1 1-1 8: 4.11 (3H, s, OCH3), 8.14 (1H, d J3,5 = 2.4 Hz, H-3), 8.51 (1H, d J5,3 = 2.7 Hz, H- 5)

13C 8: 57.95 (-0CH3), 111.91 (C-5), 112.51 (C-3 ), 135.49 (ipso, C-6), 142.54 (ipso, C- 4), 146.19 (ipso, C-1), 153.59 (ipso, C-2) MS ink: 346 (Mt 8 %), 197 (M +-0Tf, 16 %) 163

Methyl trans-2'-methoxy-4',6'-dinitrocinnamate (2.53)

OMe C Pe

A mixture of 2.52 (1.5 g, 4.3 mmol), methyl acrylate (1.56 ml, 17.3 mmol), epichlorohydrin (0.68 ml, 8.6 mmol) and Pd(PPh3)4 (750 mg, 0.65 mmol) was heated to 95 °C in a sealed flask. The reaction mixture was stirred at this temperature for 20 hours, filtered through celite and followed by evaporation in vacuo. Flash chromatography (1:1 hexane:EtOAc) afforded 2.53 as a viscous orange oil (954 mg, 78 %).

'H 8: 3.82 (3H, s, -COOCILI3), 4.04 (3H, s, -00j3), 6.66 (1H, d J2,3 = 16.1 Hz, H-2), 7.65 (1H, d J3,2 = 16.1 Hz , H-3), 7.95 (1H, d = 2.1 Hz , H-3'), 8.26 (1 H, d J5 ,3 = 2.1 Hz , H-5')

13C 8: 51.51 (-COOCH3) 57.31 (-OCH3), 108.91 (C-5'), 111.23 (C-3'), 124.07 (ipso, C- 1'), 127.76 (C-2), 132.49 (C-3), 147.67 (ipso, C-6'), 150.28 (ipso, C-4'), 159.09 (ipso, C-21) 165.95 (CO, C-1) MS m/z: 282 (Mt, 28 %), 251 (Mt OCH3, 75 %), 236 (Mt NO2, 14%), 205 (Mt OCH3 - NO2, 32 %)

7-Amino-2-hydroxy-5-methoxyquinoline (2.44)

H2N N OH

To a stirred solution of 2.53 (90 mg, 0.319 mmol) in absolute ethanol (3 ml) was added SnC12.2H20 (720 mg, 3.19 mmol). The reaction mixture was heated at 70 °C until TLC indicated consumption of the starting material. The solvent was removed in vacuo followed by the addition of 3M HCI (25 ml) and boiling of the resultant reaction mixture for 24 hours. The reaction mixture was cooled down to room temperature, poured onto ice, neutralised with a saturated sodium bicarbonate solution and extracted with dichloromethane (3 x 25 ml portions). Removal of the solvent in vacuo followed by flash chromatography (EtOAc) afforded 2.44 (27 mg, 45%) as a brown powder which decomposes on standing. 164

I I-I 8: (in DMSO-d6) 3.77 (3H, s, OCH3), 5.82 (2H, br s, 5.91 (1H, d J3,4 = 9.3 Hz, H-3), 5.96 (1H, d Js,6 =1.5 Hz, H-8), 5.98 (1H, d15,8 =1.5 Hz, H-6), 7.72 (1H, d 4,3 = 9.6 Hz, H-4), 11.21 (1H, br s, -OH)

13 C 8: (in DMSO-d6) 54.45 (-OCH 3), 89.56 (C-4a), 91.63 (C-6), 101.04 (C-8), 112.86 (C-3), 134.18 (C-4), 142.00 (C-8a), 152.19 (ipso, C-7), 156.31 (ipso, C-5), 162.64 (ipso, C-2) MS m/z: 190 (Mt, 100 %), 147 (Mt- CH3 - CO, 45%) HREI-MS m/z:190.074772 (C101410N202 requires 190.074228)

5.3.3 PREPARATION OF LAVENDAMYCIN LEFT HAND SIDE FROM A PREFORMED RING SYSTEM

8-Hydroxyquinoline-N-oxide (2.71)

To a solution of 8-hydroxyquinoline (660 mg, 4.52 mmol) in dichloromethane (4 ml) was added m-CPBA (50 %, 790 mg, 4.59 mmol). The reaction mixture was stirred at room temperature until TLC indicated total consumption of the starting material (3 hours). Solid sodium bisulfite was added carefully to destroy any unreacted m-CPBA and the reaction mixture was filtered. Solid potassium carbonate was added to the filtrate to precipitate the m-chlorobenzoic acid byproduct as its potassium salt and the mixture was filtered again. The filtrate was dried over MgSO4, the solvent was removed in vacuo and flash chromatography (1:1 hexane:EtOAc) afforded the product as light yellow crystals (549 mg, 75 %). Mp: 131 - 134 °C (lit. 8 138 - 140 °C)

I H 8: 7.02 (1H, d J7,6 = 8.1 Hz, H-7), 7.21 (2H, m, H-3 and H-5), 7.45 (1H, t J6,5 = J6,7 = 8.1 Hz, H-6), 7.75 (1H, d J4,2 = 8.4 Hz, H-4), 8.22 (1H, di2A= 5.7 Hz, H-2), 15.01 (1H, br s, -OH) 13c 8: 114.92 (C-3), 116.79 (C-7), 120.42 (C-5 and C-6), 129.72 (C-4), 130.61 (C-2), 132.29 (C-4a), 134.53 (C-8a), 154.04 (ipso, C-8) MS m/z: 161 (Mt, 100 %) 165

2,8-Diacetoxyquinoline (2.72)

OAc

Acetic anhydride (3 ml) was added to a solution of 2.71 (100 mg, 0.62 mmol) in dichloromethane (3 ml). The reaction mixture was stirred under reflux for 3 days to go to completion. The solvent was removed in vacuo and flash chromatography (1:1 hexane:EtOAc) afforded the product 2.72 as white crystals (119 mg, 78 %).

Mp: 238 - 242 °C 'H 5: 2.36 and 2.45 (2 x 3H, 2 x s, 2 x -OAc), 7.21 (1H, d J3,4= 8.7 Hz, H-3), 7.43 (1H, dd J7,6 = 7.2 Hz, th ,5= 1.2 Hz, H-7), 7.51 (1H, t J6,5 = .16,7 = 7.5 Hz, H-6), 7.73 (1H, dd J5,6= 8.1 HZ, J5,7 = 1.5 Hz, H-5), 8.22 (1H, d J4,3 = 8.7 Hz, H-4), 15.01 (1H, br s, -OH)

13C 3: 20.93 and 21.27 (2 x COCH3), 116.45 (C-3), 122.47 (C-7), 125.49 (C-5), 126.20 (C-6), 128.35 (C-4a), 139.66 (C-8a), 139.93 (C-4), 146.69 (ipso, C-8), 156.43 (ipso, C-2), 168.94 (COCH3), 169.67 (COCH3) MS m/z: 245 (Mt, 3 %), 203 (Mt - CH2CO, 3 %), 161 (M t - 2 x CH2CO, 100 %)

2,8-Dihydroxyquinoline (2.73)

OH

2,8-Diacetoxyquinoline 2.72 (73 mg, 0.298 mmol) was dissolved in methanol (4 ml) and triethylamine (0.4 ml) added to the solution. The reaction mixture was stirred at room temperature for 4 hours and the solvent removed in vacuo to furnish compound 2.73 (47 mg, 100 %) as white cubic crystals.

166

Mp: 275 -280 °C (lit. 8 > 250 °C)

1 1-1 8: (in DMSO-d6) 6.46 (1H, d .13A= 9.3 Hz, H-3), 6.95 -7.01 (2H, m, H-6 and H-7),

7.08 (1H, dd J5,6 = 7.5 Hz, 15,7= 1.5 Hz, H-5), 7.83 (1H, d J4,3 = 9.3 Hz, H-4), 10.47 (1H, br s, -OH)

13 C 8: 114.63 (C-3), 118.21 (C-7), 120.04 (C-4a), 121.93 and 122.19 (C-5 and C-6), 128.16 (C-8a), 140.49 (C-4), 143.71 (C-8), 161.46(C-2) MS m/z: 161 (Mt 100 %), 133 (M+ - CO, 67 %)

8-Allyloxyquinoline (2.75)

To a solution of 8-hydroxyquinoline (500 mg, 3.45 mmol) in acetone (10 ml) was added solid potassium carbonate (570 mg, 4.14 mmol) and allyl bromide (3.5 ml, 4.14 mmol). The reaction mixture was refluxed for 48 hours followed by filtration and removal of the solvent in vacuo. Flash chromatography (2:1 hexane:EtOAc) of the residue afforded the product (541 mg, 85 %) as an orange oil.

3: 4.82 (2H, d 5.4 Hz, -OCH2CH=CH2), 5.28 (1H, d = 10.5 Hz, OCH2CH=CH2 cis), 5.42 (1H, d .13',21= 17.1 Hz, OCH2CH=C112 trans), 6.17 (1H,

ddt fT,3, = 17.1 Hz, fry = 10.5 Hz, 12',1' = 5.4 Hz, OCH2CH=CH2), 7.01 (1H, d 17,6 = 7.2 Hz, H-7), 7.30 -7.45 (3H, m, H-3, H-5 and H-6), 8.06 (1H, dd 14,3 = 8.1 Hz,

14,2 = 1.5 Hz, H-4), 8.90 (1H, dd f2,3 = 4.2 Hz, 12,4 = 1.8 Hz, H-2)

D C 8: 69.49 (C-1'), 108.89 (C-7), 117.97 (C-3'), 119.38 and 121.24 (C-3 and C-5), 126.22 (C-6), 129.11 (C-4a), 132.79 (C-2'), 135.50 (C-4), 140.03 (C-8a), 148.96 (C-2), 153.88 (ipso, C-8) HREI-MS m/z:185.091874 (C121-11INIOI requires 185.091889) 167

8-Benzyloxyquinoline (2.76)

0

To a solution of 8-hydroxyquinoline (1.06 g, 7.31 mmol) in acetone (15 ml) was added solid potassium carbonate (1.07 g, 8.76 mmol) and benzyl bromide (1.05 ml, 8.76 mmol). The reaction mixture was refluxed for 24 hours, allowed to cool to room temperature and followed by filtration and removal of the solvent in vacuo. Flash chromatography (2:1 hexane:EtOAc) of the residue afforded the product (990 mg, 58 %) as light orange crystals.

Mp: 58 °C

1 1-1 8: 5.41 (2H, s, -CHPh), 7.01 (1H, td J= 4.5 Hz, J7,6= 6.3 Hz, H-7), 7.24 - 7.52 (8H, m, H-3, H-5, H-6 and 5 aromatic protons), 8.09 (11-I, dd J4,3 = 8.4 Hz, J4,2 = 2.0 Hz, H-4), 8.95 (1H, dd J2,3 = 4.2 Hz, J2,4 = 2.0 Hz, H-2)

I3 C 8: 70.71 (-CH2Ph), 109.84 (C-7), 119.77 and 121.49 (C-3 and C-5), 126.46 and 127.69(C-6 and 1 aromatic carbon), 127.00 (2 aromatic carbons), 128.51 (2 aromatic carbons), 129.40 (C-4a), 135.77 (C-4), 136.86 (C-1'), 140.40 (C-8a), 149.24 (C-2), 154.22 (ipso, C-8) MS m/z: 235 (M±, 61 %), 158 (MIE - C6H5, 19 %)

8-Methoxyquinoline (2.77)

To a solution of 8-hydroxyquinoline (1.05 g, 7.23 mmol) in acetone (15 ml) was added solid potassium carbonate (1.0 g, 7.24 mmol) and methyl iodide (0.45 ml, 7.23 mmol). The reaction mixture was refluxed for 24 hours. It was allowed to cool to room temperature and followed by filtration and removal of the solvent in vacuo. Flash chromatography (2:1 hexane:Et0Ac) of the residue afforded the product (815 mg, 71 %) as a light orange oil. 168

I li 8: 4.07 (3H, s, -OCH3) 7.02 (1H, dd J7,5 = 1.5 Hz, 46 = 7.8 Hz, H-7), 7.35 (111, dd

J5,7= 1.5 Hz, J5,6 = 8.4 Hz, H-5), 7.39 (1H, dd J3,4 = 8.4 Hz, J3,2= 4.2 Hz, H-3),

7.43 (1H, dd J6,5 = J6,7 = 8.1 Hz, H-6), 8.09 (1H, dd J4,3 = 8.4 Hz, J4,2 = 1.8 Hz, H-

4), 8.90 (1H, dd J2,3 = 4.8 Hz, J7,4= 1.5 Hz, H-2)

13 C 8: 55.95 (-OCH3), 107.47 (C-7), 119.45 (C-5), 121.58 (C-3), 126.62 (C-6), 129.26 (C-4a), 135.79 (C-4), 140.07 (C-8a), 149.10 (C-2), 155.27 (ipso, C-8) MS m/z: 159 (Mt, 100 %), 128 (Mt - OCH3, 32 %)

General procedure for the preparation of N-oxides (2.81 - 2.83) To a solution of the 0-protected quinoline derivatives (2.75 or 2.76 or 2.77, 2mmol) in dichloromethane (7 ml) was added m-CPBA (50 %, 4mmol). The reaction mixture was stirred at room temperature until TLC indicated total consumption of starting material followed by filtration through a thick, short silica gel column packed in 100 % ethyl acetate. The polar N- oxides were washed from the column with a 3:1 EtOAc:methanol eluent to afford the corresponding N-oxides.

8-Allyloxyquinoline-N-oxide (2.81)

The product was isolated as a dark brown viscous oil in a yield of 78 %. 1 1-1 8: 4.64 (2H, m, -OCH2CH=CH2), 5.24 (114, d J3,2 = 10.1 Hz, OCH2CH=Q12 cis),

5.56 (1H, d J3',21 = 17.4 Hz, OCH2CH=CH2 trans), 6.07 (1H, m, OCH2CH=CH2),

7.01 (1H, dd J7,6 = 7.5 Hz, J7,5 = 1.5 Hz, H-7), 7.09 (1H, dd ./3,4 = 8.4 Hz, J3,2 =

6.3 Hz, H-3), 7.24 -7.39 (2H, m, H-5 and H-6), 7.51 (1H, dd J4,3 = 8.1 Hz,

J4,2 = 0.9 Hz, H-4), 8.30 (1H, dd J2,3 = 6.0 Hz, J2,4 = 0.9 Hz, H-2) pc 8: 71.52 (C-1'), 113.94 (C-7), 117.81 (C-3'), 121.08 and 121.13 (C-3 and C-5), 125.99 (C-4), 128.57 (C-6), 132.47 (C-2'), 133.62 and 134.48 (C-4a and C-8a), 138.22 (C-2), 152.25 (ipso, C-8) MS m/z: 201 (Mt, 24 %), 184 (Mt - OH, 100%) 169

8-Benzyloxyquinoline-N-oxide (2.82)

The product was obtained as a brown oil in a yield of 52 %.

1 1-1 8: 5.28 (2H, s, -CHPh), 7.05 - 7.65 (10H, m, H-3, H-4, H-5, H-6, H-7 and 5 aromatic protons), 8.41 (1H, dd J2,3 = 6.0 Hz, J2,4 = 0.9 Hz, H-2) "C 8: 72.77 (-CH2Ph), 111.56 (C-7), 121.18 and 121.39 (C-3 and C-5), 125.49 (C-4), 127.00 (2 aromatic carbons), 127.58 (1 aromatic carbon), 128.43 (2 aromatic carbons), 128.55 (C-6), 133.68 and 134.83 (C-4a and C-8a), 136.53 (C-1'), 138.07 (C-2), 152.20 (ipso, C-8) MS m/z: 252 (W + 1, 31 %), 145 (Mt + 1 - OCH2C6H5, 17 %), 128 (M t + 1 - OCH2C6H5 - OH, 100 %)

8-Methoxyquinoline-N-oxide (2.83)

The product was obtained as a dark brown oil in a yield of 74 %.

1 H 8: 3.96 (3H, s, -OCH3) 7.01 (1H, d J7,6 = 7.5 Hz, H-7), 7.14 (1H, dd J3,4 = 8.1 Hz, J3,2

= 6.0 Hz, H-3), 7.32 (1H, d J5,6 = 8.1 Hz, H-5), 7.41 (1H,t J6,5 = J6,7 = 8.1 Hz, H- 6), 7.56 (1H, d 43 = 8.4 Hz, H-4), 8.35 (1H, d J2,3 = 6.0 Hz, H-2)

13C 8: 55.96 (-OCH3), 110.83 (C-7), 120.39 (C-5), 121.14 (C-3), 125.49 (C-4), 128.51 (C-6), 133.53 and 133.98 (C-4a and C-8a), 137.89 (C-2), 153.41 (ipso, C- 8) MS m/z: 175 (Mt, 39 %), 158 (Mt - OH, 100 %)

170

General procedure for the preparation of 2-chloroquinoline derivatives

To a solution of the N-oxide (1.5 mmol) in dichloroethane (3 ml) was added POCI3 (4.5 mmol). The reaction mixture was stirred at room temperature until TLC indicated total consumption of the starting material (< 1.5 hours). The volatile components were removed in vacuo and the residue neutralised with NaHCO3 followed by extraction with dichloromethane (3 x 20 ml portions). The combined dichloromethane fractions were dried over MgSO4 and the solvent evaporated under reduced pressure. Flash chromatograpy of the residue with the appropriate solvent afforded the product.

8-Allyloxy-2-chloroquinoline (2.84)

The title compound was obtained as a brown oil in a yield of 66 %.

1 11 5: 4.84 (2H, d 5.4 Hz, -OCH2CH=CH2), 5.30 (1H, d J3.,2' = 10.5 Hz, OCH2CH=CH2 cis), 5.44 (I H, d = 18.9 Hz, OCH2CH=CH2 trans), 6.12 (1H, m, OCH2CH=CH2), 7.01 (1H, d J7 ,6 = 7.8 Hz, J7,5 = 1.2 Hz, H-7), 7.32 -7.46 (3H,

m, H-3, H-5 and H-6), 8.03 (1H, dd J4 , 3 = 8.7 Hz, H-4)

13C 5: 68.47 (C-1'), 110.81 (C-7), 118.08 (C-3 1), 119.33 (C-5), 122.89 (C-3), 126.99 (C- 6), 128.01 (C-4a), 132.93 (C-2'), 138.74 (C-4), 139.64 (C-8a), 149.74 (ipso, C-2), 153.41 (ipso, C-8) HREI-MS m/z:220.052943 (Cl2HIINIOICII requires 220.152917) 171

8-Benzyloxy-2-chloroquinoline (2.85)

The title compound was obtained as cream crystals in a yield of 58 %.

Mp: 61 - 63 °C

1 1-1 5.39 (2H, s, -C_H_2Ph), 7.08 - 7.65 (9H, m, H-3, 1-1-5, H-6,11-7 and 5 aromatic protons), 8.11 (1H, d J4,3 = 8.4 Hz, H-4)

13 C 8: 70.99 (-CH2Ph), 111.88 (C-7), 119.59 (C-5), 123.06 (C-3), 127.05 (2 aromatic carbons), 127.45 (1 aromatic carbon), 127.82 (C-6), 128.11 (C-4a), 128.48 (2 aromatic carbons), 136.31 (C-1 1), 138.82 (C-8a), 139.72 (C-4), 149.54 (ipso, C-2), 152.93 (ipso, C-8) MS m/z: 270 (M+ + 1, 14 %), 192 (M+ - C6H5, 8 %),156 (M + - C6H5 - Cl, 19 %)

2-Chloro-8-methoxyquinoline (2.86)

one

The title compound was isolated as white crystals in a yield of 77 %.

Mp: 78 °C 1 H 8: 4.03 (3H, s, -0CL13) 7.04 (1H, dd J7,5 = 0.9 Hz, J7,6 = 7.5 Hz, H-7), 7.30 - 7.55 (3H, m, 1-1-3, H-5 and H-6), 8.02 (1H, dJ4,3 = 8.4 Hz, H-4)

13 C 8: 55.99 (-00-13), 108.79 (C-7), 119.07 (C-5), 122.89 (C-3), 127.09 (C-6), 127.84 (C-4a), 138.66 (C-4), 139.40 (C-8a), 149.69 (ipso, C-2), 154.46 (ipso, C- 8) MS m/z: 193 (Mt 96 %), 162 (M+ - OCH3, 100 %), 162 (M+ - OCH3 - Cl, 32 %), 172

2-Chloro-6-methoxyquinoline (2.123)

Me0

The title compound was isolated as white crystals in a yield of 32 %.

Mp: 58 - 66 °C 8: 3.87 (3H, s, -0013) 7.00 (1H, d J5,7 = 2.7 Hz, H-5), 7.27 (1H, d, J3,4 = 8.7 Hz, H- 3), 7.32 (1H,dd J7,8 = 9.0 Hz, J7,5= 2.4 Hz, H-7), 7.85 (1H, d J8,7= 9.3 Hz, H-8) 7.92 (1H, d14,3= 8.7 Hz, H-4)

13 C 8: 55.52 (-OCH3), 105.10 (C-5), 122.30 and 122.89 (C-3 and C-7), 127.73 (C-4a), 127.70 (C-8), 137.47 (C-4), 143.57 (C-8a), 147.79 (ipso, C-2), 157.86 (ipso, C- 6) MS m/z: 193 (Mt, 100 %), 178 (Mt - CH3, 22 %), 158 (Mt - CI, 8 %), 150 (Mt - CH3 - CO, 56 %)

4-Chloro-6-methoxyquinoline (2.124)

CI Mao

This compound was isolated as the major, but unwanted product using the experimental procedure described for the preparation of 2.123. It was obtained as white crystals in a yield of 56 %.

Mp: 62 - 66 °C

1 1-1 8: 3.93 (3H, s, -OCH3) 7.00 (1H, dJ5,7 = 2.7 Hz, 1-1-5), 7.24 - 7.48 (3H, m, H-3, H-5 and H-7), 7.97 (1H,d 47= 9.6 Hz, H-8), 8.58 (1H, d .12,3 = 4.5 Hz, H-2)

13 C 8: 55.63 (-OCH3), 101.47 (C-3), 121.23 and 123.19 (C-5 and C-7), 127.44 (C-4a), 131.19 (C-8), 140.79 (ipso, C-4), 145.05 (C-8a), 146.99 (C-2), 158.55 (ipso, C- 6) MS m/z: 193 (Mt, 100 %), 178 (Mt - CH3, 13 %), 150 (Mt - CH3 - CO, 64 %) 173

General procedure for the preparation of 2-bromoquinoline derivatives

METHOD A: To a solution of the N-oxide (1.5 mmol) in chloroform (4 ml) under argon was added POBr3 (2.25 mmol). The reaction mixture was refluxed until TLC indicated total consumption of the starting material (< 1.5 hours) and then allowed to cool to room temperature. The solid material was removed by filtration, the filtrate neutralised with NaHCO 3 and extracted with dichloromethane (3 x 20 ml portions). The combined dichloromethane fractions were dried over MgSO4 and the solvent evaporated under reduced pressure. Flash chromatograpy of the residue with the appropriate solvent afforded the corresponding product.

METHOD B: To a solution of the N-oxide (1.5 mmol) in dichloromethane under argon (6 ml) was added tetrabutylammonium bromide (TBAB) (1.5 mmol) followed by the addition of triflic anhydride (1.65 mmol). The reaction mixture was stirred at room temperature until TLC indicated total consumption of the starting material and then the solvent was removed in vacuo. Flash chromatography of the residue with the appropriate solvent afforded the corresponding product.

8-Allyloxy-2-bromoquinoline (2.87)

Br

2'

The title compound was isolated as a brown oil in a yield of 42 % using Method A and a yield of 30 % using Method B, respectively.

H ö: 4.83 (2H, d Jr,2, = 5.4 Hz, -OCH2CH=CH2), 5.30 (1H, d ha, = 10.5 Hz, OCH2CH=CH2 cis), 5.44 (1H, d 17.4 Hz, OCH2CH=CH2 trans), 6.13 (1H, m, OCH2CH=CH2), 7.07 (1H, d .17,6 = 7.8 Hz, ../7,5 = 1.2 Hz, H-7), 7.32 (1H, dd J5,6 = 8.4 Hz, J5,7 = 1.5 Hz, H-5), 7.43 (1H, t J6,5 = J6,2 = 7.8 Hz, H-6), 7.49 (1H, d ./3,4= 8.4 Hz, H-3), 7.91 (1H, d 43 = 8.7 Hz, H-4) 174

13 C 8: 69.91 (C-1'), 110.81 (C-7), 116.06 (C-3'), 119.47 (C-5), 126.38 (C-3), 127.14 (C- 6), 128.21 (C-4a), 132.98 (C-2'), 138.16 (C-4), 140.47 (C-8a), 140.85 (ipso, C-2), 153.53 (ipso, C-8) HREI-MS m/z:264.002398 (CI21-1111\1101Bri requires 264.002400)

2-Bromo-8-methoxyquinoline (2.86)

OW

The title compound was isolated as a colourless oil in a yield of 47 % using Method A.

1 11 8: 4.04 (3H, s, -OCH3) 7.07 (1H, dd J7,5 = 1.5 Hz, J7,6 = 7.5 Hz, H-7), 7.33 (1H, dd ../3,7= 1.2 Hz, iv = 8.1 Hz, H-5), 7.46 (1H, t ./6,5=./6,7= 7.8 Hz, H-6), 7.93 (1H, d A3 = 8.4 Hz, H-4)

13 C 8: 55.05 (-OCH3), 108.77 (C-7), 119.21 (C-5), 126.42 (C-3), 127.27 (C-6), 128.06 (C-4a), 138.15 (C-4), 140.17 (C-8a), 140.79 (ipso, C-2), 154.52 (ipso, C- 8) MS m/z: 237 (M+, 18 %), 206 (M+ - OCH3, 13 %), 127 (M+ - OCH3 - Br, 24 %),

2-Iodo-8-methoxyquinoline (2.117)

One

To a solution of the N-oxide 2.83 (210 mg, 1.20 mmol) in dichloromethane under argon (5 ml) was added tetrabutylanmionium iodide (TBAI) (444 mg, 1.20 mmol) followed by the addition of triflic anhydride (0.212 ml, 1.26 mmol). The reaction mixture was stirred at room temperature until TLC indicated total consumption of the starting material. The solvent was removed in vacuo and the residue flash chromatographed (1:1 hexane:EtOAc) to afford the product 2.117 as a brown oil (5 mg, 8 %).

1 H 8: 4.04 (3H, s, -0C113) 7.15 (1H, d J7,6= 7.8 Hz, H-7), 7.24 - 7.54 (3H, m, H-3, H-5, H-6), 8.05 (1H, d J4,3 = 8.4 Hz, H-4) 175

13 C 5: 55.05 (-OCH3), 110.61 (C-7), 117.63 (ipso, C-2) 119.62 (C-5), 127.25 (C-6), 128.02 (C-4a), 132.68 (C-3), 136.86 (C-4), 140.79 (C-8a), 148.51 (ipso, C- 8) MS m/z: 285 (Mt, 100 %), 127 (Mt - OCH3- 1, 44 %),

5.3.4 NITRATION OF A QUINOLINE DERIVATIVE

8-Hydroxy-5,7-dinitroquinoline (2.91)

024 OH

A solution of 8-hydroxyquinoline (2.0 g, 13.8 mmol) in 98 % H2SO4 (6.8 ml) was cooled to -5 °C followed by the addition of 98 % H2SO4 (1.4 ml) and 70 % HNO3 (2 ml). The reaction mixture was stirred at this temperature for 1 hour and poured into 9 ml ice-water. Filtration afforded the title compound as yellow crystals (2.68 g, 88 %). 9

1H 5: (in DMSO-d6) 8.26 (1H, dd J3,4 = 9.0 Hz, J3,2 = 5.4 Hz, H-3), 8.95 (1H, dd J2,3 = 5.1 Hz, J2,4= 0.9 Hz, H-4), 9.25 (1H, s, H-6), 9.82 (1H, dd, J4,3 = 8.7 Hz, J4,2 = 0.9 Hz, H-4)

13C 5: 122.62 (ipso, C-4a), 127.72 (C-3), 129.98 (C-6), 127.09, 130.59 and 137.65 (ipso, C-5, C-7 and C-8a), 141.72 (C-2), 142.28 (C-4), 161.92 (ipso, C-8) MS m/z: 235 (Mt, 100 %), 205 (Mt - NO, 38 %),

5.3.5 METHOXYCARBONYLATION OF QUINOLINE DERIVATIVES

General procedure for the methoxycarbonylation of 2-chloroquinoline derivatives

To a 50 ml glass pressure vessel equipped with a magnetic stirrer bar was added sequentially: the 2-chloroquinoline derivative (1.83 mmol), methanol (3 ml), PdC12(PPh3)2 (0.183 mmol) and triethylamine (7.3 mmol). The system was sealed, cooled down with liquid nitrogen, evacuated, purged with argon and charged with CO to 4 or 5 bar. The mixture was heated to the required 176 temperature (80 - 105 °C) for 6 hours. After cooling to room temperature, the solvent was removed in vacuo. Flash chromatography of the residue on silica with the appropriate solvent afforded the product.

Quinaldic acid methyl ester (2.120)

N COOMe

The title compound was obtained as yellow needles in a quantitative yield using the experimental methoxycarbonylation procedure described above (5 bar, 105 °C).

Mp: 75 °C 'H 8: 4.06 (3H, s, -00E3), 7.63 (I H, dt J6,7 = J6,5 = 7.2 Hz, J6,5 = 1.2 Hz„ H-6), 7.77

(1H, dt J7,8 =17,6 = 6.9 Hz, •17,5 = 1.5 Hz, H-7), 7.85 (1H, dd, .J5,6 = 8.1 Hz, 15,7 = 1.8 Hz, H-5), 8.17 (1H, d18,7 = 8.4 Hz, H-8), 8.20 - 8.32 (2H, m, H-3 and H-4) "C 8: 53.21 (-OCH3), 120.95 (C-3), 127.47 and 128.55 (C-5 and C-6), 129.27 (C-4a), 130.23 and 130.61 (C-7 and C-8), 137.24 (C-4), 147.42 and 147.79 (C-2 and C- 8a), 165.83 (COOMe) MS m/z: 187 (Mt 14 %), 129 (Mt CO2Me, 100 %)

8-Methoxyquinaldic acid methyl ester (2.121)

OMs

The title compound was obtained as white crystals in yield of 72 % using the experimental procedure described above (4 bar, 90 °C).

Mp: 92 °C

H 8: 3.99 and 4.03 (2 x 3H,2 x s, -OCH3 and COOCH3), 7.04 (1H, dd J2,6=7.8 Hz, •17,5

= 1.6Hz, H-7), 7.36 (1H, dd .15,6 = 8.4 Hz, iv= 1.2 Hz, H-5), 7.50 (1H, t, J6 , 7 = /6,5 =8.4 Hz, H-6), 8.13 - 8.23 (2H, m, H-3 and H-4) 177

13 C 8: 52.92 and 55.97 (-OCH3 and COOCH3), 107.89 (C-7), 118.93 (C-5), 121.38 (C- 3), 129.00 (C-6), 130.34 (C-4a), 136.98 (C-4), 139.24 (C-8a), 146.35 (ipso, C-2), 155.72 (ipso, C-8), 165.76 (COOCH3) MS m/z: 217 (Mt 51 %), 202 (Mt CH3, 94 %), 186 (Mt OCH3, 27 %), 156 (Mt 2 x OCH3 + H+,100 %)

5.3.6 AMINOCARBONYLATION OF QUINOLINE DERIVATIVES

N-benzyl-2-quinaldic acid amide (2.113)

0 A mixture of 2-chloroquinoline (670 mg, 3.6 mmol), benzylamine (0.57 ml, 5.3 mmol), tri-n- butylamine (0.95 ml, 4.0 mmol) and PdC1 2(PPh3)2 (77 mg, 0.9 mmol) was heated to 95 °C in a round bottomed flask under a CO atmosphere. The reaction mixture was stirred at this temperature for 1 hour, filtered through celite and followed by evaporation in vacuo. Flash chromatography (3:1 hexane:EtOAc) afforded 2.113 as white crystals (900 mg, 84 %).

Mp: 99 -101 °C

1 11 8: 4.72 (2H, d J= 6.3 Hz, -CH2Ph), 7.26 - 7.42 (5H, m, aromatic protons), 7.51 (1H,

dt 4,7 = J6,5 = 8.1 Hz, J6,8 = 1.2 Hz, H-6), 7.70 (1H, dt J7,6 = J7,5 = 8.2 Hz, J7,5 =

1.5 Hz, H-7), 7.83 (1H, dd, J5,6 = 8.1 Hz, 15,7 = 0.9 Hz, H-5), 8.04 (1H, dd, 15,7

8.7 Hz, 18,6 = 0.9 Hz, H-8), 8.27 (1H, d, J1,3 = 8.7 Hz, H-4), 8.33 (1H, d, 13,4 = 0.9 Hz, H-3) "C 5: 43.52 (-CH2Ph), 118.79 (C-3), 127.33 (C-5), 127.58 (C-6), 127.83 and 128.56 (5 aromatic carbons), 129.18 (C-4a), 129.52 (C-8), 129.92 (C-7), 137.31 (C-4), 138.16 (ipso, C-1'), 146.31 (C-8a), 149.50 (ipso, C-2), 164.26 (CONHCH2-) MS m/z: 262 (Mt 52 %), 129 (Mt 0=C—NCH2Ph, 83 %) 178

General procedure for the aminocarbonylation of 2-chloroquinoline derivatives

To a 50 ml glass pressure vessel equipped with a magnetic stirrer bar was added sequentially: the 2-chloroquinoline derivative (1.83 mmol), DMF (4 ml), PdCl2(PPh3)2 (0.183 mmol), the appropriate nucleophile and triethylamine (7.3 mmol). The system was sealed, cooled down with liquid nitrogen, evacuated, purged with argon and charged with CO (4 - 10 bar). The mixture was heated to the required temperature (80 - 105 °C) until TLC indicated consumption of the starting material. The indoles were visualised by spraying with van Urk's reagent (4- dimethylaminobenzaldehyde-HCI). After cooling to room temperature, the solvent was removed in vacuo. Flash chromatography of the residue on silica with the appropriate solvent afforded the product.

N-12-(indoly1-3)-1-methoxycarbonylethyll-2-quinaldic acid amide (2.114)

The title compound was obtained as yellow oil in a yield of 70 % (85 % on conversion) using tryptophan methyl ester trifluoroacetate as nucleophile and the general aminocarbonylation procedure (4 bar and 80 °C) as described above. [a]202 -7.42 (c, 1.0, CHCI3)

1 H 8: 3.48 (2H, m, 11-2'), 3.67 (3H, s, CO2013),- 5.16 (1H, m, H-1'), 7.0 - 7.34 (3H,

m, H-2", H-4" and H-5"), 7.31 (1H, d .17^,5^ = 6.6 Hz, H-7"), 7.50 - 7.74 (4H, m,

H- 5, H-6, H-7 and H-6"), 7.80 (1H, dd, .78,7 = 8.4 Hz, .18,5= 0.9 Hz, H-8), 7.92 - 8.02 (2H, m, 11-3 and 11-4), 8.52 (1H, br s, NH), 8.79 (114, d, J= 8.1 Hz, -NW

13C 5: 27.98 (C-2'), 52.34 (CO2CH3) 53.19 (C-1'), 109.82 (ipso, C-3"), 111.19 (C-7"), 118.56 (C-3), 118.61 C-6"), 119.41 (C-4"), 121.93 (C-5"), 122.93 (C-2"), 127.38 (C-6), 127.49 (C-5), 127.83 (C-3"a), 129.18 (C-8), 129.78 (C-7), 129.85 (C-4a), 136.08 (C-7"a), 137.25 (C-4), 146.32 (C-8a), 148.92 (ipso, C-2), 164.16 (-CONHCH2-), 164.26 (-CO2CH3) MS m/z: 373 (Mt, 5 %), 201 (Mt C9H6NICONH, 100 %), 128 (Mt CONHCH(CO2CH3)CH2C8H6N, 98 %) 179

N-12-(indoly1-3)-ethy11-2-quinaldic acid amide (2.115)

H

The title compound was obtained as white crystals in a yield of 87 % using tryptamine as nucleophile and the general aminocarbonylation procedure (4 bar and 80 °C) as described above.

Mp: 166 - 170 °C

1 1-1 8: 3.48 (2H, d = 6.9 Hz, H-2'), 3.85 (2H, m, H-1'), 7.06 - 7.22 (3H, m, H-2", H-

4" and H-5"), 7.37 (1H, dd Jr = 9.0 Hz, Jr ,5" = 0.9 Hz, H-7"), 7.57 (1H, dt J6",7"

= 6",5" = 6.9 Hz, J6",4"= 1.2 Hz, H-6"), 7.65 - 7.74 (3H, m, H-5, H-6 and H-7), 7.84 (1H, dd, .18,7 = 8.1 Hz, J8,6 = 0.9 Hz, H-8), 7.96 - 8.03 (2H, m, H-3 and H -4), 8.43 (2H, br s, 2 x NH)

13C 8: 25.59 (C-2'), 40.01 (C-1'), 111.17 (C-3"), 112.93 (C-7"), 118.68 (C-3), 118.72 (C- 6"), 119.24 (C-4"), 121.89 (C-5"), 122.09 (C-2"), 127.27 (C-6), 127.59 (C-5), 129.14 (C-3"a), 129.56 (C-8), 129.86 (C-7), 131.89 (C-4a), 136.32 (C-7"a), 137.29 (C-4), 146.31 (C-8a), 148.72 (ipso, C-2), 164.34 (-CONHCH2-) MS rtilz: 315 (Mt, 9 %), 128 (Mt CONHCH2CH2C8H6N, 100 %)

2-Butenoic acid, 8-quinolyl ester (2.116)

The title compound was obtained as clear viscous liquid in a yield of 87 % using 2.75 as the starting material and employing the general aminocarbonylation procedure (4 bar and 80 °C) as described above. Precipitation of the palladium occurred after 3 hours. 180

1 H 5: 1.88 (3H, dd = 6.9 Hz, 42' = 1.5 Hz, H-4'), 6.16 (1H, kd JrAi = 1.5 HZ, J2%3' =

15.3 Hz, H-2'), 7.17 (1H, kd = 6.9 Hz, J31,2' = 15.6 Hz, H-3'), 7.28 (1H, dd J3,2

= 4.2 Hz, J3A = 8.4 Hz, H-3), 7.34 (1H, dd J7,6 = 8.1 Hz, .17,5= 1.2 Hz, H-7), 7.41

(1H, t = J6,7 = 8.1 Hz, H-6), 7.59 (1H, dd, J5,6= 8.1 HZ, J5,7 = 1.5 Hz, H-5),

8.04 (1H, dd, J4,3 = 8.4 Hz, J4,2 = 1.5 Hz, H-4), 8.77 (1H, dd, J2,3 = 4.2 Hz, J2,4= 1.5 Hz, H-2)

13C 5: 18.35 (C-4'), 121.58 (C-3 and C-7), 121.86 (C-2'), 125.72 (C-5), 126.16 (C-6), 129.48 (C-4a), 135.98 (C-4), 141.27 (C-8a), 147.28 (C-8), 148.32 (ipso, C-2), 164.96 (-OCO(CH)2CH 3-) MS m/z: 213 (M+, 22 %), 128 (Mt CH3CHCHCOO, 100 %)

N-12-(indoly1-3)-1-methoxycarbonyl-ethy1J-8-benzyloxy-2-quinaldic acid amide (2.111b)

The title compound was obtained as light yellow oil in a yield of 60 % using tryptophan methyl ester trifluoroacetate as nucleophile and the general aminocarbonylation procedure (5 bar and 105 °C) as described above.

[a]e2 : -57.68 (c, 1.0, Me0H)

1 /1 5: 3.46 (2H, m, H-2'), 3.59 (1H, s, -COOCH3), 5.28 (2H, s, -CH2Ph), 6.95 (1H, d 2.4 Hz, H-2"), 5.99 - 7.19 (4H, m, H-7, H-7" and 2 aromatic protons), 7.32 -

7.60 (8H, m, 8 aromatic protons), 8.22 (1H, d J3,4 = 8.7 Hz, H-3) 8.27 (I H, d J4,3 = 8.4 Hz, H-4), 8.87 (I H, d J= 1.8 Hz, -NH)

13 C 5: 28.05 (C-2'), 52.29 (CO2CH3), 52.66 (C-1'), 70.99 (-CH2Ph), 109.69 (ipso, C-3"), 110.43 and 111.05 (C-7 and C-7"), 119.29 (C-3), 123.01 (C-2"), 118.59, 119.24, 119.95, 121.88, 127.63, 128.12, 128.21 and 128.67 (10 aromatic carbons), 127.19 (C-3a"), 130.62 (C-4a), 135.93 (C-7"a), 136.92 (ipso, C-1"'), 137.21 (C-4), 138.79 (C-8a), 148.04 (ipso, C-2), 154.82 (ipso, C-8), 164.21 (-CONH-), 171.95 (- CO2CH3) MS m/z: 480 (M+ + 1, 6 %), 262 (Mt NHCH(CO2CH3)CH2C8H61\11, 71 %) 181

N-12-(indoly1-3)-1-methoxycarbonyl-ethy11-6-methoxy-2-quinaldic acid amide (2.125)

Me0

The title compound was obtained as light yellow oil in a yield of 82 % using tryptophan methyl ester trifluoroacetate as nucleophile and the general aminocarbonylation procedure (4 bar and 105 °C) as described above.

-33.71 (c, 1.0, Me0H) 3.46 (2H, m, H-2'), 3.67 (3H, s, CO2CH3) 3.91 (3H, s, -OCH 3), 5.16 (1H, m, H-

1'), 7.03 (3H, m, H-5, H-2" and H-6"), 7.16 (1H, dt, ./5",4" ✓511,6" = 9.0 Hz, J5",7" -= 1.2 Hz, H-5"), 7.31 (111, d Jr,6" = 7.2 Hz, H-7"), 7.33 (114, dd J7,8 = 9.3 Hz, J7,5 =2.7 Hz, H-7), 7.63 (1H, d J4,5= 7.8 Hz, H-4"), 7.84 (1H, d 183 -= 9.3 Hz, H-8) 8.10 (1H, dJ4,3 = 8.4 Hz, H-4), 8.19 (1H, d J3,4= 8.4 Hz, H-3), 8.36 (1H, br s, - NH), 8.72 (1H, dJ= 8.1 Hz, -NH)

13C 8: 28.05 (C-2'), 52.35 (-CO2CH3), 53.15 (C-1'), 55.58 (-OCH3), 104.66 (C-5), 110.02 (ipso, C-3"), 111.16 (C-7"), 118.69 (C-4"), 119.01 (C-3), 119.48 (C-6"), 122.00 (C-5"), 122.89 (C-7), 123.05 (C-2"), 127.43 (C-4a), 130.56 (C-3"a), 131.23 (C-8), 135.69 (C-4), 136.07 (C-7"a), 142.45 (C-8a), 146.69 (ipso, C-2), 158.78 (ipso, C-6), 164.41 (-CONH-), 172.25 (-CO2CH3) MS m/z: 403 (Mt, 23 %), 201 (Mt CH(CO2CH3)CH2C8H6NI, 71 %), 186 alse - NHCH(CO2CH3)CH2C8H6NI, 27 %), 158 ((Mt CONHCH(CO2CH3)CH2C8H6N1, 68 %) 182

5.3.7 BISCHLER-NAPIERALSKI CYCLISATION

2-(3-methoxycarbonyl-P-carbolin-1-y0-quinoline (2.126)

5 4 3 2 N CO2Me 3' 4' H-N

6' A solution of amide 2.114 (150 mg, 0.402 mmol) and POCI3 (1 ml) in dry xylene (10 ml) was heated under reflux with stirring for 4 hours. The flask was cooled and the contents poured over crushed ice while stirring. The resulting solution was neutralised to pH = 8 with sodium carbonate followed by extraction with dichloromethane (3 x 30 ml portions). The combined organic fractions were dried over MgSO4 and the solvent removed in vacuo. The residue was purified by flash chromatography using hexane:EtOAc (1:1) as eluent to afford the title compound (31 mg, 23 %) as light yellow crystals. i°

Mp: 194 - 196 °C

1 1-1 8: 4.10 (3H, s, -0013),- 7.37 (1H,t 16', x' kJ' = 7.9 Hz, H-61), 7.58 - 7.86 (4H, m, H-

6, H-7, H-7',and H-8'), 7.96 (1H, d 15,6 = 8.2 Hz, H-5), 8.22 (1H, discs = 7.9 Hz,

H-5'), 8.35 (1H,d,18,7 = H-8), 8.52 (IH, d, 14,3 = 8.7 Hz, H-4), 8.94 (1H, s, H-4 1),

8.96 (1H, d 13,4 = 8.7 Hz, H-3), 12.00 (1H, br s, -NH) "C 8: 52.69 (-COOCH3), 112.32 (C-8'), 118.57 (C-4'), 119.63 (C-3), 120.90 (C-6'), 121.55 (C-5'a), 121.91 (C-5 1), 126.98 (C-6), 127.93 (C-5), 128.05 (C-4a), 128.97 (C-7), 129.09 (C-8), 129.75 (C-7'), 136.78 (C-4), 130.66, 136.62, 136.84, 137.41 and 140.94 (5 quaternary carbons), 147.15 (C-2), 157.42 (C-8a), 166.62 (- COOCH3) MS m/z: 353 (Mt, 3 %), 352 (Mt H, 100 %), 294 (Mt COOCH3, 68 %), 183

5.3.8 PREPARATION OF LAVENDAMYCIN PENTACYCLIC SYSTEM

N-12-(indoly1-3)-1-methoxycarbonyl-ethy11-8-methyloxy-2-quinaldic acid amide (2.111a)

H 2'

OMe

To a 50 ml stainless steel pressure vessel equipped with a magnetic stirrer bar were added sequentially: the 2-chloroquinoline derivative 2.86 (80 mg, 0.413 mmol), DMF (4 ml), PdC12(PPh3)2 (36 mg, 0.0413 mmol), tryptophan methyl ester trifluoroacetate (150mg, 0.455 mmol) and triethylamine (0.22 ml, 1.65 mmol). The system was sealed, cooled down with liquid nitrogen, evacuated, purged with argon and charged with CO (10 bar). The mixture was heated to 105 °C for 11 hours. After cooling to room temperature, the solvent was removed in vacuo. Flash chromatography of the residue on silica with hexane:EtOAc (1:1) afforded the product (150 mg, 89 %) as orange brown oil.

[a]p -21.51 (c, 1.0, Me0H)

1 1-1 8: 3.43 (2H, dd Jrt = 6.0 Hz, J = 1.8 Hz, H-2'), 3.61 (3H, s, CO2CF13), 3.96 (3H, s, OCH3), 5.06 (I H, dt Jp,2. = 7.8 Hz, JI',N14 = 6.0 Hz, H-1'), 7.96 - 7.09 (3H, H-2", H-5" and H-7), 7.11 (1H, d J= 2.1 Hz, H-2"), 7.25 (IH, d J7",6" = 8.7 Hz, H-7"), 7.31 (1H, d J41,51= 8.1 Hz, H-4"), 7.45 (1H, t, J6,7 = J6,5 = 7.8 Hz, H-6), 7.60 (IH, d J5',61= 7.5 Hz, H-5'), 8.14 (1H, br s, NH), 8.15 (1H, d J4,3 = 8.4 Hz, H-4), 8.21 (1H, d J.3,4= 8.4 Hz, H-3), 8.83 (1H, d, J= 7.8 Hz, -NH)

13 C 8: 27.95 (C-2'), 52.18 (CO2CH3), 53.14 (OCH 3), 55.92 (C-1'), 108.32 (C-7), 109.66 (ipso, C-3"), 111.09 (C-7"), 118.37 (C-5), 119.05, 119.09 and 119.19 (C-3, C-4" and C-6"), 121.61 (C-5"), 123.07 (C-2"), 128.19 (C-6), 131.75 and 131.88 (C-4 and C-3"a), 136.09 (C-7"a), 137.03 (C-4), 138.29 (C-8a), 147.70 (ipso, C-2), 155.39 (ipso, C-8), 164.30 (-CONHCH2-), 172.15 ( -CO2CH3) MS m/z: 404 (M'+ 1, 12 %), 201 (Mt (OCH3)C9H6NICONH, 26 %), 186 (Mt NHCH(CO2CH3)CH2C8H6N, 9 %), 158 (Mt CONHCH(CO2CH3)CH2C3H6N, 32 %) 184

2-(3-methoxycarbonyl-(3-carbolin-1-y1)-8-methoxyquinoline (2.112a)

5 4 3 2 COMB 8 N 3' OMe 4' H-N

7. 6

A solution of amide 2.111a (83 mg, 0.204 mmol) and POCI3 (0.5 ml) in thy xylene (5 ml) was heated under reflux with stirring for 2 hours. The flask was cooled and the contents poured over crushed ice while stirring. This solution was neutralised to pH = 8 with sodium carbonate followed by extraction with dichloromethane (3 x 30 ml portions). The combined organic fractions were dried over MgSO4 and the solvent removed in vacuo. The residue was purified by flash chromatography using hexane:EtOAc (3:1) as eluent to afford the title compound (18 mg, 23 %) as light yellow crystals. The yield of this reaction was improved to 60 % by stopping the reaction at 30 % conversion and recycling of the starting material.

Mp: 140 °C

1 1-1 8: 4.10 (3H, s, -0Q113), 4.26 (3H, s, COOCL13), 7.14 (1H, dd J7,6 = 7.8 Hz, J7,5 = 1.5 Hz, H-7), 7.35 (1H, ddd J67 = J6',51 = 8.1 Hz, J6 ,8 = 1.2 Hz, H-6'), 7.41 - 7.70 (4H, m, H-5, H-6, H-7',and H-8'), 7.23 (1H, d =9.0 Hz, H-55, 8.32 (1H, dJ4,3 = 8.7 Hz, H-4), 8.95 (11-1, d J3,4 = 8.7 Hz, H-3), 8.96 (1H, s, H-4'), 12.50 (1H, br s, -NH) "C 8: 52.69 (-COOCH3), 56.29 (-OCH3), 107.82 (C-7), 112.33 (C-85, 118.57 (C-4'), 119.49 (C-3), 119.55 (C-5), 120.68 (C-65, 121.61 (C-5'a), 121.93 (C-5'), 127.18 (C-6), 128.75 (C-75, 128.82 (C-4a), 136.43 (C-4), 138.75 (C-3'), 130.45 133.13, 136.53, 136.84, 137.85 (5 quaternary carbon signals), 141.29 (C-8a), 155.54 (C- 8), 166.72 (-COOCH3) MS m/z: 384 (M± + 1, 45 %) 185

5.4 PREPARATION OF MELATONIN AND DERIVATIVES

5.4.1 SYNTHESIS OF CYCLIC ENAMIDES VIA OXIDATION OF PYRROLIDINE

Pyrrolidine trimer (3.22) 1 0

6 A 25 % aqueous solution of sodium peroxodisulfate (6.66g, 28 mmol) was added dropwise to a stirred mixture (below -10 °C) of pyrrolidine (2.3 ml, 28 mmol), sodium hydroxide (2.24 g, 56 mmol) and silver nitrate (23.8 mg, 0.14 mmol) in water (30 ml) and the mixture stirred for 1.5 hours. The reaction mixture was then saturated with sodium chloride, extracted with dichloromethane( 3 x 60 ml portions) and the extract dried over MgSO4 and concentrated under reduced pressure on an ice-bath. Ether (5 ml) was added and the precipitate filtered off. The filtrate was evaporated under reduced pressure at 0 °C to give the product (1.08 g, 56 %) as a faint orange oil which could be used without further purification."

1 1-1 8: 1.75 (12H, m, H-1, H-2, H-5, H-6, H-9, H-10) , 2.25 (3H, m, H-4a, H-8a, H-12a), 2.95 (6H, m, H-3, H-7, H-11)

I3 C 8: 20.25 (C-2, C-6, C-10), 27.86 (C-1, C-5, C-9), 45.84 (C-3, C-7, C-11), 81.96 (C- 4a, C-8a, C-12a)

General procedure for the preparation of N-acyl-2-pyrrolines

A solution of the trimer 3.22 (7.2 mmol) dissolved in benzene (30 ml) was added dropwise to a refluxing solution of acid chloride (21.7 mmol) and Hiinigs base (N,N,N-ethyldiisopropylamine) (2.8g, 21.7 mmol) in benzene (10 ml). After 20 minutes the reaction mixture was allowed to cool to room temperature, diluted with dichloromethane and washed with water. The organic phase was dried (MgSO4) and the solvent removed in vacuo. Flash chromatography on silica gel (deactivated with triethylamine) using ethyl acetate and hexane (1:1) afforded the product. 186

N-Acetyl-2-pyrroline (3.17)

COCH3 The title compound 12 was obtained as a colourless oil (58 %) using the general procedure described above.

1 H 5: (exists as a rotameric pair) 1.99 and 2.04 (3H, s, OCH3), 2.56 (2H, m, H- 4), 3.76 (2H, t J5A = 9.3 Hz, H-5), 5.14 (1H, m, H-3), 6.37 and 6.85 (1H, m, H-2)

D C 8: 21.56 and 22.06 (-COCH3), 28.25 and 29.96 (C-4), 44.58 and 45.35 (C-5), 110.97 and 111.66 (C-3), 128.97 and 129.29 (C-2), 166.17 (-COCH3) HREI-MS m/z: 111.0682 (C6H9NO requires 111.0684)

N-benzoyl-2-pyrroline (3.25)

2

COC6H5

The title compound" was obtained as a colourless oil in a yield of 67% using the general procedure as described above.

1 11 5: (main isomer of the rotameric pair) 2.84 (2 H, m, H-4), 3.99 (2H, t J5A = 9.3 Hz, H-5), 5.12 (1H,m, H-3), 6.42 (1H, m, H-2 ), 7.39 - 7.46 (5H, m, aromatic protons).

13 C 8: 28.25 (C-4), 44.1 (C-5), 110.697 (C-3), 127.66 and 128.39 (5 aromatic carbons), 130.65 (C-2), 135.81 (C-1 1), 166.94 (-COPh) HREI-MS m/z:173.0845 (CI IIIHNO requires 173.0840) MS m/z: 173 (Mt, 67 %), 105 (M 4- C4H6N, 100 %), 77 (Mt C4H6N - CO, 74 %) 187

5.4.2 PREPARATION OF MELATONIN AND OTHER INDOLES USING ENAMIDES

General procedure for the preparation of indoles

A solution of enamide (0.21 mmol) and the required phenylhydrazine (0.24 mmol) in the appropriate solvent (1.5 - 5 ml) with catalyst (given at each compound's NMR data) was refluxed for 20 minutes, allowed to cool to room temperature, diluted with dichloromethane and washed with water. The organic layer was washed with saturated NaHCO3 solution, dried (MgSO4) and evaporated in vacuo. Flash chromatography on silica gel (deactivated with triethylamine) using an appropriate solvent afforded the products.

Melatonin (N42-(5-methoxy-1H-indo1-3-yHethyliacetamide) (3.1)

The title compound was obtained as white crystals in a yield of 75 % using the general procedure for the preparation of indoles from cyclic enamides. The catalyst and solvent that was employed in this case was a 25:35: 40 mixture of AcOH:EtOH:H20.

Mp: 112°C (lit. 14 116- 118 °C)

1 1-1 5: 1.90 (3H, s, -COOCH3), 2.91 (3H, t J2,1 = 6.6 Hz, H-2), 3.56 (3H,dt J1,NH = 6.3 Hz, .11,2 = 6.6 Hz, H-1), 3.83 (1H, s, 5'-OCH3), 5.61 (1H, br s, -NH), 6.84 (1H, dd

6',7' = 8.7 Hz, 4,4. = 2.4 Hz, H-6'), 6.83 (1H, d = 2.4 Hz, H-2'), 6.98 (1H, d J4',61 = 2.4 Hz, H-4'), 7.24 (1H, d = 8.7 Hz, H-7'), 8.18 (1H, br s, -NH)

13 C 5: 23.31 (-COCH3), 25.18 (C-2), 39.66 (C-1), 55.87 (-OCH3), 100.39 (C-4'), 112.43 and 112.04 (C-6' and C-7'), 112.62 (C-3'), 122.84 (C-2'), 127.73 (C-3a'), 131.57 (C-Ta), 154.09 (C-5'), 170.25 (-COCH3) MS m/z: 232 (Mt, 41 %), 173 (Mt - NH2COCH3, 100 %) HREI-MS m/z:232.12189 (C13Hi6N202 requires 232.12119) 188

N-I2-(5-methoxy-1H-indo1-3-ypethylIbenzamide (3.28)

The title compound was obtained as a brown liquid in a yield of 85 % using the general procedure for the preparation of indoles from cyclic enamides. The catalyst and solvent that was employed in this case was a 25:35: 40 mixture of AcOH:Et0H:H20.

i ll 8: 3.04 (2H, d J2,1 = 5.1 Hz, H-2), 3.74 (3H, s, -OCH3), 3.76 (2H, m, H-1), 6.41 (1H, br s, NH), 6.83 (1H, dd = 8.7 Hz, f6',4' = 2.4 Hz, H-6'), 6.98 (1H, d J2',NH = 2.1 Hz, H-2'), 7.02 (1H, dJ4,6 = 2.4 Hz, H-4 1), 7.24 (1H, d = 8.7 Hz, H-7'), 7.18 - 7.68 (5H, m, 5 overlapping aromatic resonances), 8.39 (1H, s, NH)

13 C 8: 25.13 (C-2), 40.35 (C-1), 55.69 (-0013), 100.28 (C-4'), 112.13 and 112.41 (C-6 1 and C-7'), 112.51 (C-3 1), 122.99 (C-2 1), 126.85 and 128.59 (5 aromatic carbons), 128.37 (C-3a1), 131.59 (C-7'a), 134.49 (C-1"), 154.01 (C-5'), 167.73 ( -00C6H5) MS m/z: 294 (Mt 10 %), 173 (M+ - NH2COC6H5, 100 %) HREI-MS m/z:294.13280 (C181-118N202 requires 294.13684)

N-12+1H-indol-3-yflethyllacetamide (3.26)

N yC H3 0

The title compound was obtained in a yield of 62 % using 1.2 mole equivalents of ZnC12 in xylene.

1H 8: 1.89 (3H, s, -COOCH3), 2.95 (3H, t J2,, = 6.6 Hz, H-2), 3.57 (3H,dt J1,NH = 6.6 Hz, 42 = 6.6 Hz, H-1), 5.58 (1H, br s, -NH), 7.01 (1H, d J2',NH = 2.4 Hz, H-2 1), 7.11 (1H, dd./67 = 6.9 Hz, J6',4' = 1.5 Hz, H-6'), 7.19 (1H, dd = 6.9 Hz, J5.,6: = 1.5 Hz, H-5'), 7.35 (1H, dd J41,6" = 1.2 Hz, 4%5• = 8.1 Hz, H-4'), 7.24 (1H, dd ./7,6' = 8.1 Hz, kr 1.2 Hz, H-7'), 8.28 (1H, br s, -NH) 189

13C 8: 23.29 (-COCH3), 25.17 (C-2), 39.77 (C-1), 111.32 (C-7'), 112.93 (C-3'), 118.69 (C-6'), 119.49 (C-4'), 122.09 and 122.20 (C-2' and C-5'), 127.37 (C-3a 1), 136.47 (C-Ta), 170.25 (-COCH3) HREI-MS m/z:202.10996 (C121-114N20 requires 202.11061)

N-12-0H-indo1-3-ypethylibenzamide (3.27)

The title compound was obtained in a yield of 65 % using 1.2 mole equivalents of ZnCl2 in xylene.

1 H 8: 3.07 (3H, t J2,, = 6.6 Hz, H-2), 3.77 (3H,dt ANH = 6.6 Hz, J1,2= 6.6 Hz, H-1 ), 6.35 (1H, br s, -NH), 7.00 (1H, d J2.,Nu = 2.4 Hz, H-2'), 7.11 (1H, dd J6,2. = 6.9

Hz, J6.,4' = 1.5 Hz, H-6 1), 7.19 (1H, dd i51,7 = 6.9 Hz, J5',6' = 1.5 Hz, H-5'), 7.30 - 7.39 (4H, m, 3 overlapping aromatic resonances and H-4'), 7.61 (1H, dd thig = 8.1 Hz, J7',5,= 1.2 Hz, H-7'), 7.63 - 7.68 (2H, m, 3 overlapping aromatic resonances), 8.43 (1H, br s, -NH)

13 C 8: 25.16 (C-2), 40.23 (C-1), 111.36 (C-7'), 112.76 (C-3'), 118.68 (C-6'), 119.42 (C- 4'), 122.14 and 122.22 (C-2' and C-5'), 126.85, 128.59 and 131.37 (5 aromatic CH carbons), 127.29 (C-3a'), 134.62 (C-1"), 136.48 (C-7'a), 167.65 (-00061.15) MS m/z: 264 (M+, 7 %), 173 (M+ - NH2COC6H5, 100 %) HREI-MS m/z:202.12574 (C121-114N20 requires 202.12625) 190

N-[2-(5-bromo-1H-indo1-3-yBethylIbenzamide (3.29)

Br

The title compound was obtained in a yield of 30 % using AcOH/2% Ac20 as the catalyst.

1 1-I 8: 3.02 (3H, t /2,1 = 6.6 Hz, H-2), 3.75 (3H,dt JI,NH = 6.0 Hz, J1,2 = 6.0 Hz, H-1),

6.24 (I H, br s, -NH), 7.04 (1H, d Jz,NH = 2.4 Hz, H-2'), 7.20 -7.69 (8H, 3 x m, H- 4', H-6', H-7' and 5 overlapping aromatic resonances), 8.34 (1H, br s, -NH)

13C 8: 25.09 (C-2), 40.28 (C-1), 112.80 (C-4'), 112.76 (C-7'), 121.36 (C-2'), 123.41 (C- 6'), 125.08 (C-4'), 126.85, 128.62 and 131.46 (5 aromatic CH carbons), 128.29 (C-5a'), 129.18 (C-3'a), 134.61 (C-1"), 135.05 (C-7'a'), 167.67 (-0006F15) MS m/z: 342/344 (Br79 , Br81) (Mt, 32 %), 221/223 (Mt - NH2COC6H5, 100 %) HREI-MS m/z:342.03490 (Cl2H15N20Br79 requires 342.03677) and 344.02675 (Cl2H15N20Br 81 requires 344.03085)

5.4.3 PREPARATION OF N-ACYLCARBINOLAMINES

(S) - tert- Butyl-2-pyrrolidone-5-carboxylic acid (3.38)

CO2C(CH3)3 H

To a solution of L-pyroglutamic acid (500 mg, 3.85 mmol) in tert- butyl acetate (4 ml) was added perchloric acid (60 %, 1 ml) at room temperature. The reaction mixture was stirred for 3 days at this temperature, neutralised with sodium bicarbonate and extracted with ether (3 x 25 ml portions). The combined organic fractions were dried over MgSO4 and evaporated in vacuo to afford the essentially pure product as white crystals (320 mg, 45 %).

Mp: 91 °C (lit." 93 °C) 191

1 14 5: 1.45 (9H, s, C(CH3)3), 2.18 - 2.46 (4H, 2 x m, H-3 and H-4), 4.11 (1H, m, H-5), 6.49 (1H, br s, NH)

13 C 8: 24.71 (C-4), 27.84 (-C(CH 3)3), 29.31 (C-3), 56.06, (C-5), 82.29 (-C(CH3)3), 171.19 (-CONH-), 178.18 (-COOC(CH3)3) MS m/z: 186 (Ise + 1, 73 %), 130 (M+ + 1 - C(CH3)3, 45 %) HREI-MS mh:185.10936 (C6H15NO3 requires 185.10519)

General procedure for the N-protection of pyrrolidones

To a solution of the required pyrrolidone (11.8 mmol) and sodium hydride (12.9 mmol) in tetrahydrofuran (3 ml) at -50°C was added benzyloxychloroformate (12.9 mmol) and stirred for 6 hours. After extractive work up with water and dichloromethane, the organic layer was dried (MgSO4) and evaporated in vacuo to provide a crude reaction mixture which could be used without further purification.

N-benzyloxycarbonyl-2-pyrrolidone (3.33)

o

CO2CH2C6H5

The title compound was obtained as a colourless oil in a quantitative yield using the general procedure as described above.

I I-I 8: 2.01 (2H, tt J4 , 3 = J4,5 = 7.2 Hz, H-4), 2.51 (2H, t J3,4 = 8.4 Hz, H-3), 3.79 (2H, t 15,4 = 6.9 Hz, H-5), 5.23 (2H, s, -CH2C6H5), 7.25 - 45 (5H, m, -CH2C6I15) D C 8: 17.44 (C-4), 32.69 (C-3), 46.34 (C-5), 67.93 (-CH2C6H5), 128.23, 128.41 and 128.60 (5 aromatic carbons), 135.37 (phenyl ring junction carbon), 151.54 (- CONH-), 174.09 (-COOCH2C6H5) MS mfr: 219 (Mt, 83 %), 107 (100 %) HREI-MS mh:219.08558 (C121413NO3 requires 219.08954) 192

(S)-tert- Butyl-(N-benzyloxycarbonyl)-2-pyrrolidone-5-carboxylic acid (3.39)

0 N CO2C(043) 3 CO 2CH2C6H5

The title compound was obtained as a colourless oil in a yield of 85 % using the general procedure for the protection of 2-pyrrolidones as described above.

1 H 8: (main isomer of pair of rotamers) 1.36 (9H, s, C(CH3)3), 2.00 - 2.39 (2H, m, H-4), 2.46 - 2.61 (2H, m, H-3), 4.52 (1H, dd J= 9.3 Hz, J= 2.7 Hz, H-5), 5.24 (2H, dd J = 5.1 Hz, -CH2C6H5), 7.25 - 45 (5H, m, -0-12C635)1 13C 8: 21.76 (C-4), 27.66 (-C(CH3)3), 30.91 (C-3), 59.06 (C-5), 68.16 (-CH2C6H5), 82.29 (-C(CH3)3), 128.42, 128.56 and 128.72 (5 aromatic carbons), 135.02 (phenyl ring junction carbon), 150.91 (-COCH2C6H5), 170.11 (-COOC(CH3)3), 173.27 (-CH2CONH-) HREI-MS m/z:319.14216 (CI7H2IN05 requires 319.14197)

General procedure for the preparation of N-acylcarbinolamines

A portion of the crude N-protected pyrrolidone (0.84 mmol) was dissolved in dry THF (2 ml), cooled to -78 °C and followed by the addition of 1.0 M solution of lithium triethylborohydride in THF (1m1, I mmol). After 1 hour the reaction mixture was quenched with a saturated aqueous NaHCO3-solution (2 ml), allowed to warm to room temperature and extracted with dichloromethane. The combined organic layers were dried over MgSO4, filtered and concentrated to furnish the essentially pure products.

N-(benzyloxycarbony1)-2-hydroxy-pyrrolidine (3.34)

OH N CO2CH2C6H5 The title compound was obtained as a colourless oil in a yield of 84 % using the general procedure as described above. 193

1 11 8: (main isomer of rotameric pair) 1.85 (2 H, m, H-4), 3.33 (2H, m, H-3), 4.25

(21-1, H-5), 5.12 (21-I, s, -C112C6H5), 5.48 (1H, s, H-2), 7.33 (514, m, CH2C5 15)1 13 C 5: 22.59 (C-4), 32.74 (C-3), 45.67 (C-5), 67.02 (-CH2C6H5), 81,89 (C-2), 127.85, 128.01 and 128.19 (5 aromatic carbons), 136.37 (phenyl ring junction carbon), 155.54 (-COOCH2C61-10 MS m/z: 221 (Mt, 0.7 %), 203 (Mt - H2O, 100 %) HREI-MS m/z: 221.10515 (C121-121N05 requires 221.10519)

(S)-tert-Butyl-IN-(benzyloxycarbony1)-2-hydroxy-pyrrolidiny11-5-carboxylic acid (3.40)

OH N CO2C(CF4 )3 CO2CH2C6H5

The title compound was obtained as a colourless oil in a yield of 91% using the procedure as described above.

[a] 2: -48.9 (c, 1.0, CHCI3)

1 1-1 8: (acetone- d6,) (main isomer of rotameric pair) 1.34 (9H, s, -C(CLI3)3), 1.81 - 1.94 (2H, m, H-4), 2.47 (2H, m,14-3), 2.88 (1H,br s, -OH), 4.29 (1H, m, H-5), 5.06 s, -CHC6H5), 5.59 m, H-2), 7.34 (5H, m, 5 aromatic protons)

13C 5: (acetone- d6,) 21.76 (C-4), 27.94 (-C(CF13)3), 32.33 (C-3), 60.34 (C-5), 67.12 (- CH2C6H5), 82.89 (-C(CH3)3), 128.53, 128.68 and 129.19 (5 aromatic carbons), 137.85 (phenyl ring junction carbon), 150.91 (-COCH2C6H5), 170.11 (- COOC(CH3)3), 173.27 (-C(OH)NH-) HREI-MS m/z: 321.16120 (C171-1231\105requires 321.15762) 194

5.4.4 APPLICATION OF N-ACYLCARBINOLAMINES IN THE SYNTHESIS OF INDOLES

General procedure for the preparation of indoles using N-acylcarbinolamines

A solution of p-methoxyphenylhydrazine hydrochloride (0.58 mmol) and the appropriate N- acylcarbinolamines (0.94 mmol) in an acetic acid/water/ethanol mixture (25:40:35) (3m1) was refluxed for 35 minutes. After extractive work up with water and dichloromethane, the organic layer was washed with a saturated NaHCO3 solution, dried (MgSO4) and evaporated in vacuo. Flash chromatography on silica gel (deactivated with triethylamine) using ethyl acetate-hexane (1:1) furnished the pure indoles.

N-12-(5-methoxy-1H-indo1-3-yDethyllbenzyloxycarboxamide (3.36)

H 0 N.O.A.„-C6H5

The title compound was obtained as an orange oil in a yield of > 95%.

1 H 3: 2.94 (3H, t J2,1 = 6.6 Hz, H-2), 3.52 (3H,dt ANH = 6.3 Hz, J1,2 = 6.6 Hz, H-1), 3.84 (1H, s, 5'-OCH3), 4.96 (1H, br s, -NH), 5.11 (2H, s, -CHC6H5), 6.87 (1H, dd J6,71= 8.7 Hz, J6,4' = 1.8 Hz, I4-6'), 6.93 (1H, s, H-2'), 7.00 (1H, s, 11-4'), 7.23 (1H, d J7,6' = 8.7 Hz, H-7'), 8.21 (111, br s, -NH)

13 C 5: 25.56 (C-2), 41.12 (C-1), 55.77 (-0CH3), 66.52 (-CH2C6H5), 100.37 (C-4'), 112.98 and 112.25 (C-6' and C-7'), 111.98 (C-3'), 122.93 (C-2'), 127.56 (C-3a'), 128.06 and 128.48 (5 aromatic carbons), 131.52 (C-7'a), 136.54 (phenyl ring junction carbon), 153.93 (C-5'), 156.48 (-FINCOC6H5), MS m/z: 324 (Mt 80 %), 173 (M± - H2NCH2CO2C6H5) HREI-MS m/z: 324.1466(C19H213N203 requires 324.1474)

195

(S)-tert-Buty1-11-(5-methoxy-1H-indo1-3-y1)-2-N-(benzyloxycarbamoyl)propanoate (3.42)

Me0 H 0 3 ...k.„-06H5

2' 002C(CH3)3 H The title compound was isolated as an orange oil in a yield of > 95 %.

[a]2: -11.7 (c, 1.0, CHC13)

I I-1 8: 1.35 (9H, s, C(CH3)3), 3.21 (3H, m, H-2), 3.78 (1H, s, 5'-OCH3), 4.59 (3H, m, H-

1), 5.07 (2H, s, -CHC6H5), 5.31 (1H, d, -NH), 6.82 (I H, dd = 8.7 Hz, f6',4' = 2.4 Hz, H-6), 6.95 (1H, d ./21,NH = 2.4 Hz, H-2D, 7.02 (1H, d = 2.1 Hz, H-4'), 7.23 (1H, d = 8.7 Hz, H-7'), 7.30 (51-1, m, -CH2C6H5), 7.96 (1H, br s, -NH) "C 8: 27.74 (C-3), 27.86 (C(CH3)3), 54.97 (C-3), 55.78 (-OCH3), 66.80 (-CH2C6H5), 82.00 (C(CH3)3), 100.63 (C-4D, 110.16 (C-3D, 111.84 and 112.56 (C-6 1 and C-7'), 122.45 (C-2'), 127.80 (C-7'a), 128.14 and 128.51 (5 aromatic carbons), 131.20 (C- 3'a), 136.38 (phenyl ring junction carbon), 154.21 (C-5'), 155. 85 (-HNCOC6H5), 171.12 (COOC(CH3)3) MS m/z: 324 (M+, 13 %), 273 (M+ - H2NCH2CO2C6H5, 4 %), %), 248 (M + - H2NCH2CO2C6H5- C(CH3)3, 1 %), 217 (M + - H2NCH2CO2C6H5 - C(CH3)3 - 0Me, 4 %), 160 (M+ - C6H5CO2CH2NHCHCOOC(CH3)3 + H, 100 %), HREI-MS m/z: 424.1989 (C24H28N205 requires 424.1998)

Dehydration of 3.34 to form cyclic enecarbamate 3.35

A mixture of 3.34 (210 mg, 0.93 mmol), acetyl chloride (360 mg, 4.64 mmol) and Hiinig's base (600 mg, 4.64 mmol) in benzene (5m1) was refluxed for 15 minutes, evaporated in vacuo and flash chromatographed on silica (deactivated with triethylamine) using ethyl acetate hexane (1:1) to yield the product as a colourless oil (20 mg, 98 %). 196

N-(benzyloxycarbonyI)-2-pyrroline (3.35)

CO2CH2C6H5

1 H 5: 2.63 (2 H, m, 14-4), 3.76 (2H, m, f1-5), 5.07 (111, m, H-3), 5.15 (2H, s, CH2C6H5), 6.52 (1H, td J= 1.8 and 3.9 Hz, H-2), 7.34 (511, m, CH2C6H5)

13 C 8: 28.54 (C-4), 45.15(C-5), 66.97(CH2C6H5), 108.81 (C-3), 127.88 and 128.50 (5 aromatic carbons), 128.08 (C-2), 136.56 (phenyl ring junction carbon), 163.82 (COCH2C6H5) HREI-MS m/z: 203.0899,(C12HI3NO2 requires 203.0946)

5.5 STRUCTURE ELUCIDATION OF NATURAL PRODUCTS FROM SELECTED ALOE SPECIES

5.5.1 ISOLATION OF PLICATOLOSIDE FROM ALOE PLICATILIS

Leaf exudate samples were collected near Worcester in the south-western Cape. Copious amounts of a watery exudate, which soon solidified to a pale yellow wax-like solid, was taken from three different plants, and also from leaves of various ages (old, medium and young). Analytical studies showed that the three different individuals and the various leaf samples were all practically identical. The same applies to several extracts obtained from cultivated specimens. In view of the striking morphological features of A. plicatilis, which makes it impossible to be mistaken for any other species, we deemed it unnecessarily destructive to collect a voucher specimen.

The yellow wax-like solid from Aloe plicatilis was treated with methanol and ether (2:1) and stirred for thirty minutes at 0 °C. After filtration the solvent was removed in vacuo to furnish a brown residue which contained one major compound with RI' = 0.53 (Et0Ac-Me0H-H20 - 7:2:1). This (1 g) was adsorbed on silica gel and flash chromatographed on silica gel (100 g) at low temperature (-15° C) in EtOAC:MeOH:H20 (8:1:1). The separation was monitored by TLC and analytical HPLC. 197

2,8-0,0-di-(0-D-glucopyranosy0-1,2,8-trihydroxy-3-methylnapthalene (4.1)

Mp: White crystals, 31%, 197-199°C

[a]ii - 199.1° (DMSO, c 1.0) UV nm (log 8) 284 sh (3.98), 305 (4.05), 318 (3.97), 334 (3.89)

I li 6: see Table 4.1 '3C 6: see Table 4.1 EIMS m/z: 190.0634 (C11H10 03 requires 190.0630) FABMS m/z: 515 [M+1] +

Acetylation of plicataloside 4.1

Plicataloside 4.1 (100 mg), acetic anhydride (0.5 ml) and pyridine (0.5 ml) in CH2Cl2 (3 ml) were stirred for 70 hours at room temperature. The solvent was removed under reduced pressure and the crude product chromatographed in EtOAC: CHC13 (1:1). Reverse phase HPLC of the eluate indicated the presence of two products which were not separated by chromatography on silica gel. Separation was achieved by preparative reverse phase HPLC on a Beckman Ultrasphere ODS column (C18 reverse phase, 5 mm particle size, 250 x 10 mm; flow rate 4,5 ml.mind ; 1 ml sample loop). The solvent system comprised 55% MeCN and 45% H2O (isocratic). Detection was by diode array detector, using 2 channels (A set at 275 ± 35 nm; B set at 365 ± 20 nm). Pure 4.2 and 4.3 were obtained (for NMR data, see Table 4.2). The same mixture of structures was obtained on acid (H2SO4) catalysed acetylation of plicataloside 4.1 in acetic anhydride at room temperature for 70 hours.

5.5.2 ISOLATION OF NEW COMPOUNDS FROM A. AFRICANA, A. SPECIOSA AND A. BROOMII

Leaf exudate samples from Aloe africana were collected near Ann's Villa and from A. speciosa in the Zuurberg (both on the road between Paterson and Cookhouse in the Eastern Cape) and the A. broomii samples came from a locality 130 km south of Bloemfontein. Since these characteristic species were identified in situ by the Aloe taxonomist Dr Ben-Erik van Wyk, no voucher specimens were collected.

198

The exudate samples were pre-cleaned using C,8 cartridges, with Me0H as solvent. They were then dissolved in a Me0H-H20 (1:1) mixture and studied by HPLC using a Phenomenex IB-Sil C 18 reverse phase column (5mm particle size, 250 x 4.6 mm internal diameter for analytical runs; flow rate 1 ml min'; 20 ml sample loop). The solvent system was a 30% to 60% linear gradient of Me0H in H2O over 25 min, 3 min isocratic, 100% in 2 min, 4 min isocratic. A diode array detector with two channels (A set at 275 ± 70 nm; B set at 365 ± 40 nm) was used.

5.5.2.1 CONSTITUENTS FROM A. AFRICANA

The yellow exudate was dissolved in Me0H. After filtration, the solvent was removed in vacuo to provide a brown residue which contained one unknown compound with Rf 0.33 (MeOH- CHC13, 2:8 ). Analysis of the sample with the analytical HPLC system revealed the presence of aloesin 4.4, aloin A and B 4.5, aloinoside A and B and 4.6 (Itt 21.48). A portion (1g) of the residue was flash chromatographed on silica gel (100 g) at low temperature (-15°C) in Me0H- CHC13 (1:9). The separation was monitored by TLC and analytical HPLC.

(E)-2-Acetony1-8-(2LO-feruloy1)-13-D-glucopyranosy1-7-methoxy-5-methyl-chromone (4.6)

Amorphous solid (7.8%) UV Xmir nm: 326 sh, 298, 238 sh, 217

8: 2.26 (3H, s, 11-Me), 2.68 (3H, s, 5-Me), 3.36 (2H, 2 x m, H-4' and H- 5'), 3.42 (1 H, m, H-6'a), 3.50 (1H, m, H-3'), 3.73 (1H, m, H-6'6), 3.77 (3H, s, 7- OMe), 3.80 (2H, s, H-9), 3.84 (3H, s, 6"-OMe), 4.4 (1H, t, J= 5.8 Hz, 6-0L1), 4.94 (1H, d, J= 10.1 Hz, H-1'), 5.13 (1H, d, J= 4.9 Hz, 4'-OH), 5.21 (1H, d, J- 5.4 Hz, 3'-Oth, 5.47 (1H, t, J=9.5 Hz, H-2'), 6.17 (1H, d, J= 15.9 Hz, H-2"), 6.18 (1H, s, H-3), 6.74 (1H, d, J= 8.2 Hz, H-8"), 6.86 (1H, s, H-6), 7.01 (1H, dd, J5^,9" = 1.9 Hz, fry = 8.2 Hz, H-9"), 7.2 (1H, d, J= 1.9 Hz, H-5"), 7.27 (1H, d, J = 15.9 Hz, H-3"), 9.52 (1H, s, 7"-OH). '3C 8: see Table 4.3 ES-MS m/z: 583 ([M - I f, 100 %), 407 (M" - COCH=CHC6H3(OCH3)(OH), 31 %), 389 ([M - COCH=CHC6H3(OCH3)(OH) - H2O, 7 %) 199

5.5.2.2 CONSTITUENTS FROM A. SPECIOSA

The yellow exudate was dissolved in Me0H. After filtration, the solvent was removed in vacuo to provide a yellow brown residue which contained one major unknown compound with Rf 0.65 (Me0H-CHC13 1:2). Analysis of the sample with the analytical HPLC system revealed the presence of homonataloin and 4.8 (R, 28.75). A portion (I g) of the residue was adsorbed onto silica gel and flash chromatographed on silica gel (100 g) at low temperature (-15°C) in Me0H- CHC13 (1:10). The separation was monitored by TLC and analytical HPLC.

(E)-2-Acetonyl-(2',6'-di-0,0-coumaroy1)-13-D-glucopyranosyl-7-hydroxy-5- methylchromone (4.8)

Amorphous solid (14%). UV Xmth" nm: 255 sh, 305.

I li 8: 2.24 (3H, s, 11 - Me), 2.56 (3H, s, 5-Me), 3.34 (1H, m, H-4'), 3.78 (214, s, H-9), 4.06 (1H, m, H-6'b), 4.15 (2H, 2 x m, H-3' and H-4 1), 4.52 (1H, d, J = 11 Hz, H- 6'a), 4.92 (1H, d, J= 10.2 Hz, H-1'), 5.32 (1H, d, J= 3.9 Hz, 3'-OH), 5.43 (1H, br s, 4'-OH), 5.49 (1H, t, J = 9.6 Hz, H-2'), 6.11 (1H, d, J = 16.0 Hz, H-2"), 6.15 (1H, s, H-3), 6.43 (1H, d, J= 16.0 Hz, 11-2 1"), 6.58 (111, s, H-6), 6.74 (211, d, J- 8.7 Hz, H-6" and H-8"), 6.76 (2H, d, J= 8.7 Hz, H-6"' and FI-8"), 7.25 (1H, d, J = 16.0 Hz, H-3"), 7.44 (2H, d, J= 8.6 Hz, I4-5" and 14-9"), 7.52 (2H, d, J= 8.6 Hz, H-5"' and H-9"), 7.54 (1H, d, J= 15.9 Hz, H-3"),

13 C 8: Table 4.4 ES-MS m/z: 685 ([M - 1] . , 58 %), 539 GM" - COCH=CHC6H4OH, 8 %), 521 ([ 1‘/E - COCH=CHC6H4OH - H20, 100 %),

5.5.2.3 CONSTITUENTS FROM A. BROOMII

Chopped leaves of A. broomii were extracted with Me0H (1000 ml), the extract filtered and the solvent removed in vacuo to provide a yellow brown residue. Analysis of the sample with the 200 analytical HPLC system revealed the presence of aloin A and B 4.5 as well as the unknown compounds 4.10 (Rt 30.27 min), 4.11 (Rt 21.56 min), 4.12 (RI 31.27 min) and 4.13 (R1 24.24 min). A portion (1 g) of the residue was subjected to preparative HPLC using a Phenomenex IB-

Sil C18 reverse phase column (5mm particle size, 250 x 10 mm internal diameter; flow rate 4.5 ml mind ; 1 ml sample loop). The solvent system was a 40% to 60% linear gradient of Me0H in H2O over 25 min, 100% in 2 min. A diode array detector with two channels (A set at 275 ± 70 nm; B set at 365 f 40 nm) was used. The separation was monitored by analytical HPLC.

(10R, 1'S)-6'-0-Caffeoy1-5-hydroxyaloin A (4.10)

Amorphous solid (8.6%). UV X= nm: 300 sh, 330. 1H 8: (5% DMSO-c/6 and CDC13) 2.85 (1H, t, J= 9.2 Hz, H-3'), 2.87 (1H, t, J= 9.2 Hz, H-4'), 3.02 (1H, m, H-5'), 3.24 (1H, t, J= 9.2 Hz, H-1 1), 3.26 (1H, t, J= 9.2 Hz, H-21), 3.78 (1H, m, H-6'b), 4.20 (1H, m, H-6'a), 4.52 (2H, d, Ja,b = -14.2 Hz, 11- Me), 4.68 (1H, d, J= 2.4 Hz, H-10), 6.03 (1H, d, J= 16.0 Hz, H-2"), 6.64 (1H, d, J= 8.6 Hz, H-7), 6.74 (2H, d, J= 1.2 Hz, H-2 and H-4), 6.78 (1H, d, J= 8.6 Hz,

H-8"), 6.88 (1H, dd, J8",9" = 8.6 Hz, .15, = 1.5 Hz, H-9"), 7.01 (1H, d, J= 1.5 Hz, H-5"), 7.03 (1H, d, J= 8.6 Hz, H-6), 7.28 (1H, d, J= 16.0 Hz, H-3"), 11.37 (IH, s, 8-0H), 11.73 (1H, s, 1-01-1)

13C 8: see Table 4.5 ES-MS m/z: 595 ([M - if, 100), 271 [NI" - COCH=CHC6H3(OH)(OH) - C6H,,O5, 60 %) FAB-MS m/z: 597 ([M + lf, 48 %), 433 (M + - COCH=CHC6H3(OH)(OH), 60 %), 314 (the - COCH=CHC6H3(OH)(OH) - C4H802, 16 %), 271 (433 (M + - COCH=CHC61-13(OH)(OH) - C6H1105, 100 %)

(E)-2-Acetony1-8-(2LO-caffeoy1)-p-D-glucopyranosyl-7-methoxy-5-methyl-chromone (4.11)

Amorphous solid (4.2%) UV Xmmf," nm: 302; 329 (sh) 201

1 1-1 8: 2.27 (3H, s, 11-Me), 2.66 (3H, s, 5-Me), 3.21 (1H, m, H-4' ), 3.25 (1H, m, H- 5'), 3.38 (1H, m, H-612'), 3.51 (1H, m, H-3'), 3.71 (1H, m, H-6'a), 3.79 (3H, s, 7- OMe), 3.82 (2H, s, H-9), 4.40 (1H, br s, 6 1-0H), 4.93 (1H, d, J= 10.0 Hz, H-I'), 5.17 (1H, br s, 4'-OH), 5.19 (1H, br s, 3'-OH), 5.42 (1H, t, J= 9.8 Hz, H-2'), 5.95 (IH, d, J=15.9 Hz, H-2"), 6.17 (1H, s, H-3), 6.71 (1H, d, J=8.1 Hz, H-8"), 6.84 (1H, s, H-6), 6.87 (1H, dd, J511,9" = 2.1 Hz, J811,9" = 8.1 Hz, H-9"), 6.93 (1H, d, J2.0 Hz, H-5"), 7.17 (1H, d, J= 15.9 Hz, H-3").

I3 C 8: see Table 4.6 ES-MS m/z: 569 - 1r, 100 %), 407 (M" - COCH=CHC6H3(OH)(OH), 35 %), 389 (M" - COCH=CHC6H3(OH)(OH) - H20, 8 %).

(E)-2-Acetony1-8-[(2'-0-cinnamoy1)-P-D-glucopyranosyl-7-methoxy-5-methylchromone (4.12)

Amorphous solid (3.2%) UV Xmmea?" nm: 251, 284.

1 H 8: (5% DMSO-d6 and CDCI3) 2.29 (3H, s, 11-Me), 2.69 (3H, s, 5-Me), 3.41 (1H, m, H-5' ), 3.63 (1H, m, H-4'), 3.66 - 3.82 (3H, m, H-3', H-6'a and H-6b'), 3.81 H-3), 6.12 (1H, d, J= 15.5 Hz, H-2"), 6.55 (1H, s, H-6), 7.29 (3H, m, H-6", H-7" and H-8"), 7.36 (IH, d, J= 15.4 Hz, H-3"), 7.37 (2H, m, H-5" and H-9").

13C 8: (5% DMSO-d6 and CDC13) see Table 4.6 ES-MS m/z: 537 ([M - lf, 100 %), 407 (M" - COCH=CHC6H5, 5 %), 389 (M" - COCH=CHC6H5 - H20, 65 %).

(10R, 1'S)-5-Hydroxyaloin A (4.13)

Amorphous solid (16%). UV eme$1c 1 nm:268, 298, 365

202

8: (5% DMSO-d6 and CDCI3 2.84 (IH, m, H-5'), 2.85 (I H, t, J = 9.2Hz, H-2'), 2.96 (1H, t, J= 9.2 Hz, H-3'), 3.28 (1H, t, J= 9.2 Hz, H-4'), 3.33 (1H, m, H-6'b), 3.35

(1H, t, J= 9.2 Hz, H-1'), 3.44 (1H, m, H-6'a), 4.56 (1H, d, Ja,b = -14.6 Hz, H-11), 4.69 (2H, d, J= 2.4 Hz, H-10), 6.68 (1H, d, J= 8.6 Hz, H-7), 6.78 (1H, d, J = 1.5 Hz, H-2), 6.94 (1H, d, J= 1.5 Hz, H-4), 7.04 (1H, d, J= 8.6 Hz, H-6), 11.35 (1H, s, 8-OH), 11.68 (1H, s, 1-OH).

DC 8: (5% DMSO-d6 and CDCI3) see Table 4.5 ES-MS m/z: 433 ([M - 1] -, 55 %), 313 (M" - C4H804, 8 %), 271 (M - - C6H1105, 100 %). FAB-MS m/z: 435 ([M + 1] +, 65), 272 (M+ - C611,105, 100 %), 255 (M+ - C6F11,05- OH, 60 %)

5.6 ISOLATION OF NEW COMPOUNDS FROM WILD GINGER, SIPHONOCHILUS AETHIOPICUS

Rhizomes and roots were collected from a commercial nursery. Since S. aethiopicus is the only indigenous member of the family in South Africa and since it has such a unique and distinctive morphology no voucher specimen was collected.

The fleshy crushed roots of several plants of S. aethiopicus were steam distilled for one hour to obtain a pale yellow semi-crystalline mixture which contained (TLC) compound 4.15 as main constituent. Extraction of the roots with cold ethanol for 5 days and evaporation of the solvent furnished a residue which again contained mainly 4.15. Flash chromatography of a portion (0.5 g) of the residue on silica gel (100 g) in toluene and 2 % triethylamine afforded two main compounds. 203

(4aa,5[3,8aP)]-3,5,8a-Trimethy1-4,4a,8a,9-tetrahydronaphtho12,3N-furan-8(5H)-one (4.15)

White crystals (7.8%) Mp: 60 °C

[a]O +107.4° (CDC13, c, 1)

1 11 5: see Table 4.7

13 C 5: see Table 4.8 EI-M m/z: 230 (Mt, 100 %), 215 (Mt - CH3, 90 %), 197 ((M t - CH3 - , 13), 187 (56)

(4acr.,513,843)]-2-Hydroxy-3,5,8a-Trimethyl-4,4a,8a,9-tetrahydronaphtho[2,3N-furan- 8(5H)-one (4.16)

White crystals (14 %) Mp: 109 °C +84° (MeOH, c, 1.0)

1 H 5: see Table 4.7

13 C 8: see Table 4.8 FAB-MS m/z: 247 (Mt + 1, 100 %)

(4aa.,5[3,8a13)]-2-Acetoxy-3,5,8a-trimethyl-4,4a,8a,9-tetrahydronaphtho[2,36]-furan-8(5H)- one (4.17)

A solution of compound 4.16 (100 mg), acetic anhydride (0.5 ml) and triethylamine (0.5 ml) in CH2Cl2 (3 ml) was stirred for 70 hours at room temperature. The solvent was removed under reduced pressure and the crude product chromatographed in EtOAC: CHC13 (1:1) to furnish the compound 4.17 as a colourless oil in a quantitative yield.

H 5: see Table 4.7 13 C 5: see Table 4.8 204

5.7 REFERENCES

D.D. Perrin and W.L.F. Armarego, Purification of Laboratory Chemicals, 3rd Ed. Pergamon: Oxford, 1988. B.M. Fumiss, A.J. Hannaford, P.W.G. Smith, A.R. Tatchell, Vogel's Textbook of Practical Organic Chemistry, 5th Ed, Longman Scientific and Technical: New York, 1989. R. F. Heck, Palladium Reagents in Organic Synthesis, Academic: New York, 1985. W.C. Still, M. Kahn and A. Mitra, J. Org. Chem., 1978, 43, 2923. J. Rigaudy and S.P. Klesney, Nomenclature of Organic Chemistry, Sections A, B, C, D, E, F and H, Pergamon: Oxford, 1979. H. Hartenstein and D. Sicker, J. Prakt. Chem., 1993, 335, 103. Beilstein 21, 77. J.P. Phillips, E.M. Barrall and R.Breese, Kentucky Acad. Sci., 1956, 17, 135. P. Sutter and C.D. Weis, J. Heterocyclic Chem., 1986, 23, 29. C. Barbier, A. Joissans, A. Commercon, J.-F. Riou and F. Huet, Heterocycles, 2000, 53, 37. K. Ogawa, Y. Nomura, Y. Takeuchi and S. Tomoda, ./ Chem. Soc. Perkin Trans I, 1982, 3031. J.K. Stille and Y. Becker, J. Org. Chem., 1980, 45, 2139. G.A. Kraus and K. Neuenschwander, J. Org. Chem., 1981, 46, 4791. J. Szmuszkovicz, W.C. Anthony and R.V. Heinnzelman, J. Org. Chem., 1960, 25, 857. E. Taschner, C. Wasielewski and J.F. Biemat, Liebigs Ann. Chem., 1961, 646, 119. H.W. Rauwald and A. Beil, Z. Naturforsch., 1993, 48c, 1 . ACKNOWLEDGEMENTS

I would like to express my sincere thanks to the following people:

Professor Cedric Holzapfel - teacher par excellence who generously shared his endless

knowledge with great enthusiasm.

Dr. Louis Fourie of Potchefstroom University for recording accurate masses.

My parents for their steadfast belief in me over the years and especially for taking care of

our children during school holidays.

Daniel for continuous encouragement, endless patience and help with the final

preparation of the manuscript.

Finally, Daniel Jr. and Gerhard for bearing with me in times of trouble.