Total Synthesis of 1,2- Diamine Contained Alkaloids, Schizozygine, Vallesamidine and Strempeliopine (Chapter 1 and 2)

Towards Schizozygine Type Alkaloids: Total Synthesis

of (+)-Vallesamidine and (+)-Strempeliopine

Xiangyu Zhang

A thesis submitted in partial fulfilment of the requirement for the degree of

Doctor of Philosophy

Department of Chemistry

University College London

Declaration

I, Xiangyu Zhang, confirm that the work presented in this thesis is my own. Where information has been derived from other sources, I confirm that this has been indicated in the thesis.

Signed:

Date:

Abstract

As a classic and powerful tool for carbon-carbon bond formation, the nitro-Mannich reaction has shown its versatility in drugs and natural products syntheses. The 1,2- diamine structure, a reduced moiety from nitro-Mannich adduct, is widely present in naturally occurring alkaloids and this feature suggested the potential application of nitro- mannich reaction in such alkaloids synthesis. This thesis showcases the nitro-Mannich reaction as a key strategic reaction through studies towards the total synthesis of 1,2- diamine contained alkaloids, schizozygine, vallesamidine and strempeliopine (Chapter 1 and 2).

Initial studies on the schizozygine molecule (Chapter 3) generated a diastereoselective nitro-Mannich reaction on -branched nitroalkanes to synthesise complex -nitroamines with three contigurous chiral centres and syn,anti stereochemistry.

This reaction was followed by a reductive cyclisation to achieve the functionalised piperidine ring C. Although the subsequent manipulation towards advanced shcizozygine intermediate was unsuccessful, the nitro-Mannich/reductive cyclisation sequence provided methodology for highly functionalised piperidine ring synthesis.

A second generation route using nitro-Mannich reaction was accompanied by other nitro group chemistry, Michael addition, Tsuji-Trost allylation and nitro group reduction/C-N coupling reaction, to realise the quick and concise preparation of an A/B/C ring intermediate. An unusual and novel [1,4]-hydride tansfer/Mannich type cyclisation was carried out to build the ring E. The resulting A/B/C/E ring intermediate was used divergently to complete the total synthesis of (+)-vallesamidine (Chapter 4) and (+)-

14,15-dehydrostrempeliopine (Chapter 5) as well as three other unnatural analogues.

These natural and unnatural products could be candidates for drug discovery research and the route would be applicale for the synthesis of schizozygine and related molecules. ii

Statement of impact

Natural products are chemical compounds or substances produced by living organisms and many of them possess impressive biological activities. Of the approved pharmaceuticals joining the market between 1981 and 2010, 64% were related to natural products. The chemical synthesis of natural products can save the use of precious and sometimes scarce natural resources, provide more material than nature can produce and can facilitate drug discovery. This thesis developed a new route to a late stage synthetic divergent intermediate from which vallesamidine, strempeliopine and schizozygine type alkaloids could be prepared. To exemplify this the total syntheses of alkaloids (+)- vallesamidine and (+)-14,15-dehydrostrempeliopine were completed. This route highlighted nitro group chemistry and intramolecular C-H functionalization and showcased their diversity in multi-target synthesis. The route contributes a notable strategy to the chemical community that will potentially be used for other alkaloid syntheses. This enabling synthetic route will enable the evaluation of rare structures for biological studies and have impact and inspire drug discovery to tackle disease.

iii

Acknowledgment

I would like to thank my supervisor, Prof. Jim Anderson, for his guidance in my research during the last four years. The PhD research journey was full of challenges and no achievement could be made without his mentorship and inspiration.

I also want to thank the support from NMR (Dr. Abil Aliev), mass spectroscopy (Dr.

Kersti Karu) services and Dr Merina Corpinot for the single crystal X-ray structure determination.

My PhD study was funded by UCL-Chinese Scholarship Council (CSC) joint scholarship and I appreciate this funding scheme for providing me a chance to carry out my PhD research at UCL.

Finally, I would like to thank the strong backing and support from my Family, especially my mom and my girlfriend Minshan.

List of abbreviations

Ac = Acetyl Aq. = Aqueous

Ac2O = Acetic anhydride All = Allyl Alloc = Allyloxycarbonyl Ar = Aryl 9-BBN = 9-Borabicyclo[3.3.1]nonane BMS = Borane-dimethylsulfide complex Bn = Benzyl Boc = tert-Butoxycarbonyl n-Bu = n-Butyl s-Bu = sec-Butyl t-Bu = tert-Butyl Bz = Benzoyl brsm = % yield based on recovered starting material CAN = Ceric ammonium nitrate Cbz = Benzyloxycarbonyl Cod = Cyclooctadiene Cp = Cyclopentadienyl 18-crown-6 = 1,4,7,10,13,16-Hexaoxacyclooctadecane CSA = Camphorsulphonic acid DAST = Diethylaminosulphur trifluoride dba = Dibenzylideneacetone DBAD = Di-tert-butyl azodicarboxylate DBN = 1,5-Diazabyciclo[4.3.0]non-5-ene DBU = 1,8-Diazabyciclo[5.4.0]undec-7-ene DCC = 1,3-Dicyclohexylcarbodiimide DCE = 1,2-Dichloroethane DCM = Dichloromethane

DDQ = 2,3-Dichloro-5,6-dicyano-1,4-benzoquinone DEAD = Diethyl azodicarboxylate DIBAL = Diisobutylaluminium hydride DIBAL-H = Diisobutylaluminium hydride DIPEA = Diisopropylethylamine DMAP = 4-Dimethylaminopyridine DME = 1,2-Dimethoxyethane DMF = N,N-Dimethylformamide DMP = Dess-Martin periodinane DMPU = 1,3-Dimethyl-3,4,5,6-tetrahydro-2(1H)- pirimidone DMS = Dimethylslfide DMSO = Dimethylsulfoxide DPA = Diisopropylamine DPPA = Diphenylphosphoryl azide dr = Diastereomeric ratio EDCI = 1-Ethyl-3-(3- dimethylaminopropy)carbodiimide hydrochloride EDTA = Ethylenediaminetetraacetic acid EE = Ethoxyethyl EOM = Ethoxymethyl e.e. = Enantiomeric excess e.r. = Enantiomeric ratio EDG = Electron donating group EWG = Electron withdrawing group HMDS = Hexamethyldisilazane HMPA = Hexamethylphosphoramide HOBt = 1-Hydroxybenzotriazole Hoveyda-Grubbs 2nd Gen catalyst (1,3-Bis-(2,4,6-trimethylphenyl)-2- imidazolidinylidene)dichloro(o- isopropoxyphenylmethylene)ruthenium vi

IPA = Isopropyl alcohol Im = Imidazole IBX = 2-Iodoxybenoic acid KHMDS = Potassium bis(trimethylsilyl)amide LAH = Lithium aluminium hydride LDA = Lithium diisopropylamide LHMDS = Lithium bis(trimethylsilyl)amide mCPBA = meta-Chloroperoxybenzoic acid M.S. = Molecular sieves Ms = Methanesulphonyl NaHMDS = Sodium bis(trimethylsilyl)amide NBS = N-Bromosuccinimide NIS = N-Iodosuccinimide NMM = N-Methylmorpholine NMO = N-Methylmorpholine-N-oxide Ns = p-Nitrophenylsulphonyl Para/ortho/meta (p-, o-, m-) = (1,4)/(1,2)/(1,3)-Substitution on benzene PDC = Pyridinium dichlorochromate PCC = Pyridinium chlorochromate Pg = Protective group Ph = Phenyl PMB = p-Methoxybenzyl PMP = p-Methoxyphenyl PPTS = Pyridinium p-toluensulphonate n-Pr = n-Propyl Pr = Propyl iPr = iso-propyl PTSA = p-Toluenesulphonic acid Py = Pyridine quant. = Quantitative rac = Racemic

vii

rt = Room temperature TBAF = Tetra-n-butylammonium fluoride TBDMS/TBS = tert-Butyldimethylsilyl TBDPS = tert-Butyldiphenylsilyl TEA = Triethylamine TEBAC = Triethylbenzylammonium chloride TEMPO = (2,2,6,6-Tetramethylpiperidin-1-yl)oxyl radical TES = Triethylsilyl Tf = Trifluoromethanesulfonyl TfO = Trifluoromethanesulfonate

Tf2O = Trifluoromethanesulfonyl anhydride TfOH = Trifluoromethanesulfonic acid TFA = Trifluoroacetic acid TFAA = Trifluoroacetic anhydride Thexyl = 2,3-Dimethyl-2-butyl THF = Tetrahydrofurane TIPS = Triisopropylsilyl TMEDA = N,N,N',N'-Tetramethylethylendiamine TMG = Tetramethylguanidine TMS = Trimethylsilyl p-Tol = p-Toluyl or 4-methylbenzyl o-Tol = o-Toluyl or 2-methylbenzyl TPAP = Tetra-n-propylammonium perruthenate Tr = Trityl, triphenylmethyl Troc = 2,2,2-Trichloroethoxycarbonyl p-Ts = p-Toluenesulphonyl p-TsOH = p-Toluenesulphonic acid

viii

Table of Contents

Declaration ...... i

Abstract ...... ii

Acknowledgment ...... iii

List of abbreviations ...... v

Introduction ...... 1

Chapter 1. Applications of the nitro-Mannich reaction in organic synthesis ...... 2

1.1 Synthesis of molecules containing the 1,2-diamine moiety ...... 4

1.2 Nitro-Mannich/reductive denitration sequences in drugs and natural product

synthesis ...... 12

1.3 Nitro-Mannich/Nef reaction sequences in organic synthesis ...... 15

1.4 Research aim ...... 18

Chapter 2. Targets background and synthetic studies ...... 20

2.1 Monoterpene indole alkaloids ...... 20

2.2 Schizozygine, vallesamidine and strempeliopine: isolation and potential

activities ...... 22

2.3 Schizozygine, vallesamidine and strempeliopine: absolute configuration (AC)

determination ...... 25

2.4 Schizozygine, vallesamidine and strempeliopine: proposed biosynthesis ...... 27

2.5 Schizozygine, vallesamidine and strempeliopine: synthetic studies ...... 30

Results and discussion ...... 51

Chapter 3. Studies on the total synthesis of schizozygine ...... 52

3.1 Retrosynthesis ...... 52

3.2 Feasibility and literature precedents ...... 52

3.3 Synthetic studies ...... 58

3.4 Chapter conclusion ...... 85

Chapter 4. New attempt towards schizozygine: total synthesis of (+)-vallesamidine

...... 87

4.1 Revised retrosynthesis ...... 87

4.2 Feasibility and literature precedents on key transformations ...... 89

4.3 Synthetic studies ...... 100

4.4 Chapter conclusion ...... 130

Chapter 5 Studies towards the total synthesis of (+)-strempeliopine ...... 132

5.1 Retrosynthesis of lactam ring F ...... 132

5.2 Feasibility and literature precedents ...... 133

5.3 Synthetic studies ...... 135

5.4 Chapter conclusion ...... 160

Conclusions and future work ...... 161

References ...... 163

Introduction

➢ Nitro-Mannich reaction in organic synthesis

➢ Monoterpene indole alkaloids

➢ Vallesamidine, strempeliopine and schizozygine:

current studies and limitations

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

Our group has been focused on the development of the nitro-Mannich reaction for many years and has a long-term goal to use this reaction in complex molecule synthesis. As one of the applications of the nitro-Mannich reaction is to prepare 1,2-diamine compounds, we proposed that this strategy could be involved in the synthesis of alkaloids containing this sub-structure. This PhD thesis will describe research in the total synthesis of vallesamidine and schizozygine alkaloids using the nitro-Mannich reaction.

This chapter will provide a short introduction of the nitro-Mannich reaction and highlight the applications of this reaction in the literature, based on further transformations of the nitro-Mannich adducts. Then the background of vallesamidine and schizozygine alkaloids, including isolation, biosynthesis, reported chemical total synthesis and synthetic limitations will be introduced.

Chapter 1. Applications of the nitro-Mannich reaction in organic synthesis

Carbon-carbon bond formation using readily prepared C-nucleophiles by deprotonation adjacent to a C=X group is one of the most important methods in organic chemistry.1

This transformation mainly involves four types of reactions: Aldol, nitro-Aldol (Henry),

Mannich and nitro-Mannich (aza-Henry) reactions (Scheme 1). However, far less attention was put on the nitro-Mannich reaction until recent decades. Compared with the first three reactions, the combination of a less nucleophilic nitronate and less electrophilic imine (nitro-Mannich) is more difficult in the absence of any activation (e.g. Bronsted acid, Lewis acid (transition metal catalysis) or H-bond (organocatalysis)). Along with 2

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

the development of transition metal catalysis and organocatalysis, the nitro-Mannich reactions were rapidly investigated and many methods were developed for diastereo- and enantioselective reactions.2

The product from a nitro-Mannich reaction, -nitroamine 1, could be further transformed to 1,2-diamino (nitro reduction, 1 to 4), mono amino (denitration, 1 to 2) and

-amino carbonyl compounds (Nef reaction, 1 to 3). These diverse intermediates have showed good applications in target molecule synthesis.

Scheme 1. Nitro-Mannich reaction

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

Scheme 2. Reactions of -nitro amines.

1.1 Synthesis of molecules containing the 1,2-diamine moiety

The 1,2-diamine moiety is widely represented in natural products and drugs (Scheme 3), in cyclic and acyclic skeletons. As a useful method to construct the 1,2-diamine structure, the nitro-Mannich reaction has been reported in several natural product and drug synthesis.

Scheme 3 Natural products and drugs containing 1,2-diamine moiety

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

Oseltamivir 5, commercially available as its phosphate salt (Tamiflu), is an antiviral drug treating influenza A and B. In 2010, Lu’s group reported the synthesis of

Oseltamivir from diethyl D-tartrate 12 (Scheme 4).3 The chiral imine 13 was synthesized from diethyl D-tartrate 12 and then a nitro-Mannich reaction was conducted to yield the nitroamine 14 in 86% yield with a dr of 10:1. Manipulations of the amino protecting group from sulfoxide to acetate gave 15 and then IBX oxidation of the alcohol afforded the aldehyde 16 without any epimerisation. A Michael addition followed by an intramolecular Horner–Wadsworth–Emmons reaction with DBU/LiCl was carried out in one-pot to give the functionalized cyclohexene 17 as a 3:2 mixture of diastereomers. To improve the diastereoselectivity, Hayashi’s strategy4 was used here by treating 17 with tolSH in the presence of Cs2CO3 to give compound 18 as a single diastereomer. The mechanism involves Michael addition of tolSH followed by epimerisation of the nitro group into the thermodynamically more stable positions. After that, reduction of the nitro group and removal of thiol group by elimination gave the target molecule.

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

Scheme 4. Synthesis of (-)-Oseltamivir.

In 2002, Shibasaki’s group reported the synthesis of CP-99994 (8) and ICI-

199441 (9) using an asymmetric nitro-Mannich reaction.5 Molecule CP-99994 is a competent antagonist of substance P, which can be used in different inflammatory diseases’ treatment. Molecule ICI-199441, a highly potent -opioid agonist, can provide useful analgesics with less side effects. The synthesis of CP-99994 was commenced with an asymmetric nitro-Mannich reaction of imine 20 and nitro compound 21 catalyzed by

(S)-ALB (20 mol%) 22 and KO-t-Bu (18 mol%) at -40 °C giving 23 in 40 % yield and

97 % ee after purification (Scheme 5). Removal of the TBS group followed by Dess-

Martin oxidation and cyclisation gave the key intermediate dihydropiperidine 24.

Epimerisation through a nitrosilylation/deprotection sequence was applied on 24 to give

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

cis isomer 25. The nitro group was reduced, and the resulting free amine was alkylated by reductive amination with o-anisaldehyde. Upon treatment with LiAlH4, the diphenyl phosphinoyl group was removed at last to yield CP-99994 (8) in 30% yield.

Scheme 5. Synthesis of CP-99994.

Similarly, the synthesis of ICI-199441 (9) also started with an asymmetric nitro-

Mannich reaction (20 to 28) in the presence of YbKH2[(R)-binapthoxide]3 27 (Scheme 6).

A simple reduction of the nitro group was performed, and the resulting amine 29 was subjected to the double reductive amination to the pyrrolindine compound 30. Hydrolysis of the diphenylphosphinoyl group under acidic conditions followed by methylation of the free amino group provided 31, which was simply treated with acyl chloride 32 to achieve the synthesis of ICI 99441 (9) as its hydrochloride salt.

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

Scheme 6. Synthesis of ICI 199441

Scheme 7. Asymmetric nitro-mannich reaction and synthesis of nemonapride. 8

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

The same group in 2010 described a heterobimetallic Cu/Sm/dinucleating Schiff base complexes that catalysed the syn selective nitro-Mannich reaction (Scheme 7).6

Under the optimal condition, the authors synthesised -nitro amine 38 from 36 with an excellent syn selectivity but a moderate enantioselecyivity (80% ee). The nitro group reduction followed by removal of TBDPS group delivered the alcohol 40 that was subsequently recrystalised to give the enantiopure material (>99% ee). An intramolecular displacement gave pyrrolidine 41 and condensation with mixed anhydride 42 afforded nemonapride 10, which is used in clinic as an antipsychotic agent.

During the studies on DPP4 inhibitors for diabetes treatment, a research team at

Abbott proposed that lactam structures such as 46 would be good targets for their investigation. Therefore, a series of piperidinone-constrained phenylethylamines were designed and synthesised. Both racemic and enantioselective synthesis were developed and shown in Scheme 8.7 As the target structure contains the 1,2-damine moiety, the key step involved a multicomponent, nitro-Mannich/lactamisation cascade reaction. The enantioselective synthesis used an asymmetric Michael addition of diethylmalonate to the nitroalkene 47. The nitromalonate 48 was treated with paraformaldehyde and various alkylamines in refluxing ethanol to afford the lactams 49. The Krapcho decarboxylation was then carried out to give decarboxylated lactam 50 as the major anti-diastereoisomer.

Reduction of the nitro group to free amine gave the final targets 51. Experiments indicated excellent selectivity (1000 fold on DPP4 than other related peptidases) and potent inhibition.

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

Scheme 8. Synthesis of piperidinone-constrained phenylethylamines as DPP4 inhibitors

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

Scheme 9. Synthesis of proposed structure of Piperazirum

Piperazirum was isolated from Arum palaestinum Boiss and the structure was originally proposed as 58.8 Anderson’s group reported a concise synthesis of 58 using a diastereoselective nitro-Mannich reaction as a key transformation (Scheme 9).9 An 1,4- reduction/nitro-Mannich cascade reaction of 52 and imine 53 gave the nitroamine 54 with full conversion and an excellent dr over 95:5. The nitro group was then reduced to a free amine and coupled with the carboxylic acid 56, followed by an intramolecular imine formation/tautomerisation to provide the piperazinone 57 in 69% yield. Hydrogenation from the least hindered face and PMP group removal finally gave the desired product 58.

The structure was confirmed by single crystal X-ray diffraction experiment, but the NMR data did not match the reported spectra from isolation, which indicated that the original proposed structure was not correct and required revision.

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

1.2 Nitro-Mannich/reductive denitration sequences in drugs and natural product synthesis

Without taking the 1,2-diamine motif into account, the nitro-Mannich reactions can be regarded as a method to form a carbon-carbon bond. In this case, the nitro group is an

‘auxiliary group’ or ‘traceless agent’ to direct the stereoselective formation of carbon- carbon bond. This group can then be easily removed by reductive denitration. Ultimately, with reapect to bond disconnection in retrosynthetic analysis (Scheme 10), a possible disconnection for any amine could start from the functional group addition of a nitro group to the  position (60), which would then allow a nitro-Mannich disconnection of the C-C bond to an imine 61a or imine precursor 61b and a nitroalkane 62.

Scheme 10. Nitro-Mannich/denitration strategy

In 2008, Dixon’s group reported a formal synthesis of (3S,4R)-paroxetine 67, an antidepressant drug, using a nitro-mannich/lactamisation strategy (Scheme 11).10 The enantioenriched nitro compound 64 was synthesised from nitroalkene 63 via an organocatalysed Michael addition of dimethylmalonate. A one-pot nitro-

Mannich/lactamization was conducted and was followed by a one-pot radical denitration and decarboxylation under thermal AIBN/n-Bu3SnH conditions to give lactam 66, a known intermediate that could be converted to paroxetine according to the literature.

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

Scheme 11. Formal synthesis of Paroxetine

Scheme 12 Total Synthesis of (-)-Nakadomarin A

A more complex example utilising this synthetic strategy from the same group was demonstrated in the total synthesis of (-)-Nakadomarin A, a complex 8/5/5/5/15/6 hexacyclic marine alkaloid showing various bioactivities (Scheme 12).11 Thiourea 68 catalysed conjugate addition of 69 gave desired product 70 in 57% yield and 91:9 dr after 13

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

a reaction time of 8 days. The nitro-Mannich/lactamisation was applied again to give 72 in 68% yield. Reductive denitration of 72 afforded key intermediate 73 from which the whole synthesis was completed in only three more steps.

The Johnston’s group described a nitro-Mannich/denitration protocol to enantioselectively synthesise -amino acids (Scheme 13).12 Using the catalyst 77, imine

75 reacted with nitro ester 76 to give a diastereoisomeric mixture of 78. Fortunately each diastereoisomers showed the same excellent level of enantioselectivity from the selective imine addition. Therefore, regardless of the diastereo-selectivity of the nitro-Mannich reaction, the overall protocol after radical denitration gave identical amino ester products without loss of optical purity. This two-step sequence was applied to the preparation of

80 with 75% yield and 88% ee. Upon treatment with thionyl chloride in methanol, global hydrolysis/deprotection occurred followed by in situ methyl ester formation of the carboxylic acid. Protection on the free amine with Boc2O afforded enantioenriched 81.

Only the racemic form had previously reported by Wasserman in the total synthesis of

(+/-)-chaenorhine 82.

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

Scheme 13. Formal enantioselective synthesis of -amino acids using the nitro-

Mannich/denitration protocol.

1.3 Nitro-Mannich/Nef reaction sequences in organic synthesis

Another type of manipulation of the nitro group is the Nef reaction, which is the conversion of a nitro group to a carbonyl group.13 This reaction greatly expands the scope of the nitro group applications as the carbonyl compounds have much more opportunities for further transformations. Therefore, using nitroalkanes to enable carbon-carbon bond formation, followed by Nef reaction can be a powerful strategy to generate various carbonyl compounds and applications utilising this methodology have been reported.14

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

Scheme 14. Nef reaction

Readily available enantio- and diastereo-enriched -nitro amine, products from the nitro-Mannich reaction, can be converted into enantiomerically pure -amino carbonyl molecules (83 to 84) by the Nef reaction. In contrast controlling the stereoselectivity of the -amination of the carbonyl compounds (85 to 84) is usually challenging.

Scheme 15. Preparation of a-amino carbonyl compounds

Palomo and co-workers reported an asymmetric nitro-Mannich reaction of Boc

15 protected imine 86 using Zn(OTf)2/(-)-NME catalysis system in 2005 (Scheme 16). The authors took one of the products 88 and performed the Nef reaction to deliver the -amino acid 89. After esterification, HPLC analysis on compound 90 indicated the same enantiomeric purity as 88 (97% ee). No other analogues were tested but this protocol could be a useful method to synthesise enantioenriched amino acids.

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

Scheme 16. Nitro-Mannich/Nef reaction sequence to give phenyl -amino acids

In 2014, Foubelo, Yus and co-workers reported a diastereoselective nitro-

Mannich reaction on chiral N-tert-butylsulfinyl imines 91, promoted by NaHCO3

(Scheme 17).16 This methodology worked well for aryl, alkyl imines with nitromethane and nitroethane in moderate to good yields and diastereoselectivities. The nitromethane adducts 92 were then treated with NaNO2, AcOH in warm DMF to give the corresponding

-amino acids 94 in moderate yields and above 95% purity (by GC). The epimeric mixture of nitroethane adducts 93 were subjected to potassium hydroxide in methanol, followed by oxidation with potassium permanganate at 0 oC to give the single -amino ketones 95 without epimerisation at the chiral centre adjacent to the carbonyl group.

A more impressive example was reported by Dixon’s group in their total synthesis of Manzamine A (101).17 The authors developed an intramolecular reductive nitro-

Mannich reaction between the nitroalkane and the amide group in 96 to give the cyclic nitro amine 97 with a dr of 83:17 (Scheme 18). Then a TiCl3 promoted Nef ketone 98 and oxime 99 in 56% and 21% yields respectively. The ketone intermediate 98 was then

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

treated with 3-butenylmagnesium bromide to provide the desired tertiary alcohol moiety

100 needed for the target manzamine A 101.

Scheme 17. Nitro-Mannich/Nef reaction to alkyl/alryl -amino acids

Scheme 18. Dixon’s synthesis of Manzamine A

1.4 Research aim

To expand the scope of the nitro-Mannich reaction in complex natural product synthesis, we decided to focus on monoterpene indole alkaloids that contained 1,2-diamine in their

Chapter 1. Applications of nitro-Mannich reaction in organic synthesis

core structures. We began to develop a general synthetic route to the target structures vallesamidine (102), strempeliopine (103) and schizozygine (6) molecules (Scheme 19) that involved the nitro-Mannich reaction as a key step. The route could provide a new strategy and disconnection for other related molecules.

Scheme 19. Synthetic targets

Chapter 2. Targets background and synthetic studies

2.1 Monoterpene indole alkaloids

The indole alkaloids are one of the largest family of alkaloids with more than 4100 structurally diverse compounds discovered. 18 Their fascinating structures, stereochemistry, bio- and chemical synthesis and impressive bioactivities (e.g. anticancer, anti-malarial and anti-arrhythmic etc.) have attracted the attention of organic and medicinal chemists. According to their biogenesis, indole alkaloids can be classified into: i) simple indole derivatives; ii) -carboline derivatives; iii) semi-terpenoid indoles and iv) monoterpenoid indoles.19 The monoterpene indole alkaloids are the most important type and are characterised structurally by an indole moiety (from tryptophan 104) and a C9 or

C10 fragment derived from secologanin 106 (Scheme 20). All terpene indole alkaloids are derived from the coupling of tryptamine 105 and secologanin 106 via a Pictect-

Spenglar reaction, de-glycosylation to remove the glucose part and self-condensation to give the 4,21-dehydrogeissoschizine intermediate 110. Rearrangements of 110 finally achieves the corynanthe, iboga and aspidosperma alkaloid classes with unique carbon skeletons (Scheme 20a). While many other skeletons have been elucidated, they can always be identified to one of the above three types through further rearrangements.20

Chapter 2. Targets background and synthetic studies

Scheme 20. Monoterpene indole alkaloids

Chapter 2. Targets background and synthetic studies

2.2 Schizozygine, vallesamidine and strempeliopine: isolation and potential activities

Isolation: The alkaloid schizozygine (6) was one of a small class of alkaloids, which was first isolated by Renner and co-workers in 1963 from Schizozygia caffaeoides (Boj.)

Baill, a monotypic genus of plant belonging to the Apocynaceae family in East Africa.21

This plant was used as a traditional medicine to treat some skin diseases in Kenya. 22

Other analogues, including schizogaline (117), schizogamine (118). New analogues like

3-oxo-14,15-epoxyshcizozyginen (119) and 6,7-dehydro-19-hydroxyschizozygine

(120) were also isolated and structurally identified in 200223 and 200424 respectively

(Scheme 21).

Scheme 21. Schizozygine and vallesamidine alkaloids

Chapter 2. Targets background and synthetic studies

Vallesamidine (102) was isolated from Vallesia dichotoma Ruiz et Pav in 1965, along with 27 other alkaloids.25 Compared with schizozygia, which only yielded two types of skeletons, vallesia contains a much wider structural diversity. Strempeliopine

(103) was isolated from Strempeliopsis strempelioides K. Schum in 1980 26 and andrangine (121) was isolated from Crioceras dipladeniiflorus in 1974.27 Additionally, all the above alkaloids belong to the Apocynaceae family or aspidosperma class, and this could suggest similar biosynthetic pathways to these alkaloids.

Bioactivities: The bioactivity of vallesamidine (102) has not been determined due to the insufficient quantity of materials obtained from the early stage of isolation. However, some more common alkaloids were also isolated from Vallesia dichotoma Ruiz et Pav, including reserpine (treatment of high blood pressure) and aspidospermine (diuretic, vasoconstriction, hypertensive, and respiratory stimulant properties28), which suggests that vallesamidine may also possess similar bio-activities.

For the schizozygine alkaloids (6, 117 to 120) and strempeliopine (103), preliminary biological studies have shown both anti-fungal and anti-bacterial activities.

Schizozygine (6) and Strempeliopine (103) possess moderate minimum inhibitory concentrations (> 500 g/mL) to all the tested organisms (Table 1) while 120 was much more active towards the tested fungi, performing better than the commercial drug ketoconazole.29

Chapter 2. Targets background and synthetic studies

Table 1. Bioactivities of selected schizozygine type alkalids

tested organisms MIC (g/mL)

6 103 120 ketoconazole ampicillin

Fungi

Trichophyton mentagrophytes > 500 > 500 < 1.95 6.25 N.T.

Microsporum gypseum > 500 > 500 1.95 6.25 N.T.

Epidermopgyton floccossum > 500 > 500 > 1.95 25 N.T.

Trichophyton tonsurans > 500 > 500 3.9 50 N.T.

Trichophyton interdigitale > 500 > 500 < 3.9 25 N.T.

Cladosporium cladosporioides > 500 > 500 7.8 25 N.T.

Cladosporium harbarum > 500 > 500 15.6 50 N.T.

Candida albicans > 500 > 500 7.8 25 N.T.

Bacteria

Escherichia coli > 500 > 500 > 250 N.T. 20

Staphyloccoccus aureas > 500 125 > 500 N.T. 12.5

Bacillus subtilis > 500 62.5 > 500 N.T. 12.5

Pseudomonas aeruginosa > 500 > 500 > 500 N.T. 25

Other biological tests30of these indoline alkaloids revealed moderate to good anti- plasmodial activities. Specifically, in terms of the inhibition of D6 and W2 strains of plasmodium falciparum, Schizozygine (6) showed an IC50 of around 33.6 (D6) and 26.8

M (W2) and hydroxyl-schizozygine (120) and epoxischizozygine (119) gave 19.1, 38.3

M for D6 and 29.1, 59.1 M for W2, repectively.

Chapter 2. Targets background and synthetic studies

2.3 Schizozygine, vallesamidine and strempeliopine: absolute configuration (AC) determination

The absolute configuration of (-)-vallesamidine (102) was determined by an X-ray diffraction experiment31 of its N-methyliodide salt 122 (Scheme 22). The structure of (-

)-andrangine (121) was elucidated through chemical correlation and its absolute configuration was determined by its conversion to 124 via an epoxide ring opening and a rearranged elimination gave product 124. The optical data (specific rotation) of the reduced product was in agreement with compound (+)-102, which suggested an absolute configuration opposite to that of (-)-vallesamidine.32

In 1982, Hájíček et al reported a synthesis of (-)-strempeliopine (103) from the aspidoaperma type compound (+)-18-methylenevincadifformine (125) (synthetic details will be discussed in section 2.5). 33 In this synthesis, the enantiopure (+)-18- methylenevincadifformine (125) was resolved from (+/-)-125 using (+)-(2R,3R)-tartaric acid (1:1) and its absolute configuration was determined by comparison of the ECD spectra (electronic circular dichroism spectra) data with (+)- and (-)-vincadifformine

(126) and (-)-tabersonine (115) (Table 2). Because the absolute configuration of (-)-126 was known in the literature, data from (+)-125 was consistent with that from (+)-126, the enantiomer of naturally occurring (-)-vincadifformine (126) and thus the absolute configuration of (+)-125 was confirmed as that in Scheme 22. Enantioenriched (+)-125 was transformed to 127, followed by allyl group manipulation to afford (-)-stremepliopine

20 (103). Although the specific rotation data of synthetic (-)-103 ([]578 -25.4 (c 1.8,

20 MeOH)) did not fit the reported figure ([]D -120 (MeOH)), Hájíček believed that the original data was reported with a misprint but unfortunately there was not enough of the

Chapter 2. Targets background and synthetic studies

authentic sample of (-)-stremepliopine (103) available to run a new optical rotation measurement. In 2018, Qin and co-workers reported an asymmetric total synthesis of strempeliopine (103) and obtained the specific rotation of -24 (c 0.23, MeOH),48 which was consistent with Hájíček’s observation.

Scheme 22. AC determination of vallesamidine, andrangine and strempeliopine

Table 2. ECD data of (+)-125 and naturally occurring (+)-and (-)-126 and (-)-115

alkaloids max [nm] ( )

(+)-125 324.5 (+31.3) 285.0 (-4.3) 236.0 (-17.5)

(+)-126 325.0 (+32.3) 285.0 (-2.4) 237.0 (-16.6)

(-)-126 324.0 (-35.7) 286.0 (+3.2) 237.0 (+17.2)

(-)-115 324.5 (-24.7) 288.0 (+6.1) 238.5 (+18.2)

Chapter 2. Targets background and synthetic studies

In 2007, Hájíček and co-workers used density functional theory calculations

(DFT) of vibrational circular dichroism (VCD), electronic circular dichroism (ECD) and optical rotation to determine the absolute configuration of (+)-schizozygine (6). The calculated specific rotations of both enantiomers of schizozygine under three different methods indicated that (2R, 7S, 20S, 21S)-6 was in a better agreement with the experimental data of (+)-6 (Scheme 23).

Scheme 23. AC determination of (+)-schizozygine using DFT calculation

2.4 Schizozygine, vallesamidine and strempeliopine: proposed biosynthesis

In general, all terpene indole alkaloids are believed to be synthesised biologically from strictosidine (107) and the bisosynthesis of strictosidine (107) involves an enzyme

(strictosidine synthase) catalysed Pictect-Spengler reaction between tryptamine (105) and secologanin (106) (see Scheme 20). Secologanin (107) itself is an iridoid terpenoid

Chapter 2. Targets background and synthetic studies

natural product and its biosynthesis (seco-iridoid pathway) was fully characterised and reported in 2014.34 How more structurally diverse indole alkaloids (e.g. corynanthe, ajmalan, quinoline, iboga and aspidosperma alkaloids) are derived from from strictosidine

(107) via biosynthesis is one of the most important questions in this area and recent studies have been summarised and reviewed.20

The vallesamidine, strempeliopine and schizozygine alkaloids belong to the aspidosperma class of alkaloids and their structures are hypothesised to come from the rearrangement of the aspidospermine skeleton. To exemplify the biosynthesis of the aspidosperma alkaloidsit is informative to look at the biosynthesis tabersonine (115). The proposed biosynthesis starts from strictosidine (107). It is first transformed into 4,21- dehydrogeissoschizine (110), via a deglycosylation (by -glucosidase), an intramolecular imine formation. Then a series of structural rearrangements and fragmentations delivers dehydrosecodine (132). Finally, an intramolecular Diels-Alder cycloaddition forms the aspidosperma alkaloid tabersonine (115).

Chapter 2. Targets background and synthetic studies

Scheme 24. Proposed biosynthesis of aspidosperma alkaloid tabersonine 115

Scheme 25. Le Men’s semi-synthesis of vallesamidine from tabersonine

In 1971, Le Men et al 35 reported that treatment of tabersonine (115) with hydrochloric acid afforded decarboxylated intermediate 133. A reductive rearrangement

Chapter 2. Targets background and synthetic studies

o of 133 in the presence of Zn and CuSO4 in acetic acid over 100 C gave the intermediate

137. Finally, N-methylation, followed by C-14,15 double bond reduction delivered vallesamidine (102). The biosynthesis of vallesamidine has not been disclosed but Le

Men’s synthesis suggests the possible biosynthetic route of vallesamidine from an aspidosperma skeleton and this proposal could also be considered in the schizozygine alkaloids’ biosynthesis.

2.5 Schizozygine, vallesamidine and strempeliopine: synthetic studies

Currently the total synthesis of schizozygine (6) has not been reported with more attentions being focused on the total synthesis of vallesamidine (102) and strempeliopine

(103.)

2.5.1 Reported synthesis on vallesamidine

After Le Men’s semi-synthesis of vallesamidine (102) from the reductive rearrangement of tabersonine (115), no other synthesis of this alkaloid was reported until Heathcock disclosed a landmark synthesis of (+/-)-vallesamidine in 1989 (Scheme 26). 36 The synthesis commenced with a Michael addition of 2-ethylcyclopentanone (138) to acrylonitrile. The resulting adduct 139 was reduced via hydrogenation to induce an intramolecular imine formation to afford the key intermediate 140. Treatment of 140 with o-nitrocinnamic acid (141) in refluxing dioxane initiated a diastereoselective

Michael addition/lactamisation cascade and formed a tricyclic intermediate 143, with the correct relative stereochemistry (dr > 95:5), via a proposed transition state 142.37 After nitro group reduction, the aniline 144 was subjected to an NBS mediated cyclisation and

Chapter 2. Targets background and synthetic studies

hydrolysis (AgNO3/MeOH-H2O) to yield a mixture of hydroxy and methoxy intermediates 146. The removal of the hydroxy/methoxy group and N-methylation to 147 were carried out in one-pot by treatment with sodium cyanoborohydride in aqueous acetic acid in the presence of formaldehyde. The last step of lactam reduction yielded the target

(+/-)-vallesamidine (102) in an impressive 8 overall steps.

Scheme 26. Heathcock’s total synthesis of racemic vallesamidine

After Heathcock’s work, Bergter et al reported an asymmetric synthesis of

Heathcock’s intermediate 140 (Scheme 27). 38 This work involved an asymmetric

Michael addition of chiral imine 148, prepared from 2-ethylcyclopentanone (138) and R-

(+)-1-phenylethylamine, to methyl acrylate to give the alkylated product S-(+)-149 in

1 70% yield. The H NMR using Eu(tfc)3 as a chiral shift reagent indicated an enantiomeric excess of 90%. The ketone in 149 was protected, and the ester group was then converted 31

Chapter 2. Targets background and synthetic studies

into the free amine 152. This was then exposed to 10% HCl aqueous solution to give

Heathcock’s intermediate 140 in enantioenriched form.

Scheme 27. Bergter’s asymmetric preparation of Heathcock’s intermediate 1.140

During his studies on vallesamidine (102),37 Heathcock discussed another type of ring construction through an intramolecular C-2 alkylation (Scheme 28). This approach was proved to be inapplicable because earlier work by Herlay-Mason39 in 1967 had reported that treatment of compound 153 with acid afforded the aspidospermine structure compound 157. This process may initially come from a C-2 alkylation, but the resulting intermediate 154 was labile and quickly underwent rearrangement to 155 which was trapped by a C-3 addition to achieve the aspidospermine type skeleton 157.

Scheme 28. Unsuccessful attempt to synthesise vallesamdine through C-2 alkylation 32

Chapter 2. Targets background and synthetic studies

Inspired by the results above, Okada et al reported a radical cyclisation to reach vallesamidine precursor 156.40 Commercially available (but expensive, 1g/136.96 USD,

ACROS chemicals, website visited 6 Aug. 2019) (S)--hydroxymethyl--butyrolactone was used as the starting material and 158 was prepared through a protection/double alkylation/deprotection sequences. Then 8-steps of functional group manipulations were required to give key building block 163 that was condensed with tryptamine to initiated a Pictet-Spenglar cyclisation to give a 1:1 mixture of diastereomers 164 and 165, from which the desired product 165 was isolated with 30% yield. The hydroxyl group of 165 was converted into the benzene selenide radical precursor 166, which was subjected to a n-Bu3SnH/AIBN radical cyclisation to afford vallesamidine precursor 156 in an excellent

91% yield.

Chapter 2. Targets background and synthetic studies

Scheme 29. Okada’s total synthesis of (-)-vallesamidine

Cycloadditions are highly efficient method to rapidly construct complex poly- cyclic systems and have been used in various natural product syntheses.41 In 1998, Padwa et al reported a formal synthesis of vallesamidine (102) by making Heathcock’s intermediate 143 through an intramolecular 1,3-dipolar cycloaddition of carbonyl ylide.42

A model study using compound 169 was carried out first (Scheme 30). Treatment of diazocompound 169 with a catalytic amount of rhodium (II) perfluorobutyrate afforded the carbonyl ylide 170 which underwent a 1,3-dipolar cycloaddition to provide 171. Ring

Chapter 2. Targets background and synthetic studies

opening of the furan, mediated by TMSOf, generated the tricyclic compound 172. Further transformations to 174 required a Barton-McCombie deoxygenation and a hydrolysis/decarboxylation sequence. Following the same route, compound 175 was converted into Heathcock’s intermediate 143.

Scheme 30. Padwa’s formal synthesis of vallesamidine

In 2017, Qin’s group reported a total synthesis of (-)-vallesamidine 102, together with another 32 monoterpene indole alkaloids using a visible light photocatalysed cascade radical cyclisation as a key step.43 Deprotection of 180 generated a free amine which then 35

Chapter 2. Targets background and synthetic studies

treated with 181 in the presence of acetic acid to give a one-pot imine condensation- lactamisation cascade to obtain the dehydro-lactam 182 in 75% yield over two steps

(Scheme 31). Nitro group reduction and formation of sulfonamide 183 was followed by silyl group removal and oxidation of the resulting alcohol to give conjugated aldehyde

184, which would be the radical acceptor in the key radical cascade cyclisation.

Intermediate 184 was exposed to blue LED in the presence of a catalytic amount of Ir(dtbbpy)(ppy)2PF6 to give nitrogen radical 185 which underwent the first 5-exo-trig cyclisation to intermediate 186 and then a second 6-exo-trig cyclisation to 187 which finally gave tetracyclic compounds 188 and 189 as a mixture of diastereomers at C-20 (dr

1:1.5). The desired diastereoisomer 189 was obtained in 49% yield. After protection of the aldehyde, the indoline in 190 was deprotected so that the resulting free indoline could be oxidised to indole 191 in excellent yield. The N-methylation and acetal hydrolysis of

191 regenerated the aldehyde (192) and a SmI2 mediated radical cyclisation, similar to that in Okada’s synthesis (166 to 156, Scheme 29), gave cyclised product 193. The redundant hydroxyl group was removed via Barton/McCombie deoxygenation to give the intermediate 147, which first appeared in Heathcock’s synthesis, and the target (-)- vallesamidine (102) was obtained after catalytic lactam reduction.

Chapter 2. Targets background and synthetic studies

Scheme 31. Qin’s total synthesis of vallesamidine

Chapter 2. Targets background and synthetic studies

2.5.2 Reported synthesis on strempeliopine

The alkaloid strempeliopine (103) is structurally very close to the schizozygine alkaloids and its total synthesis has been completed by Hájíček33, 44a Padwa46 and Qin48. Hájíček’s synthesis combined Kuehne’s biomimetic synthesis of aspidosperma alkaloids45 (Scheme

32-a) and Le Men’s semi-synthesis of vallesamidine (102)35 (Scheme 25) from tabersonine (115) and then to develop a biomimetic route to strempelipine (103). To summarise Kuehue’s approach, compound 195 was firstly synthesised from tryptamine

(105) with methyl pyruvate (194) through a Pictet-Spengler reaction and then condensed with aldehyde 196 to give a spiro-enaminonium intermediate 197. A Hofmann fragmentation of 197 generated the indoleacrylic ester 198 and a subsequent intramolecular Diels-Alder (IMDA) cycloaddition was conducted to deliver the aspidospermane alkaloid vincadifformine (126).

Hájíček et al changed the aldehyde from 196 to 199 and the condensation/Hoffman fragmentation/IMDA cascade on 195 was carried out to give compound 200 in 50% yield. Hydrolysis of the ester quantitatively yielded the corresponding carboxylic acid 201 and then a decarboxylation reaction proceeded to afford indolenine compound 202. With this intermediate in hand, Le Men’s conditions were adopted to induce the reductive rearrangement and a mixture of 203 and 204 was obtained (ratio not mentioned). Formylation of this mixture yielded the compound 205 in 33% yield from 202. Oxidative cleavage of the olefin in 205 by ozonolysis in aq. HCl and methanol initially achieved the indoline aldehyde 206 and then a formamide

Chapter 2. Targets background and synthetic studies

deprotection/oxidative cyclisation sequence using hydrogen peroxide was performedo complete the total synthesis of (+/-)-strempeliopine 103.

Scheme 32. Hájíček’s total synthesis of (+/-)-strempeliopine

Chapter 2. Targets background and synthetic studies

Scheme 33. Intramolecular 1,4-dipolar cycloaddition of cross-conjugated heteroaromatic

betaines

After the synthesis of (+/-)-vallesamidine 102 using a 1,3-dipolar cycloaddition of a carbonyl ylide (Scheme 30), the Padwa group developed an intramolecular 1,4- dipolar cycloaddition of cross-conjugated heteroaromatic betaines to construct poly- heterocyclic skeletons (Scheme 33) This was then applied to the total synthesis of (+/-)- strempeliopine. 46 Generally, the cross-conjugated heteroaromatic betaine (e.g. 208) could be generated through a condensation between a monoprotic amide or thioamide and a 1,3-bis-electrophile (e.g. carbon suboxide).The subsequent 1,4-dipolar cycloaddition with an internal olefin could lead to the divergent polyring system. Based on this strategy, double alkylation of -valerolactam 210 provided compound 211 and 1,2-dibromination of the terminal olefin, followed by a double elimination gave the alkyne 212 (Scheme 34).

Sonogashira coupling of 212 with 1-iodo-2-nitrobenzene 213 smoothly gave 214 in good

72% yield. As the intramolecular 1,4-dipolar cycloaddition required a cis-alkene to

47 . achieve the correct stereochemistry, Trost’s method (Pd2(dba)3 CHCl3, 1,1,3,3- tetramethyldisiloxane, AcOH) was used to obtain the cis-olefin with a 10:1 cis/trans ratio.

Changing of the protecting group from TBS to acetyl 215, followed by treatment with

Lawesson’s reagent gave the thioamide 216. Exposure of 216 to carbon suboxide 218, generated in situ by zinc reduction of dibromomalonyl dichloride 217, resulted in the cross-conjugated heteroaromatic betaine 219. Heating 219 to 200 oC induced the intramolecular 1,4-dipolar cycloaddition and the resulting cycloadduct 220 subsequently 40

Chapter 2. Targets background and synthetic studies

decomposed to 221 in 31% yield, by loss of COS. This Heathcock type intermediate (cf

143 in Scheme 26) was then converted to indoline 222 via Heathcock’s route to vallesamidine 102, and the remaining steps to (+/-)-strempeliopine 103 were straightforward.

Scheme 34. Padwa’s total synthesis of (+/-)-strempeliopine

Chapter 2. Targets background and synthetic studies

In 2018, after reporting the asymmetric synthesis of (-)-vallesamidine 102 via a key step of photocatalytic cascade radical cyclisation (Scheme 31), Qin’s group utilised the same strategy to complete the asymmetric total synthesis of (-)-strempeliopine 103

(Scheme 35).48

For the purpose of constructing the indoline fused lactam ring F (see 103, Scheme

35), the synthesis towards strempeliopine required the radical cyclisation precursor 223.

Photocatalytic radical cyclisation gave 224 in 75% yield, but with barely any diastereocontrol (3:2 dr at C-20). The two diastereomers 226 and 227 were isolated after oxidation to the indole and reduction of the lactam to the piperidine ring. Although the minor isomer 227 possessed the correct stereochemistry for the completion of the synthesis of (-)-strempeliopine (103), the major isomer 226 was used for the synthesis of memebrs of the eburnane alkaloids ((-)-eburnaminol, (+)-larutenine, (-)-terengganensine

B). With 227 in hand, nitrile group hydrolysis in conc. HCl/MeOH led to lactam formation and removal of benzyl group. Oxidation to aldehyde 228 and then SmI2 mediated radical cyclisation furnished the last five-membered carbocycle (229) and the remaining hydroxyl group was removed via a two-step, radical deoxygenation to give (-

)-strempeliopine 103.

Chapter 2. Targets background and synthetic studies

Scheme 35. Qin’s asymmetric total synthesis of (-)-strempeliopine

Chapter 2. Targets background and synthetic studies

2.5.3 Reported synthetic studies towards schizozygine

Compared with vallesamidine (102) and strempeliopine (103), schizozygine (6) has an additional double bond between C-14 and C-15. However, the reported total syntheses of vallesamidine (102) and strempeliopine (103) did not involve the construction of a C-

14,15 double bond, instead they installed a single bond from the beginning and this limited the direct use of these reported routes for the synthesise schizozygine (6).

In 2005, Hájíček reported an unsuccessful attempt to introduce the C-14,15 double bond via C-15 hydroxylstrempeliopine (237). Following the same route of making (+/-)- strempeliopine (103) (see Scheme 32), compound 233 was prepared from 230 while the

IMDA process gave 233 in a low yield of 22% (Scheme 36). After ester hydrolysis and decarboxylation, the resulting compound 234 was subjected to the reductive rearrangement to give 235. The15-ketostrempeliopine (236) was then prepared through an oxidative cleavage/cyclisation reaction. Reduction of the ketone yielded the alcohol

237, but the dehydration of this alcohol via tosylation failed to introduce the final double bond (other dehydration methods were not mentioned). The unsuccessful dehydration through -elimination was probably due to the conformation of the molecule disfavouring an anti-periplanar conformation requires.

Chapter 2. Targets background and synthetic studies

Scheme 36. Hájíček’s unsuccessful attempt on C14,15 double construction

Based on this model studies, Hájíček then moved on to another synthetic plan towards (+/-)-schizozygine (6) (Scheme 37). 49 It is worth noting that the condensation/IMDA process from 241 to 242 performed much better than that from 230 to 233 (Scheme 36). After the reductive rearrangement, key intermediate 244 was obtained. Compared to compound 234, the presence of the dioxolane ring caused many problems. The oxidative cleavage of the allyl group in either 244 or its formamide under ozonolysis conditions failed to provide the corresponding aldehyde. Lemieux-Johnson oxidation of 244 led to the isolation of ‘back-rearranged’ aspidosperma type structure

245. The resulting diol was not cleaved but the primary hydroxyl group cyclised onto the

C-15 ketone to form a pyran ring. When the free indoline was protected as its formamide, dihydroxylation using OsO4 exclusively yielded 246 without rearrangement.

Additionally, changing the oxidation method to Sharpless dihydroxylation conditions resulted in a clean conversion back to 243 (80%) and did not touch the allyl group. The authors also proposed that nosylation of indoline nitrogen could stabilise the molecule in

Chapter 2. Targets background and synthetic studies

oxidation step. Surprisingly, treatment of 244 with 2-nitro-benzenesulfonyl chloride (2-

NsCl) and i-PrNEt2 directly afforded 243, presumably due to a fragmentation/ rearrangement of 247 and the formation of intermediate 250. We have mentioned that

(Scheme 28), without reducing agent (e.g. Zn), the C-3 addition to indole was much more favoured and only aspidosperma type structure could be obtained.

Scheme 37. Hájíček’s studies towards schizozygine (6)

Chapter 2. Targets background and synthetic studies

Scheme 38. Proposed solution on Hájíček’s work

Despite these unexpected results the transformation from 244 to 246 provided the possibility to carry out the oxidative cleavage of the 1,2-diol derived from the allyl group and continue the synthesis. However, no more updates have been reported since 2011.

For example, other protecting groups (e.g. Boc, CO2Me) could be investigated so that the protected product 251 could undergo reduction of the C-15 ketone and then oxidative cleavage to the aldehyde 254. Alternatively, dihydroxylation of 251 would yield compound 252 and the pyran ring could also be ring opened under reductive conditions.

Among the schizozygine alkaloids, the structurally rearranged analogue, isoschizogamine (255) was also isolated and attracted much more attention as a synthetic target than schizozygine (6) itself.

Chapter 2. Targets background and synthetic studies

Scheme 39. Isoschizogamine and proposed biosynthesis relationship with schizozygine

Currently, several total syntheses have been reported on isoschizogamine (255).

The first total synthesis was reported by Heathcock in 1999,50 through modifying the route from his previously reported synthesis on vallesamidine (102) (Scheme 40). A new cyclic imine 263 was prepared and its reaction with 265 through a Michael addition/cyclisation cascade reaction sequence gave tricyclic product 266 in good yield and diastereoselectivity (74% yield, dr 88:12). Dehydration of 266 using Martin’s sulfurane proceeded smoothly to afford the olefin compound 267. The nitro group was then reduced and the resulting aniline 268 underwent lactam reduction and cyclisation to give aminal 269 after work-up. Unfortunately, the C-2 stereochemistry was incorrect but treatment of 269 with aqueous acid at room temperature provided the desired 270 as the

Chapter 2. Targets background and synthetic studies

thermodynamically more stable major product. Warming the acidic solution to reflux hydrolysed the dioxolane to the aldehyde and in situ cyclisation with the indoline amine gave hemiaminal 271. Final PDC oxidation afforded isoschizogamine 255 in 27% overall yield from 267.

Scheme 40. Heathcock’s total synthesis of (+/-)-isoschizogamine 255

Looking at Heathcock’s syntheses of vallesamidine (102) and isoschizogamine

(255), treatment of 144 with NBS followed by AgNO3 in methanol afforded the 1,2- diamine moiety (146) (Scheme 41), while in contrast treatment of the similar intermediate

268 with LiAlH4 yielded the aminal 270 (Scheme 40). Conversion of 144 to 146 is an

Chapter 2. Targets background and synthetic studies

elegant and divergent strategy, but Heathcock did not report using the conditions NBS then AgNO3/MeOH on intermediate 274 to synthesise a schizogamine type skeleton 273.

Scheme 41. Divergent transformation to vallesamidine and isoschizogamine from Heathcock’s

tricyclic intermediate

Results and discussion

➢ Studies on the total synthesis of schizozygine

➢ Total Synthesis of (+)-vallesamidine

➢ Studies towards the total synthesis of (+)-strempeliopine

Chapter 3 Studies on the total synthesis of schizozygine

Chapter 3. Studies on the total synthesis of schizozygine

3.1 Retrosynthesis

The retrosynthesis of schizozygine (6) was shown in Scheme 42. The C-14,15 double bond in ring D could be constructed through a ring closing metathesis and an intramolecular amide condensation could form ring F. The -vinyl ester moiety in precursor 274 could be generated from allyl alcohol 275 via a Johnson-Claisen rearrangement while the allyl alcohol could be achieved from ketone 276. A nitro group reduction/intramolecular C-N coupling sequence would be applied on 277 to synthesise indoline 276. For ring E synthesis, a novel phosphine catalysed cyclisation from vinyl ketone 278 was proposed and the vinyl ketone moiety could be converted from ester 279.

Piperidine ring C was planned to be prepared from 280 through a reductive amination and this highly functionalised -nitro amino intermediate could be synthesised from a nitro-

Mannich reaction with compound 281, precedence for which was known in the literature.

3.2 Feasibility and literature precedents

The metathesis reaction on amine substrates in many cases are problematic because the basic nitrogen can chelate to the metal centre and sequester the catalyst. Impressive results were reported on the total synthesis of meloscine 28551 and indicated that the higher complexity of some molecules could ‘hide’ the nitrogen atom and decrease the chance of catalyst posioning caused by amine substrates. Specifically, late-stage ring closing metathesis reactions were performed on complex intermediates 284 and 286 to

Chapter 3 Studies on the total synthesis of schizozygine deliver the corresponding cyclised products 285 and 287 with excellent yields. The amine nitrogen did not show a negative effect on the catalysis (Scheme 43).

Moreover, in Batch’s synthesis,52f the compound 286 was synthesised through the

Johnson-Claisen rearrangement of allyl alcohol 290, which was constructed by HWE olefination/reduction sequence of ketone 289 (Scheme 43). These transformations provided excellent precedents and could support our proposal in late-stage manipulations

(Scheme 42).

Scheme 42. Retrosynthesis of schizozygine

Chapter 3 Studies on the total synthesis of schizozygine

Scheme 43. Precedents for ring closing metathesis in alkaloids synthesis and John-Claisen

rearrangement to -vinyl ester moiety.

As one of the key steps, the phosphine catalysed cyclisation to the five-membered ring E (278 to 277) was novel and only one example was reported based on the same idea but with different substrates. 52 As shown in Scheme 44, upon exposure to tributylphosphine (Bu3P), substrate 291 underwent a 5-endo-trig type cyclisation to give pyrrolidine 293 while using DBU exclusively yielded 292 through a 5-exo-trig pathway.

In terms of mechanism (Scheme 45), the transformation from 291 to 293 is not a real 5- endo-trig cyclisation. It was postulated that tributylphosphine first performed a 1,4- addition on the conjugate ester and the resultant enolate 294 underwent an intramolecular proton transfer to generate the more stable carbanion 295. Ring closure then occurred

Chapter 3 Studies on the total synthesis of schizozygine through a 5-exo-tet cyclisation to afford the pyrrolidine 293. Our proposed transformation on nitrovinylketone 278 would follow a similar mechanism by forming a more stable nitronate from an enolate and then cyclisation to the cyclopentanone product

(Scheme 45).

Scheme 44. Reported phosphine catalysed 5-endo-trig cyclisation

Scheme 45. Proposed mechanism of phosphine catalysed 5-endo-trig cyclisation

Chapter 3 Studies on the total synthesis of schizozygine

The preparation of complexed nitro amine 280 with syn-anti stereochemistry would come from a nitro-Mannich reaction on 281. Our group has experience in diastereoselective nitro-Mannich reactions to synthesise syn-anti -nitroamines. For example, starting from various nitrostyrenes 296, a cascade conjugate addition/nitro-

Mannich reaction method was developed to achieve various -nitroamines 298 with the syn-anti isomer as the major product (Scheme 46).53

Scheme 46. Reported conjugate addition/nitro-mannich reaction to syn,anti -nitroamines

Instead of using conjugate addition to the nitrostyrene to generate the corresponding nitronate with a chiral centre at the -position, the chiral nitronate (e.g.

300) could be prepared by deprotonation of the chiral nitroalkane 299. In general, for a chiral nitronate with two substituents of different sizes, treatment of the nitronate with an imine and acid (e.g. trifluoroacetic acid, TFA) at low temperature, induces the nitro-

Mannich reaction and achieves the kinetic syn,anti product (RL syn to NO2 and NO2 anti to R’), following the proposed closed, chair transition state TS-I (Scheme 46). Therefore, under the same nitro-Mannich condition, compound 281, in which the aryl group is larger

Chapter 3 Studies on the total synthesis of schizozygine than the carbon chain, could undergo the desired transformation with the correct control of stereochemistry.

Moreover, when using 2-bromonitrostyrene as the starting material, our group developed a reductive nitro-Mannich sequence to synthesise anti-nitroamines (Scheme

47).77 Conjugate reduction of the styrene followed by addition of an imine and acid promoter gave, after protection of the resultant amino group as its trifluoroacetamide, nitroamine 303. Reduction with zinc powder and hydrochloric acid in EtOAc/EtOH gave the TFA migrated amine product (a), which upon treatment with KOH hydrolysis followed by Pd(0) catalysed intramolecular C-N coupling reaction to afford indoline structure 304. This well-developed approach provided us with the precedent to support the total synthesis plan of schizozygine (277 to 276, Scheme 42).

Scheme 47. Reductive nitro-Mannich reaction and subsequent manipulation to indolines

Chapter 3 Studies on the total synthesis of schizozygine

3.3 Synthetic studies

The synthesis commenced by converting piperonal 305 to the corresponding - unsaturated aldehyde 282 via a bromination, Horner–Wadsworth–Emmons (HWE) olefination, DIBAL-H reduction and MnO2 oxidation sequence (Scheme 48). An asymmetric conjugate addition of nitromethane catalysed by the pyrrolidine (Jørgenson-

Hayashi) catalyst 308 was conducted54 and then the aldehyde 309 was protected to the dioxolane 310 (98% ee),i the precursor for the nitro-mannich reaction.

Scheme 48. Preparation of chiral nitroalkane 310

With the nitroalkane 310 in hand, we then investigated the nitro-Mannich reaction.

Based on published results, the nitronate should be fully and irreversibly prepared before adding the imine to perform the nitro-Mannich reaction. Therefore, following the developed protocol in group, n-BuLi was used first. Although the pKa (thermodynamic

i The ee of Compound 310 was measured using chiral HPLC by C. Rundell (C. Rundell, PhD Thesis,

University College London, 2016).

Chapter 3 Studies on the total synthesis of schizozygine acidity) of the proton adjacent to the nitro group is low (~10) and substituted nitroalkanes

(compared with nitromethane) exhibit lower pKas due to increased stabilisation of the nitronate, the kinetic acidity shows the opposite trend. As shown in Table 3, with one methyl group on nitromethane (i.e. nitroethane), the kinetic acidity is 6 times lower than that of nitromethane and 2-nitropropane gives 100 times lower kinetic acidity than nitromethane. The steric issue from substituents shows a predominant effect on the proton removal and this effect is much more significant than carbonyl compounds. For complex nitroalkane 310 we thought that the deprotonation might be challenging and would need to be carefully investigated.

3.3.1 Kinetic Nitro-Mannich reaction

Indeed, the initial attempt of the nitro-Mannich reaction (310 to 313) did not proceed very well. Deprotonation was conducted with n-BuLi in THF at -78 oC and the reaction time was screened from 30 min to 3 hours. With 30 min deprotonation time no conversion was observed. A longer 3 hours deprotonation at -78 oC followed by imine

312 addition gave nearly full conversion, but afforded the crude reaction mixture containing several products with only a small amount of desired nitroamine 313,

1 identified in the H NMR by comparing the peak of CHNO2 [ 5.35 ppm (dd, J = 10.3,

4.5 Hz, 1H),] with that of a model compound 314 [ 5.06 ppm (dd, J = 11.6, 2.4 Hz)]

(Scheme 49). In addition, halogen-lithium exchange was also possible in the reaction ii, thus we then tested other basic conditions for the nitro-Mannich reaction.

[ii] Complications from the halogen-lithium exchange was later confirmed by nitro-Mannich reaction of non-bromo substrate 322, see page 63.

Chapter 3 Studies on the total synthesis of schizozygine

Table 3. Kinetic and thermodynamic acidity of nitroalkanes

-1 -1 Nitroalkanes Kinetic acidity k (M min ) Thermodynamic acidity (pKa)

CH3NO2 238 10.2

CH3CH2NO2 39.1 8.5

(CH3)2CHNO2 2.08 7.7

Scheme 49. Initial attempt on nitro-Mannich reaction

Use of the LDA/DMPU system did not give good results. The 1H NMR of the crude reaction mixture showed low conversion and no desired nitroamine. The nitro-

Mannich reaction using LiHMDS as a base was also tested. Repeated experiments showed consistent results of 40% to 50% conversion but to give a mixture of unknown compounds or undesired isomers.

A deuteration experiment was carried out to check if the deprotonation was taking place. Deprotonation of 310 with LiHMDS from -78 oC to -10 oC for 2.5 hours was quenched with deuterated acetic acid (AcOD) followed by work up with deuterated water

1 (D2O). The H NMR showed around 40% deprotonation. However, increasing the temperature on deprotonation from -10 oC to room temperature in 2.5 hours did not improve the deuteration.

Chapter 3 Studies on the total synthesis of schizozygine

Table 4. Deuteration experiment to quantify the extent of deprotonation

Base Equivalent (to 3.37) Temperature Deuteration (%)

LiHMDS 1.5 -78 to -10 oC 40

LiHMDS 1.5 -10 oC to rt 40

Additionally, the PMP protected imine can result in the formation of a very basic nitrogen anion (pKa > 30). Potentially this could re-deprotonate the amines like i-Pr2NH or (TMS)2NH to generate the amide base again (Scheme 50) which could cause additional complications. We then attempted the nitro-Mannich reaction under catalytic amounts of

LiHMDS, NaHMDS or a strong organic base BEMP but all of these attempts gave no reaction.

Scheme 50. Proposed catalytic nitro-Mannich reaction

At this stage, we decided to go back to the n-BuLi system as the n-BuLi could give full deprotonation and the resulting butane by-product is volatile. To avoid halogen- lithium exchange and reduce the steric hindrance, we planned to prepare the non-bromo

Chapter 3 Studies on the total synthesis of schizozygine nitro compound 322 (Scheme 51) and install the bromine at a later stage or consider using direct annulation through C-H functionalization.

A literature search for the bromination (Scheme 51) gave more than 50 precedents with good yields and regioselectivity. Moreover, intramolecular C-N coupling via Pd (II) catalyzed C-H activation developed by Yu55 and his co-workers can regioselectively construct the indole structure (315 to 316) which led us to consider the indoline synthesis without the bromine (Scheme 51).

Scheme 51. Literature precedents on direct amination and C-N coupling and preparation of the

single enantiomer non-bromo nitroalkane 322

Chapter 3 Studies on the total synthesis of schizozygine

Scheme 52. Preparation of racemic nitroalkane 322

Therefore, we went back to the first step and took piperonal 305 to prepare the new precursor 322 following the same reaction sequence as for the preparation of 310

(Scheme 48). To save the expensive asymmetric catalyst usage, racemic 322 was also synthesised (Scheme 52) and used in the later research unless stated.

The nitro-Mannich reaction with imine 325 was carried out on 322 using the n-

BuLi-TFA protocol (n-BuLi, -40 °C, 50 min; imine, -78 °C, 30 min then TFA, - 78 °C to rt) and 90% conversion was observed from 1H NMR as a mixture of two major isomers with a ratio of 2:1 (Scheme 53). The HRMS (ESI) gave the correct molecular weight and formula but the relative stereochemistry was difficult to define.

Chapter 3 Studies on the total synthesis of schizozygine

Scheme 53. Nitro-Mannich on non-bromo nintroalkane

The nitro-Mannich reaction of 322 with iminoester 312 yielded a clean mixture of diastereomers 327 with a ratio of ~2:3 (Scheme 53). The stereochemistry of the minor isomer was confirmed to be the desired syn-anti product, compared with reported similar compound 314 (Scheme 49). This mixture was then subjected to the reductive cyclisation using the precedented conditions of TMSOTf(cat.)/Et3SiH system, originally developed by Noyori.56 Even though this experiment gave a very poor yield (< 20%), two cyclised isomers were isolated and NMR experiments suggested syn-anti and anti-anti piperidines respectively, according to the coupling constants exhibited by the CHNO2 proton. The anti-anti piperidine 328-a has a favoured conformation with three substituents equatorial and three contiguous protons axial. Therefore, the proton  to the nitro group would be predicted to be a doublet of doublet (dd) with large coupling constants and this was consistent with the observed J values (dd, 9.2 and 11.2Hz). For syn-anti piperidine 328- b, the favoured conformation has the aryl group equatorial. Although this conformer has

Chapter 3 Studies on the total synthesis of schizozygine two axial substituents, the hyperconjugation between the lone pair on nitrogen and the anti-bonding orbital of the C-CO2Et bond provides additional stabilization (n-* interaction).57 In this case, the proton  to the nitro group should be a doublet of doublet or apparent triplet with small coupling constants. The 1H NMR spectrum of this isomer showed the splitting pattern that matched the prediction of syn-anti piperidine very well

(app.t, 2.5 Hz).

The poor diastereoselectivity of this nitro-Mannich reaction decreased the efficiency of the total synthesis and needed to be improved. We first changed the protecting group in 322 from the dioxolane to a dimethyl acetal 329 because we also observed that the dioxolane was difficult to hydrolyseiii and the dimethyl acetal should be easier. Treatment of 322 with camphor sulfonic acid (CSA) in refluxing methanol gave

329 in 94% yield. Subsequent nitro-Mannich reaction of 329 with imine 312 afforded a mixture of two major diastereomers (2:1 dr) with over 90% conversion in favour of the syn-anti isomer 330 (Scheme 54).

Scheme 54. Nitro-Mannich on dimethyl acetal analogue 329

We have discussed the origin of the diastereoselectivity in Scheme 46, based on the proposed transition state TS-I. This transition state follows the Felkin-Ahn model of

iii Various acid hydrolysis conditions were attempted but failed to afford corresponding aldehyde.

Chapter 3 Studies on the total synthesis of schizozygine carbonyl group (331) addition because electrophile trapping on electron rich polar alkene, such as a nitronate anion, also fits a Burgi-Dunitz like angle of attack (332) and the published examples in the group53 also matched this model very well (Scheme 55).

Currently, there is no proper model to explain the diastereoselectivity of the nitro-

Mannich reaction on the nitronate bearing an adjacent chiral centre. The anti-selectivity in acid promoted nitro-Mannich reaction through a chair like transition state is widely accepted and the controversial point is how the chiral elements control the facial selectivity of the nitronate to approach the imine. The staggered, Felkin-Ahn type model proposed by us explained the observed syn-anti selectivity while some other examples considered an eclipsed model with minimised 1,3-allylic strain (333)58 to account for selectivities. For example, protonation of nitronate 334 with acetic acid at -100 oC in

THF gave nitroalkane 335 with 95% dr. The facial selectivity was explained by the conformation as shown with minimised 1,3-allylic strain and the protonation occurred from the less hindered top face. Similarly, Barbas reported a one-pot conjugate addition/nitro-Mannich protocol to synthesise functionalised piperidines 341.59 When the chiral nitroalkane 339 was formed, treatment with imine 340 in the presence of acetic acid and triethylamine afforded the piperidine 341 through the proposed transition state, based on minimised 1,3-allylic strain (TS-II type).

We have known that, from the deprotonation of nitroalkanes, the reactions on nitroalkanes are highly dependent on the steric effect and the kinetic rates are dramatically decreased on substituted nitroalkanes. On the other hand, the nitronates have a much lower reactivity than enolates and thus the outcome of the nitro-Mannich reaction at low temperature would be more affected by external or intermolecular steric repulsion with imines.

Chapter 3 Studies on the total synthesis of schizozygine

For the above mentioned nitronate protonation, the in-coming proton is very small and there is no significant energy difference when approaching the nitronate through either TS-I or TS-II. Therefore, 1,3-allylic strain predominated and controlled the actual diastereoselectivity. For a bigger electrophile like an imine, reaction at room temperature could provide enough energy to overcome the steric repulsion with RS group and ultimately the 1,3-allylic strain became the main factor to keep the favoured nitronate conformation in the transition state.

However, for the nitro-Mannich reaction at low temperature, the steric repulsion between the nitronate and in-coming imine would be a major factor. Under this circumstance, the Felkin-Ahn-like model could create more space for the imine and the additional antiperiplanar effect from C-RL could stabilise the transition state. In model

333, although the 1,3-allylic strain was minimised, steric repulsion between the imine and the RS group may make this transition state model less likely.

Chapter 3 Studies on the total synthesis of schizozygine

Scheme 55. Analysis of stereoselectivity of nitro-Mannich reactions.

In our case, we assumed that the aryl substituent was the largest group and indeed, the syn-anti product was the major isomer in the reaction, although the dr (2:1) was moderate.

As the size of the protecting group would probably not make much difference due to it being someway from the reactive centre, we proposed that activation of the imine by

Lewis acid could help tighten the TS and affect the diastereoselectivity. The original conjugate adiition/nitro-Mannich methodology (Scheme 46) involved a catalytic amount of Cu(OTf)2 and a stoichiometric amount of Zn(TFA)2, which may have affected the diastereoselectivity of the subsequent nitro-Mannich reaction according to the transition

Chapter 3 Studies on the total synthesis of schizozygine states proposed. We then investigated if the Lewis acid could improve the nitro-Mannich reaction.

Firstly, Cu(OTf)2 was used in the desired nitro-Mannich reaction. A stoichiometric amount of Cu(OTf)2 was dissolved into THF followed by addition of imine

312 in THF at low temperature to prepare a solution of “activated” imine 342 (Scheme

56). This solution was then slowly added to the solution of nitronate 324 in THF, which was prepared by deprotonation of 322 using n-BuLi in advance, at -40 °C. To our delight, the 1H NMR of the crude reaction mixture showed ~4:1 dr and 327 could be isolated from unreacted starting material by chromatography as a mixture of two diastereomers and a small amount of excess imine (Fig. 1a). This condition was then repeated on dimethyl acetal compound 329 and the syn-anti isomer 330 was isolated as a single diastereomer in 65% yield (Scheme 56 and Fig. 1b for 1H NMR spectra).

The copper chelated iminoester 342 had a higher activity (electrophilicity) then

TFA activated one and the reaction could be carried out at low temperature without warming up. In addition a metal centred transition state may be tighter and more sensitive to non-bonded interactions in competing developing transition states. Whatever the exact reason, better diastereoselectivities were provided and made the transformation more potent for our total synthesis.

Chapter 3 Studies on the total synthesis of schizozygine

Scheme 56. Cu(OTf)2 mediated nitro-Mannich reaction

desired isomer diastereomer (syn-anti)

1 Figure 1. H-NMR of Cu(OTf)2 mediated nitro-Mannich on 322 and 329 70

Chapter 3 Studies on the total synthesis of schizozygine

To further mimic the reaction of 296 to 298 shown in Scheme 46, we then used

Zn(TFA)2 as a Lewis acid to mediate the nitro-Mannich reaction (Scheme 57). As

Zn(TFA)2 was not commercially available, a THF solution of Zn(TFA)2 was prepared by adding two equivalent of TFA to a solution of ZnEt2 in THF. The nitro-Mannich reaction was conducted by adding the imine solution to the pre-made nitronate intermediate in

THF at -78 °C, then Zn(TFA)2 solution was added slowly. This method gave 330 with good dr of 7:1 and 70% isolated yield at gram scale.

Scheme 57. Zn(TFA)2 mediated nitro-Mannich reaction

In this protocol, the addition rate of the Zn(TFA)2 solution was crucial as a fast addition gave a worse dr down to 2.5:1 but a good dr was reproduced by keeping the addition rate slowiv. This was probably due to the change of the internal temperature when the Zn(TFA)2 solution was added. Therefore, we may predict that an even higher dr would be obtained if the transfer of the Zn(TFA)2 solution could be conducted through a “cold to cold” method via a cannula if applicable.

iv A consistent addition rate was achieved by using a syringe pump (5.0 mL/15 min)

Chapter 3 Studies on the total synthesis of schizozygine

3.3.2 Reductive cyclisation to piperidine ring C

With the syn-anti nitro-amine 330 in hand, we then turned to the reductive cyclisation to the piperidine ring C (328). We knew that the TMSOTf (cat.)/Et3SiH conditions gave a very poor yield (Scheme 53) and thus more conditions were screened. The Lewis acid

BF3·OEt2 was then used on 330 instead of TMSOTf and the desired syn-anti piperidine

328 was obtained in 18% yield as well as a side product 344 in 19% yield (Scheme 58).

Mechanistically, BF3·OEt2 activation could transform the dimethyl acetal 330 to the oxonium ion 345, cyclisation by the amino group (345 to 346, then 347) and further reduction to 328. Compound 344 was probably formed by competing reduction of the oxonium ion 345 by Et3SiH before it was cyclised.

Scheme 58. Reductive cyclisation and proposed mechanism for the formation of undesired 344

To avoid the undesired formation of 344, we planned to perform the hydrolysis/cyclisation first (330 to 346 or 348, Scheme 59) then reduce the resulting intermediate to the desired piperidine 328. To our surprise, the dimethyl acetal 330 was robust to different aqueous and non-aqueous acid conditions (Table 5). Other attempts, for example, Lewis acids like TMSOTf and TESOTf/2,4,6-collidine system did not push 72

Chapter 3 Studies on the total synthesis of schizozygine

the cyclisation or hydrolysis to occur. The conditions of BF3·OEt2 in CH2Cl2 afforded the cyclised product tetrahydropyridine (THP) 348 but in only 25% isolated yield plus degradation. Changing the solvent to THF did not give any desired product but caused slow degradation. The BBr3 mediated cyclisation gave a better result with 35% yield of

348 after chromatography. Additionally, 330 was also cyclised to 348 in 20 % yield in refluxing toluene (120 °C) with a catalytic amount (0.2 eq.) of pyridinium p- toluenesulfonate (PPTS). Changing the amount of PPTS did not improve the yield of the cyclisation.

Scheme 59. Hydrolysis/cyclization of 330

Table 5. Screen of conditions for hydrolysis/cyclisation

acid or Lewis acid solvent temperature time yield

TFA (1.0 eq.) THF r.t. 6 h n.r.

TFA (1.0 eq.) THF 60 °C 6 h n.r.

TFA (ex.) H2O/THF r.t. 6 h n.r.

TFA (ex.) H2O/CHCl3 r.t. 6 h n.r.

CSA (1.0 eq.) THF r.t. 6 h n.r.

TESOTf/collidine CH2Cl2 -10 °C to r.t. 6 h n.r.

TMSOTf (1.0 eq.) CH2Cl2 -10 °C to r.t. 1 h c.m.

BF3·OEt2 (3.0 eq.) CH2Cl2 -20 °C to r.t. 1 h 348 25 %

BF3·OEt2 (3.0 eq.) THF -78 °C 1 h slow degradation

BBr3 CH2Cl2 -78°C 1 h 348 35 %

Chapter 3 Studies on the total synthesis of schizozygine

TsOH (0.5 eq.) PhMe 90 to 120 °C 3 h degradation

PPTS (0.5 eq.) PhMe 120 °C 5 h 348 20 %

PPTS (1.0 eq.) PhMe 120 °C 1.5 h 348 23 % r.t. = room temperature; n.r. = no reaction; c.m. = complexed mixture;

We also observed that the tetrahydropyridine 348 degraded during bench storage at room temperature and we thought that the tetrahydropyridine structure was unstable during column or aqueous work-up probably due to enamine hydrolysis. Thus, a one-pot

BBr3 mediated cyclisation, followed by reduction using Et3SiH afforded the piperidine

. 328 in 50 % overall yield (Scheme 60). The BF3 OEt2 mediated cyclisation needed a higher reaction temperature to -10 °C but the overall yield after reduction went up to 65 % at one-gram scale.

Scheme 60. Reductive cyclisation under sequential cyclisation/reduction protocol.

3.3.3 Ethyl ester manipulation to vinylketone

At this stage, the next goal was the ester manipulation of piperidine 328 to the vinylketone

351 (Scheme 61). The first strategy to form Weinreb amide 349 did not work with the starting material recovered and no epimerization was observed in this reaction. We then

Chapter 3 Studies on the total synthesis of schizozygine tried the direct addition of vinylmagnesium bromide (vinylMgBr) to ester 328, based on the possibility that the adjacent nitrogen atom could stabilise intermediate 350 by coordination, similar to the way a Weinreb amide is believed to work. Unfortunately, there was no reaction even with the additive CeCl3 and other minor modifications.

Scheme 61. Vinylketone 351 synthesis through ester addition

Other transformations to prepare the vinylketone 351 were attempted. Weinreb amide can also be synthesised by amide coupling from carboxylic acids. Accordingly the hydrolysis of ester 328 was investigated. The LiOH/THF-H2O conditions hydrolysed the ester slowly but with epimerization of either C-1 or C-2 (Scheme 62). A more nucleophilic but less basic saponification using potassium trimethylsilanolate (TMSOK) was attempted. Although the ester 328 was consumed quickly by TLC, 1H NMR analysis of the crude hydrolysed product was complicated. Assuming some potential formation of the desired carboxylic acid potassium salt 353, the crude TMSOK hydrolysed product was treated with N,O-dimethylhydroxylamine hydrochloride under amide coupling condition (EDCI/HOBt/DIPEA). The Weinreb amide 349 was not found, but instead an elimination product 355. A possible mechanism could be that after treatment of the

Chapter 3 Studies on the total synthesis of schizozygine carboxylic acid potassium salt 353 with EDCI/HBOT, an activated ester 354 was obtained which reacted with amine and underwent a decarboxylative elimination to give 355.

Alternatively, the carboxylic salt 353 may be unstable upon stirring and directly undergo the decarboxylative elimination.

Scheme 62. Attempt on Winreb amide 2.74 synthesis through ester hydrolysis-amide coupling

We then sought to investigate the synthesis to vinylkentone 351 from aldehyde

356. Attempted DIBAL-H reduction to 351 gave a degraded product which could not be identified. The ester group in 328 was reduced smoothly to the corresponding alcohol

358 with LiBH4 in 90% yield.

Chapter 3 Studies on the total synthesis of schizozygine

Scheme 63. Attempted vinylketone synthesis through aldehyde

However, re-oxidation of the primary alcohol 358 to aldehyde 356 was problematic. A series of conditions were tested but none of them gave satisfactory results

(Table 6). Hypervalent iodine oxidation using IBX or DMP (entry 1 and 2) gave complex mixtures (c.m.). Metal oxidations such as perruthenate (entry 3) and PDC (entry 10) also produced complex mixtures. These negative, non-chemoselective oxidations might be complicated from products derived from the electron rich PMP group. Different DMSO mediated oxidations were carried out. Parikh–Doering oxidation (entry 4) and Pfitzner–

Moffatt type oxidation (entry 5) led to no reaction, probably due to the difficulties in activating the hindered alcohol. The Swern oxidation, using highly active (COCl)2, gave around 10 % conversion (entry 6) but higher loadings of the reagent performed worse.

Similar results were observed when TFAA was used as an activator (entry 8 and 9).

Table 6. Screened oxidation conditions on compound 358

Entry Condition Time Result

1 IBX (3.0 eq.), EtOAc, 80 °C 1.5 h c.m. 1

2 DMP (3.0 eq.), CH2Cl2, rt 1.0 h c.m.

3 TPAP, NMO, 4A M.S. CH2Cl2, rt 8 h c.m.

Chapter 3 Studies on the total synthesis of schizozygine

2 4 SO3·Py, DMSO, Et3N, CH2Cl2, rt 16 h n.r.

5 EDCI, DMSO, TFA·Py, CH2Cl2, rt 16 h n.r.

(COCl)2 (1.2 eq.), DMSO (1.2 eq.), Et3N 6 1.5 h 10% conv. (5.0 eq.), CH2Cl2, -78°C to rt

(COCl)2 (5.0 eq.), DMSO (8.0 eq), Et3N 7 2 h c.m. (16.0 eq.), CH2Cl2, -78°C to rt

TFAA (3.0 eq.), DMSO (5.0 eq.), Et3N 8 3 h 40% conv. (10.0 eq.), CH2Cl2, -78°C to rt

9 TFAA (7.0 eq.), DMSO (11.0 eq.), Et3N 2.5 h c.m.

(10.0 eq.), CH2Cl2, -78°C to rt

10 PDC, CH2Cl2, rt 16 h c.m. c.m. complexed mixture; n.r. no reaction;

Scheme 64. Protective groups exchange

As we could not oxidise the alcohol 358, we then tried to change the protecting group to a potentially more stable one. Oxidative cleavage of the PMP group using CAN afforded the amine 359 in high yield (> 80%) but attempted re-protection using Boc2O led to the N-Boc, de-nitro compound 360 as a major product. Presumably the Boc group

Chapter 3 Studies on the total synthesis of schizozygine

makes the ester -proton more acidic (see 361) so that Et3N can mediate the -elimination of the nitro group. A quick protection using a TFAA/pyridine protocol protected the amine 359 to 362 cleanly at 0 oC and avoided any elimination. Unfortunately, the ester group in 362 was still unreactive towards the nucleophile attack as the Weinreb amide synthesis failed and compound 363 was isolated as a major product. Additionally, direct

Grignard reaction using vinylMgBr on 362 failed, giving only degradation.

At this stage, we had to think about other methods to construct the vinylketone

351 other than manipulating the ester 328. After searching the literature, we turned our attention to the Meyer-Shuster rearrangement (Scheme 65). In general, the Meyer-

Shuster rearrangement occurs on secondary and tertiary propargylic alcohols (364 to 365) and the rearrangement of primary propargylic alcohol is rare. Reported examples60 on primary propargylic alcohol involve Au(I) catalysis (366 to 367) or Hg(OTf)2 catalysed rearrangement of its acetate (369 to 367) but both methods have limited substrate scope.

Our plan was to prepare imine 374 first, followed by nitro-Mannich reaction to

375 (Scheme 66). The TBS group would then be removed to give the propargylic alcohol, which will be the precursor for the Au(I) catalysed or Hg(OTf)2 catalysed Meyer-Schuster rearrangement to give vinylketone 377.

Chapter 3 Studies on the total synthesis of schizozygine

Scheme 65. Meyer-Shuster rearrangement

Scheme 66. Revised plan towards -nitro vinylketone intermediate

Chapter 3 Studies on the total synthesis of schizozygine

The synthesis of 374 was straightforward by following the reported procedure from 2-butyn-1,4-diol 378 to the aldehyde 373 (Scheme 67).61 The aldehyde 373 was then treated with p-anisidine and anhydrous Na2SO4 to give the imine 374, which was isolated in only 25% yield and was used immediately after chromatographic urification.

Scheme 67. Preparation of imine 374.

Scheme 68. Synthesis of piperidine 376 with propargylic alcohol

The nitro-Mannich reaction was carried out with 329 and the nitro amine 375 was obtained with 7:1 dr and over 80% yield (Scheme 68). Reductive cyclisation of 375 under

Chapter 3 Studies on the total synthesis of schizozygine

the optimised BF3·OEt2 and Et3SiH conditions gave the piperidine 376 with concomitant removal of the TBS. However, purification of 376 was difficult because of a side product with similar polarity. By analysing the 1H NMR spectra of this mixture, the side product was elucidated as 377 according to the observed two singlet peaks at 6.65 and 6.23 ppm which were assigned as H-1 and H-2 (Scheme 68). A plausible mechanism for the formation of 377 involves the trapping of the cyclised iminium ion 378 by a Pictet-

Spengler type reaction to give 377. This double cyclisation was not observed in 330 and the ease of this process could be substrate dependent.

Because of the difficulties in isolating 376 and 377, this mixture was subjected to acylation condition (acetic anhydride/pyridine) and 379 was isolated in 53% yield.

Unfortunately, the rearrangement catalysed by Hg(OTf)2 did not occur and increasing the catalytic loading of Hg(OTf)2 to one equivalent caused the degradation of the starting material (Scheme 69).

Scheme 69. Attempted Meyer-Shuster rearrangement

Chapter 3 Studies on the total synthesis of schizozygine

3.3.4 Detour from a vinyl ketone: ring E synthesis via Dieckmann cyclisation

Due to the difficulties in forming the desired vinylketone, an alternative route was proposed to synthesise the ring E through Dieckmann cyclisation (Scheme 70).

Scheme 70. Alternative disconnection for ring E synthesis

As shown in Scheme 70, installing another ester group  to the carbonyl group

(381) would make the five membered-ring E accessible through a Dieckmann cycclisation from 382. The synthesis of 382 would involve a Michael addition of 328 with methyl acrylate. Based on this idea, piperidine 328 was treated with methylacrylate in the presence of Triton-B as catalyst to give 382 in 35-40% yield (Scheme 71). However, an attempted Dieckmann cyclisation with KHMDS failed to deliver the cyclised product

381 and only epimerization of 328 was observed. In hindsight it was not surprising that the enolate, generated by treatment with KHMDS, underwent a fragmentation through a retro-Michael addition pathway to give 328 as a mixture of diastereomers after protonation.

Chapter 3 Studies on the total synthesis of schizozygine

Scheme 71. Attempted ring E synthesis via Michael addition/Dieckmann cyclisation sequence

An obvious measure to avoid the retro-Michael addition type fragmentation was to reduce the nitro group to the free amine 384. It turned out to be unsuccessful and the nitro group in 382 survived the Zn/HCl reduction. No free amine 384 or cyclised lactam

385 was observed from the crude reaction mixtures.

Scheme 72. Unsuccessful reduction on nitro group

Besides the Michael addition of 328, a Tsuji-Trost allylation was also tested.

Although the mass spectroscopy experiment detected the existence of the product. The reaction mixture was complicated with several other compounds. We suggested that the piperidine nitrogen might have a negative effect as the ‘basic nitrogen’ issue is a common problem in Lewis acid and transition metal catalysis chemistry.

Chapter 3 Studies on the total synthesis of schizozygine

Scheme 73. Tsuji-Trost allylation on piperidine 328

3.4 Chapter conclusion

At this stage we decided to re-analyse the whole synthetic plan and look for alternative modifications to continue the synthesis. The studies that were attempted showed that intermediate 328 had some drawbacks towards manipulations (Scheme 74). The sterically hindered para-methoxyphenyl (PMP) group prohibited most nucleophilic addition reactons to the ester group. The only successful ester reduction yielded a primary alcohol, which was unable to be re-oxidised to the aldehyde. Elimination of the nitro group was also observed when the molecule was treated with bases (TMG, DBU and

TBAF) and this limited functionalisation on C-2. Moreover, some Lewis acid (Meyer-

Shuster rearrangement) or metal catalysed (Tsuji-Trost allylation) transformations also failed, most likely due to the basicity of the piperidine nitrogen.

Chapter 3 Studies on the total synthesis of schizozygine

Scheme 74. Summary of attempts on ring E synthesis.

Chapter 4. New attempt towards schizozygine: total synthesis of (+)-vallesamidine

4.1 Revised retrosynthesis

Based on unsuccessful work discussed in Chapter 3, we concluded that the ring C intermediate 328 needed to be structurally modified to allow a different synthetic approach. The ester group should be removed to prevent undesired -elimination. The piperidine nitrogen would be ‘protected’ as a lactam and compound 386 was finally decided as a new ring C intermediate (Scheme 75).

Scheme 75. Structural modification of 328

At this stage, the University’s licence to purchase piperonal 305 (catalogue 1 controlled chemicals that are used to manufacture the controlled substances/illicit drugs) had expired and thus we changed the prior targets from schizozygine (6) to vallesamidine

(102) and strempeliopine (103), the latter only differing from schizozygine (6) by the deletion of the dioxolane ring substituent (Scheme 19). The synthesis of vallesamidine

(102) could provide a ‘model’ study towards schizozygine and prove that the synthetic route would be divergent to both type of molecules through C-14,15 olefin construction.

Changing the ring C intermediate to lactam 386, required the development of new synthetic strategy to build the ring E. As no functional group would be installed on C-21

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

(see 386, Scheme 75) then the carbon-carbon bond formation on C-21 would only be achieved through a C-H functionalisation pathway (Scheme 76). A PhD project in the group done by O. Ware 62 developed a [1,5]-hydride transfer/nitro-Mannich cascade reaction to synthesise tetrahydroquinoline structures (Scheme 76) and a similar methodology could provide a potential solution (see Section 4.2 for detailed review on

[1,n]-hydride transfer cyclisation).

Scheme 76. [1,5]-hydride transfer/nitro-Mannich type cyclisation

Therefore, a revised synthetic route was proposed (Scheme 770. The 14,15- dehydro precursor 387 could be synthesised by a ring closing metathesis on 388 and C-

20 substituents could be manipulated from malonate 389. For ring E synthesis, we proposed a [1,4]-hydride transfer (HT)/Mannich type cyclisation, an intramolecular redox, C-H activation process on piperidine ring C (see 390). An intramolecular C-N coupling reaction would be applied to synthesise ring B (391) and the precursor 392 bearing a quaternary carbon (C2) could be prepared through a Tsuji-Trost allylation of nitro lactam 393 (refer to 386 in Scheme 75), a known skeleton that could be assembled through a nitro-Mannich/lactamisation cascade from nitro malonate 394.

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 77. Revised retrosynthesis on schizozygine

4.2 Feasibility and literature precedents on key transformations

4.2.1 [1,n]-hydride transfer/cyclisation

The [1,n]-hydride transfer (HT)/cyclisation, also regarded as internal redox neutral C-H functionalisation, is an efficient method of ring formation and has received much attention from synthetic chemists.63

A HT/cyclisation requires the substrate to possess both a hydride donor and hydride acceptor. The transferable hydrides (hydride donor) are normally at benzylic or anomeric position (e.g. ethers, thioethers and amines) as the hyperconjugation effect weakens these C-H bonds and raises the HOMO energy. The hydride acceptors are

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

usually good electrophiles, such as aldehydes/ketones, electro-withdrawn conjugate alkene systems (Michael acceptors), alkynes or carbocations. Another criterion for successful HT/cyclisation processes is that the good overlap between the hydride s orbital and the * orbital of the acceptor moiety must be accessible.

Therefore, under certain activation (thermal, organo-catalysis or metal catalysis), the hydride is intramolecularly transferred to the acceptor and forms a zwitterion intermediate containing a nucleophile/electrophile pair. The ring closure occurs via carbon-carbon bond formation to complete the whole process. For brevity, only nitrogen containing HT/cyclisation reactions will be discussed, and other substrates will be summarised with selected examples.63b

Scheme 78. General scheme for [1,n]-hydride transfer/cyclisation

Before widespread systematic studies on HT/cyclisation processes in the literature, the unexpected intramolecular, ‘through space’ hydride transfer had been observed during some complex molecules’ synthesis. A typical example was reported in 1976 by a

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

research team in Merck during the synthesis of novel structures for drug discovery.64 An unusual structural rearrangement was observed on compound 396 by heating in methanol at 150 oC and a highly caged ring skeleton 398 with a new carbocycle formed was isolated in 81% yield (Scheme 79). A Deuterium-labeling confirmed the HT/cyclisation mechanism that the polar protic solvent could facilitate the transformation.

Scheme 79. Unexpected HT/cyclisation

Another interesting and relavent hydride transfer in natural product synthesis was contributed by Heathcock from the synthesis of daphniphyllum alkaloids.65 After an intramolecular Diels-Alder reaction (IMDA) in acetic acid at room temperature, dihydropyridinium salt 399 was transformed into intermediate 400 (Scheme 80). Raising the temperature to 80 oC smoothly led to the cyclised intermediate 401 and the resulting tertiary carbon cation was captured by an intramolecular [1,5]-hydride transfer to give iminium ion 402. A simple hydrolysis gave the secondary amine 403.

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 80. Hydride transfer in the total synthesis of daphniphyllum alkaloids

The early studies on [1,n]-HT/cyclisation using amines as hydride donors could be traced back to the 1980s when Reinhoudt and co-workers reported thermal [1,5]-

HT/cyclisation on imine compound 404 to aminal skeleton 405 (Scheme 81).66 Few studies in this field were investigated until recent decades, coincidently along with the emergence of C-H activation as a defined research topic.

Scheme 81. [1,5]-HT/cyclisation to cyclic aminal compounds.

Generally, all amine [1,n]-HT/cyclisations can be classified into endo- and exo- processes, based upon the cyclisation type.

Endo-type HT/cyclisation: For most widely studied anisidine substrates (see 406,

Scheme 82), [1,5]-hydride transfer leads to a zwitterion intermediate 407. Subsequent 92

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

cyclisation proceeds through a 6-endo-trig transition state and therefore these transformations belong to the endo-type. Selected examples with various hydride acceptors were shown in Scheme 82. Different Lewis acids were also screened and worked quite well in most of the cases on this substrate type.67 It is worth noting that the example from 415 created a divergent skeleton construction method by using different catalyst. Normal Lewis acid Sc(OTf)3 catalysed the standard [1,5]-HT/cyclisation to 416.

Changing the catalyst to alkyne-philic Gold (I) (i.e. IPrAuOTf) catalyst first led to the formation of furan intermediate 418. The highly electrophilic carbon cation induced the

[1,5]-HT/cyclisation to occur, the furan being nucleophile in the process to give product

417.68

In 2009, a research team from Pfizer, led by Ruble and Hurd, reported a synthesis of PNU-286607, an antibacterial agent, using the methodology of thermal [1,5]-

HT/cyclisation (Scheme 83).69 To the best of our knowledge, this is the only example of the application of a [1,n]-HT/cyclisation in drugs/natural products synthesis.70

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 82. Endo-type [1,5]-HT/cyclisation and selected examples

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 83. Synthesis of antibacterial agent PNU-286607.

Exo-type HT/cyclisation: The exo-type HT/cyclisation requires a different type of substrate like 421 (Scheme 84). After the hydride transfer, the resulting zwitterionic intermediate 422 completes the ring closure through an exo-trig pathway to give 423.

Compared with endo-type HT/cyclisation, the exo-type has received much less attention and selected examples are summarised in Scheme 85. It is worth noting that the feasibility of using amines as hydride donors in HT/cyclisation lies in the basicity of nitrogen atoms.

In anisidine type substrates, the basicity of amines is dramatically decreased by delocalisation with aromatic systems. For non-anisidine type substrates, amines can sequester the Lewis aid and inhibit the catalytic process and this issue must be considered when designing the substrate.

Scheme 84. Exo-type [1,n]-HT/cyclisation

In 2005, Sames et al reported the [1,5]-hydride shift/cyclisation of saturated heterocycles (pyran, furan and pyrrolidine) (Scheme 85a).71 The -branched substrates generated spiro-compounds (424 to 425) and the -branched substrates delivered fused 95

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

bicyclic products (426 to 427). For the pyrrolidine 426, the amine was protected using an electron-withdrawing protecting group to minimise any interaction with the Lewis acid. Examples of [1,4]-HT/cyclisations are even less common, with the most impressive study being reported by Akiyama et al. It showed that for non-anisidine type substrates bulky amines normally gave excellent yields while the less hindered amine, especially cyclic amines (piperidine), provided much less yields or even no reaction.

Scheme 85. Selected examples of exo-type [1,n]-HT/cyclisation

For our retrosynthesis of vallesamidine (102) (Scheme 77), the [1,4]-

HT/cyclisation was proposed on 2-substituted piperidine 390 with malonate as the

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

hydride acceptor. This type of substrate (see simplified model 433, Scheme 86) has not been studied. The work by Sames71 and Saá72 disclosed the [1,6] and [1,5]-

HT/cyclisation on 2-substituted piperidines 430 with alkynes as hydride acceptors and these impressive results indicated conformational possibilities when applying [1,4]-

HT/cyclisation (Scheme 86).

Scheme 86. [1,n]-HT/cyclisation of 2-branched piperidine substrate

4.2.2 Tsuji-Trost allylation and nitro-Mannich/lactamisation reaction

Tsuji-Trost allylation: Compared with other methods for the alkylation of nitroalkanes

(excluding addition to C=X bonds), the Tsuji-Trost allylation is the most reliable and robust (Scheme 87) with extensive methodological studies on regio-, diastereo- and enantioselective transformations.73

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 87. Mechanism of Tsuji-Trost allylation of nitroalkanes

One example using this transformation in the total synthesis studies is highlighted here as precedent for our synthesis. Shinada and co-workers reported studies towards the total synthesis of tetrdotoxin, by synthesising the bicyclic intermediate 460 (Scheme88).74

The Diels-Alder reaction between furan 451 and nitroalkene 452 afforded the bicyclic skeleton 454 and 455 with a ratio of 5:1. After reduction of the ketone and ester functions, followed by protection of the primary alcohol, the major compound 456 was subjected to

Tsuji-Trost allyation conditions and the allylated product 457 was obtained in good yield.

In addition, the nitro group was also reduced to an amine under standard conditions.

Nitro-Mannich/lactamisation reaction: Our synthesis of piperidinone compound through a nitro-Mannich/lactamisation reaction was discussed in Chapter 1 (see Scheme

8, 11, 12). Literature precedent suggests that the nitro-Mannich/lactamisation from 461 to 462 usually gives a mixture of diastereomers (Scheme 89).7 This issue is not a problem in our strategy because both isomers would generate the identical nitronate intermediate

463 under basic conditions required for subsequent Tsuji-Trost allylation.

Diastereoselectivity would then be controlled by the reactive conformation of the substrate 463.

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 88. Precedent of Tsuji-Trost allylation on nitroalkane

Scheme 89. Nitro-Mannich/lactamisation to give the piperidinone structure

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

4.3 Synthetic studies

4.3.1 Nitro-Mannich/lactamisation and Tsuji-Trost allylation to functionalised piperidinone ring C

The synthetic studies were carried out in racemic form first. The preparation of nitroester

466 commenced with the HWE olefination of 2-bromobenzaldehyde 464 followed by a

Michael addition with nitromethane.75 Meanwhile, the compound 323 was also used in the investigation as we still planned to attempt the indoline synthesis via C-H activation

(see Scheme 51).

Scheme 90. Preparation of nitroester 466

The nitro-Mannich/lactamisation was first conducted on 323 using benzylamine

76 and aqueous formaldehyde solution in warm iPrOH/H2O solvents (Scheme 91).

Unfortunately, only 7% of the desired product 467 was isolated. Changing aqueous formaldehyde to paraformaldehyde significantly improved the yield to give full conversion and a clean mixture of two diastereomers (dr 75:25). Without further purification, the crude mixture 467 was directly subjected to well precedented Tsuji-Trost

100

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

o 74 allylation conditions (allyl acetate, DBU, PdCl2(PPh3)2, DMF, 80 C) to give the allylated product 468 as a single diastereomer in 55% isolated yield over two steps.

Scheme 91. Nitro-Mannich/lactamisation and Tsuji-Trost allylation of dioxolane substrate 323

Repeating the above nitro-Mannich/lactamisation sequence on 466 with benzylamine afforded the piperidinone 469 in 62% isolated yield with a diastereomeric ratio of 65:35 (Scheme 92). As the Tsuji-Trost allylation of 467 was fast, it suggested that such harsh conditions were probably not necessary. Therefore, instead of using the

74 condition adopted from Shinada’s work, Pd(PPh3)4 was used and the reaction was performed at room temperature to give the product 471 in 30 minutes in good 81% yield.

Lowering the temperature to 0 oC did not affect the reaction time but increased the yield to 91%. For the ease of future N-protective group removal, the p-methoxybenzylamine was used to give lactam 470 in 69% yield and the allylation was just as good, proceeding in 92% yield (472). The success of this transformation supported our previous hypothesis that the piperidine nitrogen had a negative effect on palladium catalysed reactions and

‘piperidine to lactam’ strategy made the catalytic allylation accessible.

101

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 92. Optimised nitro-Mannich/lactamisation, Tsuji Trost allylation sequence

4.3.2 Attempt on direct C-N bond formation to indoline structure by C-H activation.

Before continuing our synthesis with 472, we used the remaining material of 468 to test the previous idea of direct intramolecular C-N bond formation to construct the indoline structure. Compound 468 was first treated with zinc dust and hydrochloric acid to reduce the nitro group to the free amine 473 quantitatively without further purification (Scheme

93). However, the protection of the amino group to the necessary sulfonate (474 or 475) was problematic and the starting material was recovered in most cases. The failure to install the directing group disabled the C-N coupling through C-H functionalisation. At this stage, we were curious to how hindered the amino group was and the attachment of several other protective groups was investigated. The Boc protection did not afford the desired compound, instead giving a compound proposed as the isocyanate 476. The small and highly reactive trifluoroacetic anhydride (TFAA) was used and the TFA protected

477 was isolated in 60% yield.

102

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 93. Unsuccessful installation of directing group on tertiary amine 470

4.3.3 Towards vallesamidine: C-N coupling reaction to ring B

Continuing the synthesis with 472, nitro group reduction with zinc dust and hydrochloric acid in warm isopropanol (50 oC) caused the removal of the bromine atom. Changing the reaction conditions to zinc dust and hydrochloric acid in cooled ethanol/ethlyl acetate (0 oC) gave the desired free amine 478 in 100% yield without purification (Scheme 94).

Scheme 94. Nitro group reduction and intramolecular C-N coupling reaction

103

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

To establish the C-N coupling reaction to 479, the conditions we had used

77 previously for similar coupling (of Pd(PPh3)4/t-BuONa) were applied and an average yield of 65% was obtained when brand new palladium catalyst was used. However, the performance of this palladium catalysed C-N bond formation was highly dependent on the quality of the palladium catalyst and the reaction could not be repeated with the same batch of catalyst after several months even though the catalyst still performed fine in the

Tsuji-Trost allylation. Moreover, the large molecular weight of the catalyst meant that a large mass was required upon scaled-up. Therefore, we decided to investigate the copper catalysed Ullman type C-N coupling conditions (Table 7).

Table 7. Conditions for intramolecular C-N coupling reaction 478 to 479

entry condition conv.; yield (%)

o 1 Pd(PPh3)4 (5 mol%), t-BuONa (2 eq.), PhMe, 90 C, 16 h 100; 65

o 2 CuI (5 mol%), Cs2CO3 (2 eq.), DMF, 45 C, 16 h <10; n/a

o 3 CuI (5 mol%), Cs2CO3 (2 eq.), DMF, 80 C, 16 h 56; 40

o 4 CuI (10 mol%), Cs2CO3 (2 eq.), DMF, 95 C, 20 h 75; 55

CuI (10 mol%), L-proline (20 mol%), K3PO4 (2 eq.), 5 100; >90 DMSO, 95 oC, 1 h

Normally intramolecular coupling reactions are easier than intermolecular examples due to the closer proximity of the reactive partners. Thus, a ligand free condition of CuI/Cs2CO3 was first subjected to the amine 478 in anhydrous DMF in inert atmosphere. 78 Under literature temperature of 45 oC gave a slow reaction and less than

10% conversion after 16h. Raising the temperature to 80 oC improved the conversion to

56% and the desired indoline product was isolated in 40% yield. Further modifications were carried out by increasing catalyst loading (5.0 mol% to 10.0 mol%) and the reaction 104

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

time (24 h) and finally a better yield of 55% was achieved on gram scale. This slow but clean reaction led us to consider adding ligands to accelerate the process.

Extensive studies on copper catalysed C-N bond coupling reactions promoted by various ligands79 such as N,N ligands, N,O ligands, and O,O ligands (Figure 2) have been reported. From these reported results, we first tested amino acids as ligands, according to the work of Ma,80 to see if an acceleration could be achieved. Surprisingly, exposure of amine 478 to CuI (10 mol%), L-proline (20 mol%) and K3PO4 (2.0 eq.) in DMSO at

95 oC quickly led to completion in one hour and the crude product 479 after simple work- up was judged to be clean enough for the next step.

Figure 2. Reported ligands for Ulmann type C-N coupling reactions

105

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Figure 3. 1H NMR spectra comparison between crude (upper) and purified (lower) indoline 479

4.3.4 Enantioselective synthesis of A/B/C ring intermediate (479)

To establish an enantioselective synthesis, the preparation of enantioenriched nitro ester

480 was required. The -nitro esters 480, or the equivalents of malonate 481, are normally obtained from the Michael addition of either nitromethane to the conjugate esters/malonates (method 1) or malonates to nitroalkenes (method 2) (Scheme 95).

Scheme 95. Two methods to prepare enantioenriched nitroesters/malonates 106

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

The direct asymmetric Michael addition of nitromethane to conjugate esters

(method 1) has not been reported due to the lack of a rigid interaction between the ester group and the chiral catalysts (organocatalyst/Lewis acid/Bronsted acid etc). Conversely conjugated malonates are more feasible for asymmetric activation, due to two point binding to an asymmetric catalyst, and the Michael addition with nitroalkanes has been reported by Hajra,81 Palomo,82 but these methodologies were limited to either moderate enantioselectivity or substrate scope, especially for ortho-substituted analogues (482 to

483, Scheme 94).

Michael addition of malonates to nitroalkenes (method 2) have been more extensively investigated with different organocatalysts and metal centred catalysts (484 to 483). Considering the reaction scale, substrate scope, enantioselectivity and catalysts’ cost, we chose the Evans’ method that uses chiral diamine-nickel (II) complex 487 as the catalyst (Scheme 95).83 The original research screened various substituted nitroalkenes and used a low catalyst loading (1 mol%) when scale-up experiments (20 mmol) were conducted. Moreover, this bench stable catalyst 487 could be synthesised efficiently in three-steps from (R,R)-N,N’-cyclohexane-1,2-diamine tartaric acid salt 485 which was suitable for our desired large scale reactions (Scheme 96).

107

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 96. NiBr2-diamine complex catalysed asymmetric Michael addition

We first tested the Michael addition using tert-butylmalonate, chosen for the subsequent ease of decarboxylation to the nitro ester, which worked well with nitrostyrene in 97% yield and 95% ee (Scheme 97). However, tert-butylmalonate did not add to the 2-bromo nitrostyrene 395 under these conditions, presumably due to the steric hindrance. Ultimately, diethylmalonate was used and gave the desired product 394 in up to 90% yield at 20 to 40 mmol scale. Chiral HPLC analysis indicated an enantiomeric ratio (e.r.) of 95:5. With the nitro malonate 394 in hand, we found that the

108

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

decarboxylation of malonate to ester 492 was problematic and the materials degraded

o upon treating with NaCl in DMSO/H2O at 165 C.

Scheme 97. Preparation of enantioenriched nitromalonate 394

Similar problems have been described in the literature. 84 Encouragingly decarboxylation of compound 493, the product of nitro-Mannich/lactamisation, was found to work well (Scheme 98).7 Therefore, the nitro malonate 394 was treated with paraformaldehyde, 4-methoxy benzylamine in hot ethanol (80 oC) to conduct the nitro

Mannich/lactamisation and after simple work-up the crude product 495 was directly heated with NaCl in DMSO/H2O to give the decarboxylated lactam 470 in 85% overall yield with a mixture of diastereomers at C-2 (85:15) after chromatography. Without separating these two diastereomers, Tsuji-Trost allylation at C-2 to give 472 with low catalyst loading (1.0 mol%) and excellent yield (86%) as a single diastereoisomer. The subsequent nitro group reduction quantitatively yielded the free amine 478 and the intramolecular C-N coupling reaction worked smoothly to give 479 in 84% isolated yield

(90% crude).

109

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 98. Synthesis of enantio-enriched A/B/C ring intermediate 479

4.3.5 Unexpected difficulties in indoline protection

Successful construction of the A/B/C ring system led us to consider the manipulation of the allyl group towards the ring E synthesis. To avoid any undesired rearrangement from the indoline structure, the protection of the indoline nitrogen should be necessary.

However, it proved difficult to install a protecting group. The Boc protection was 110

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

investigated first and no protection was observed under Boc2O/DMAP (cat.)/ Et3N conditions. Precedent on Boc protection of sterically hindered indoline 497 used excess

DMAP in acetonitrile (MeCN) had given 96% yield (Scheme 99).85 These conditions were attempted on 479, but 90% of starting material was recovered. Full deprotonation using sodium hydride (NaH) failed to give any product and potassium bis(trimethylsilyl)

86 amide (KHMDS)/Boc2O conditions (see precedent substrate B, Scheme 98) gave 496 in 60 to 70% yield at small scale, but the yield dropped dramatically upon scale-up.

Scheme 99. Boc protection of indoline 479

Another widely used protecting group in indole alkaloid synthesis is a methylcarbamate. Treatment of 479 with methylchloroformate at room temperature for 24 to 48 hours did not give any product (Scheme 100). Changing the solvent from dichloromethane to 1,2-dichloroethane to enable a higher reaction temperature improved the process, but the reaction did not go to completion after 48 hours. Optimised reaction conditions used neat chloroformate at reflux temperature (65 oC) and gave the best result to give 499 in 88% yield.

111

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 100. Protection of indoline to methylcarbamate

The major drawback of this condition was the usage of large amount of methylchloroformate. Although the excess chloroformate could be recovered by distillation, the highly volatile nature (vapour pressure at 20 oC, 4.8 psi/0.33 atm)v of this reagent may result in safety concerns at large scale. Ultimately, a one-pot, two-step protocol was devised that relied upon treating indoline 479 with highly reactive triphosgene solution and the resulting chloroformamide 500 was stirred in methanol at refluxing temperature to give 499 in 87% yield (Scheme 100). The overall process was much safer, practical and the high yield was also maintained.

v Vapour pressure obtained from Sigma-Aldrich: https://www.sigmaaldrich.com/catalog/product/aldrich/m35304?lang=en®ion=GB visited 2 July 2019

112

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

N 1

Figure 4. X-ray structure of 499

A single crystal of 499 was grown from toluene and X-ray diffraction experiment confirmed the structure and relative stereochemistry. The crystal structure of 499 shows an unexpected conformation in that the PMB group is elucidated towards the concave face of the bicyclic skeleton, presumably due to stabilising - interaction between the phenyl and indoline aromatic systems. In this conformation it can be seen that the indoline nitrogen (N1) is highly hindered and this may explain its inert reactivity.

4.3.6 [1,4]-HT/cyclisation and malonate manipulation

To continue the synthesis the conversion of the lactam to the piperidine now posed an unforeseen difficult chemoselective reduction due to the presence of the methyl carabamate group. To circumvent this the amide functional group was first transformed to the thioamide 501 in 93% yield by reaction with Lawesson’s reagent. Methylation of the thioamide, followed by reduction of the resultant thio-imide using sodium borohydride (NaBH4) gave piperidine (-)-502 in excellent yield (Scheme 101).

113

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 101. Lactam reduction to piperidine 502 and initial oxidative cleavage

The oxidative cleavage of the terminal alkene in 502 under Lemieux–Johnson oxidation condition (cat. OsO4, NaIO4) or standard ozonolysis condition led to degradation, most probably due to the basic piperidine amine (Scheme 101). Trost et al encountered similar problems during their studies on the total synthesis of (-)- perophoramidine (Scheme 102).87 Normal ozonolysis of the allyl group in compound

504 failed and led to degradation and this was blamed upon the oxidative instability of the electron rich and basic amidine moiety. A modified method was then developed that involve initial treatment of 504 with camphorsulfonic acid (CSA) to ‘protect’ the basic nitrogen in situ. Subsequent ozonlolysis proceeded well and the desired aldehyde 505 was subjected directly to a reductive amination to give 506 in good yield.

Scheme 102. Trost’s example on ozonolysis via in situ protonation of basic nitrogen

114

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Inspired by this, ozonolysis of 502 under Trost’s conditions gave the dimethylacetal product 507 in 53% yield due to the presence of acid and methanol

(Scheme 103). This reaction was then slightly modified by using trifluoroacetic acid

(TFA) to in situ protect the piperidine nitrogen and reducing the ozonide with Me2S without adding methanol and the desired aldehyde 503 was afforded in 56% yield. An anhydrous work-up method was then introduced which involved simply adding triethylamine to neutralise the reaction mixture followed by direct chromatography and the aldehyde 503 was isolated in 92% yield.

Scheme 103. Pre-acidified ozonolysis on 502.

The Knoevenagel condensation of aldehyde 503 with dimethylmalonate in the presence of pyridine/acetic acid gave the conjugated malonate 508 in 72% yield.

Successful preparation of 508 provided us the chance to investigate the key transformation of ring E construction. To our delight, exposure of 508 to catalytic

o Yb(OTf)3 in toluene at 100 C built the ring E smoothly in good yield (80%) (Scheme

103). The structure was confirmed by mass spectroscopy and NMR experiments. In the

1H NMR, a new, singlet peak at 5.47 ppm and the disappearance of the original H-21 signals [ 3.41 (d, J = 13.2) and 3.31 (d, J = 13.2)] in 509 indicated that the reaction had occurred at the desired position. In addition, the cyclisation was also confirmed by

115

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

HMBC experiment that showed the proton we had assigned as H-21 [ 5.47 (s)] correlated with C-20 and the carbonyl carbons of the ester groups.

Scheme 104. Preliminary investigation of the [1,4]-HT/cyclisation

Mechanistically we can account for such an efficient transformation. The high reaction temperature allows access to the less stable reactive conformation 508-a. Under the activation of the Lewis acid, the axial hydrogen on C-21, anti-periplanar to the adjacent lone pair electrons on nitrogen, can be transferred to the conjugated malonate giving both iminium and enolate ions (510) (Scheme 105). Enolate attack of the iminium carbon via a Mannich type reaction completes the five-membered ring E synthesis.

Instead of using malonate as the hydride acceptor, the nitroalkene 511 was also prepared, but the [1,4]-HT/cyclisation failed to deliver the cyclised product 512, presumably due to insufficient activation. Concurrent studies in our group at the time confirmed that this transformation is much more difficult and very substrate dependent.62

116

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 105. Proposed mechanism and conformation analysis on [1,4]-HT/cyclisation on 508

With the intermediate 509 in hand, we then started investigations into manipulations of the malonate group. Krapcho decarboxylation of 509 gave a complex mixture and deprotection of the methyl carbamate was also observed (Scheme 106).

Saponification using KOH in ethanol proceeded slowly with degradation. Ultimately, only global reduction of both the malonate esters and the methyl carbamate using LiAlH4 in refluxed THF worked and yielded diol 516 quantitatively.

Scheme 106. Attempts at malonate manipulation

117

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

The fused ring A/B/C/E-ring system made the two hydroxyl groups sterically different and gave an opportunity for a stereoselective mono-protection. Treatment of

516 with TBSCl afforded the mono-protected 517 (Scheme 107). The selectivity was arbitrarily assigned as in 517, because the two methylene groups could not be distinguished in 1H NMR spectra. Subsequent oxidation on the remaining alcohol was not straightforward. The Dess-Martin oxidation, IBX oxidation (in DMSO) and PCC oxidation led to rapid degradation, presumably due to the piperidine nitrogen. We speculated that a heterogenous oxidation may slow down the degradation so the heterogenous IBX oxidation using EtOAc as solvent was then attempted.88 To our delight, the oxidation in EtOAc at refluxing temperature proceeded smoothly, but gave a mixture of C-20 epimers (10:1). The oxidation reaction eliminated one of the two methylene groups, so a ROESY experiment was able to show that the major component had a correlation between C21-H [ 3.36 (1H, s)] and the remaining methylene group [

3.50 (1H, d, J = 9.7) and 3.28 (1H, d, J = 9.7)]. This confirmed that the major isomer was the desired product 518.

Scheme 107. Mono protection and oxidation of diol 516

118

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Figure 5. 1H NMR of aldehyde 518

A possible mechanism was proposed for the formation of the epimers from a single diastereomer 517 (Scheme 108). After the IBX oxidation, the product 518 could undergo a retro-Mannich/Mannich reaction and this process could cause the epimerisation on C-20. As the undesired epi-518 suffered a large steric repulsion between the silyl group and the PMB group, compound 518 still predominated. Alternatively, before the

IBX oxidation, the silyl group exchange could occur first and the resulting epi-517 with the less hindered hydroxyl group was quickly oxidised to epi-518. Potentially both pathways could be occurring during the oxidation and equilibration could ultimately lead to a consistent ratio (10:1 for 518).

119

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 108. Epimerisation in IBX oxidation.

The next transformation was the olefination of the aldehyde in 518. This hindered aldehyde was inert to Wittig reaction and no reaction was observed even with a large excess of phosphonium ylide. Olefination with the Petasis reagent, dimethyltitanocene

(Cp2TiMe2), was considered because this olefination reagent can react with a wide range of carbonyl groups (e.g. aldehydes, ketones, esters, carbonates etc.) to give terminal alkenes, even in sterically hindered cases.89 The Petasis reagent was prepared according to the literature by addition of methylmagnesium chloride to dichlorotitanocene and stored in the dark as a solution in toluene (15~17% w/w) (Scheme 109). With freshly prepared Petasis reagent in hand, olefination on aldehyde 518 was conducted and the desired alkene 520 was isolated as a mixture of C-20 epimers in 75% yield, based on recovered starting material and the ratio of epimers (10:1) was unchanged.

120

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 109. Petasis olefination of hindered aldehyde 518

The next step required the removal of the PMB group, followed by N-allylation to provide the precursor for ring closing metathesis. Oxidative deprotection conditions

(CAN, MeCN-H2O) on 520 caused decomposition of the molecule and attempts at an acylation strategy (ClCO2CH(Cl)CH3) did not perform well, with most of the starting material recovered. We then looked at N-PMB deprotection of malonate 509 to 521 or functional group exchange to 522 or 523, but this also proved problematic (Scheme 110) with the same observations as for 520. We postulated that the steric hindrance around the piperidine nitrogen atom was exacerbated by formation of ring E and we therefore decided to attempt the protecting group interchange/deprotection on the piperidine nitrogen prior to the ring E synthesis.

Scheme 110. Attempted PMB deprotection

121

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Direct deprotection of 502 was still unsuccessful so PMB exchange by N- acylation was attempted. Among different acylating reagents, allylchloroformate was used first as decarboxylative allylation could be carried out later to give the N-allylated product. As a model reaction, compound 502 was treated with allylchloroformate in DCE

(1,2-dichloroethane) at 80 oC and gave the desired product 525 in 76% yield, which provided an impressive result for this strategy. Due to potential regio-selectivity problems towards oxidative cleavage between the two allyl groups in 525, we decided that the aldehyde compound 503 would be a better entry for PMB group exchange. To our surprise, PMB group exchange on 503 was much slower and never went to completion under the same conditions. Increasing the amount of chloroformate to solvent level (50 eq.) raised the yield to 68%. Repeating the reaction under neat conditions did not improve the yield and finally the average yield on gram scale was consistent at 55%.

The difference in reactivity between olefin 502 and aldehyde 503 suggested that the aldehyde group may inhibit the PMB deprotection by an electronic effect.

Considering the conformers 502 and 503 (Scheme 111), in compound 503, the axial aldehyde chain could participate in a stronger non-bonding interaction (hyper conjugation) between the lone pair of electrons on the piperidine nitrogen and the aldehyde  orbital. This may account for the reduced reactivity of 503 in this reaction compared to 502, in which the hyper conjugative interaction would energetically be much less, or absent, due to the higher * energy of the less polar alkene.

122

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 111. Successful PMB deprotection before ring E synthesis

Taking this idea to the extreme, the piperidine nitrogen could attack the aldehyde to form an aminol 527 (Scheme 112), which would also serve to inhibit the PMB exchange, but we did not observe this intermediate or any side products derived from it.

Scheme 112. Predicted side reaction pathway (not observed)

The detrimental hyper conjugation effect was negated by formation of the protected primary alcohol 530 by ozonolysis of olefin 502, followed by reduction with

NaBH4 and silyl protection of the alcohol 529 directly after simple work-up in 80% yield over 2 steps (Scheme 113). The PMB group exchange on 530 was significantly improved 123

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

with a shorter reaction time (0.5 h) and excellent yield (92%). This result supported the stereoelectronic effect (n-* interaction) that impeded the reactivity of 503. Swern oxidation of 531 gave a mixture of the desired aldehyde 526 and silyl enol ether 529

(3.5:1). The silyl enol ether could be easily hydrolysed to the corresponding aldehyde by treatment with 2M HCl to give 526 in high yield (90% in total). This redox sequence

(502 to 531 to 526) required two more steps than the conversion of 502 to 526 but was more efficient in both time and overall yield (Scheme 113).

Scheme 113. Revised route to efficient preparation of 526.

Subsequent Pd(0) catalysed decarboxylative N-allylation of 526 smoothly afforded allylated product 533 in 84% yield (Scheme 114). Instead of using piperidine/AcOH conditions that we had previously used (Scheme 104), aldehyde 533 was converted to the conjugated malonate 534 in excellent 96% yield in the presence of

L-proline. The [1,4]-hydride transfer/cyclisation proceeded smoothly as before to give

535 with 70% yield on gram scale. Global reduction using LiAlH4 gave the

124

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

corresponding diol 536 in 90% yield and the subsequent mono protection afforded TBS protected product 537 in 64% yield.

Scheme 114. Preparation of N-allyl A/B/C/E ring intermediate

125

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

It is worth noting that a more concise route to the proposed metathesis precursor

388 via bis-oxidation of diol 536 followed by Wittig olefination according to Feldman’s strategy in meloscine (285) (Scheme 43) was attempted. However, the oxidation step

(Dess-Martin and IBX) encountered rapid degradation and gave complex mixtures.

Like the transformation of 517 to 518 (Scheme 107), heterogeneous oxidation of

537 using IBX also gave a mixture of C-20 epimers (539), but with a worse ratio of 5:1

(Scheme 115). Increasing the concentration of the reaction (0.06 M to 0.12 M, based on alcohol) led to a ratio of 3:1. Compared to compound 517, the smaller size of the allyl group may account for the higher thermodynamic ratio of epimers due to a reduction in steric strain of the major isomer. The isomeric mixture was then subjected to Petasis olefination and this time it was found that direct chromatography after concentration was problematic and the excess titanium residues could not be separated completely from the products. Multiple chromatographic purification did provide a better purity but with a significant reduction of yield (34 to 60% yield depending on scale and repetition of chromatography). Luckily, Payack et al had reported an example of a Petasis olefination on kilogram scale90 and it was mentioned that the titanium residues could be removed by adding hot aqueous methanol with several hours stirring which converted the soluble titanium residues into an insoluble precipitate. Filtration of these solids would remove most of the titanium byproducts.

Scheme 115. IBX oxidation/Petasis Olefination sequence

126

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Following this protocol, after the disappearance of starting aldehyde 538, aqueous

o methanol (9:1, MeOH:H2O) was added and stirred at 60 C for 2-3 hours. After filtration of the precipitates, the solution was concentrated and then passed through a short pad of silica gel to give the crude olefin product 539 which was judged to be pure enough to go on to next step. The crude yield remained at 55%, but the reaction now amenable to larger scale.

Figure 6. 1H NMR of crude 539 after aq. MeOH treatment (upper) and chromatographic

purification (lower)

4.3.7 The Ring closing metathesis to ring D and the completion of total synthesis of (+)-vallesamidine

The ring closing metathesis of the 539 and epi-539 mixture catalysed by Hoveyda-Grubbs

2nd generation catalyst gave both cyclised epimers (540 and epi-540) that could be

127

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

separated (Scheme 116).vi The stereochemistry of the major product 540 was confirmed by a NOESY experiment.

Scheme 116. Ring closing metathesis of 539 and epi-539

The driving force of both the Petasis olefination and olefin metathesis relies upon the conversion a metal alkylidene (R2Ti=CH2 or LnRu=CHR) to a more stable metal oxo complex upon reaction with a carbonyl group or a carbon-carbon double bond. We speculated that the titanium oxo by-product species may be inert towards the olefin metathesis reagent. Also, a large change in polarity on silica gel between 539 and 540 indicated that the cyclised product could potentially be much more easily separated from excess Petasis reagent and titanium oxo by-products. Therefore, a one-pot protocol of

Petasis olefination/ring closing metathesis could be feasible. To explore this aldehyde

538 was treated with dimethyltitanocene in hot toluene in the dark. Once the aldehyde had been consumed (TLC) the reaction mixture was diluted with toluene to ~ 0.02 M and

Hoveyda-Grubbs 2nd generation catalyst added to conduct the ring closing metathesis.

After quenching with aq. methanol and silica gel chromatography, the cyclised product

vi The epi-540 cannot be obtained purely as the product was always mixed with Ru residues.

128

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

540 was isolated as a single isomer in 50% yield from 537 (Scheme 117). This telescoped procedure avoided the lengthy purification protocol after the Petasis olefination step and was more efficient in time and overall yield.

Scheme 117. One-pot Petasis olefination/ring closing metathesis

At this stage, the key goal of this synthetic route involving synthesis of the C14,15 olefin had been achieved. The remaining manipulations to (+)-vallesamidine (102) were straightforward (Scheme 118). The silyl group removal to give alcohol 541 was high yielding but required heating to 65 oC. Oxidation to the aldehyde 542 with IBX proceeded cleanly. Wittig reaction on aldehyde 542 was unsuccessful but use of the Petasis reagent once again furnished the bis-alkene 387, albeit in moderate 65% yield. Finally, hydrogenation in the presence of Pd/C reduced both alkenes in 387 to the target molecule

(+)-vallesamidine (102). The characterisation data and optical data was consistent with the literature.40, 43

129

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

Scheme 118. Completion of the total synthesis of (+)-vallesamidine (102)

4.4 Chapter conclusion

An enantioselective total synthesis of (+)-vallesamidine was completed. This synthesis made full use of the transformations of nitroalkanes, including Michael addition, nitro-

Mannich, Tsuji-Trost allylation and nitro group reduction/C-N coupling reaction, to quickly assemble the A/B/C ring skeleton. Then a [1,4]-hydride transfer/Mannich type cyclisation was applied to construct the five membered ring E. Although the

[HT]/cyclisation has been widely investigated, it is the first time that this strategy has been used in complex alkaloid synthesis. Finally, a tandem Petasis olefination/ring closing metathesis was developed to synthesise ring D. The C-14,15 double bond construction potentially links the vallesamidine/strempeliopine alkaloids and schizozygine alkaloid syntheses.

The late stage manipulation involved several redox processes while the bis- oxidation/olefination of diol 536 was unsuccessful, the step-wise approach to generate intermediate 541 provides an opportunity for further structural modifications and

130

Chapter 4 New attempt towards schizozygine: total synthesis of (+)-vallesamidine

functional group manipulations towards other members of the alkaloid family and diverse natural product like analogues for biological evaluation in the future.

131

Chapter 5 Studies towards the total synthesis of (+)- strempeliopine

5.1 Retrosynthesis of lactam ring F

Compared to vallesamidine (102), strempeliopine has an additional lactam ring F and the retrosynthesis needed only minor alterations. The fully saturated (+)-strempeliopine

(103) would come from the C-14,15 dehydro precursor 543 (Scheme 119). The lactam ring F would be synthesised by halogen-lithium exchange/lactamisation sequence of 544.

Ring-closing metathesis would achieve ring D and the halogen atom could be installed by functional group manipulation of the hydroxyl group of 545, which could be prepared from diol 546. Compound 546 is the reduced product of the common intermediate 535 that was used in the synthesis of vallesamidine.

Scheme 119. Retrosynthesis of (+)-strempeliopine to common intermediate 535

132

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

5.2 Feasibility and literature precedents

5.2.1 Chemoselective reduction of malonate

Compared with the reduction of 535 to 536, the synthesis of 546 requires a chemoselective reduction of the malonate moiety in the presence of the methyl carbamate

(Scheme 120). There is a reactivity difference between the ester and carbamate functionality, and we anticipated that a chemoselective reduction could be achieved via temperature and time control. Although there is no reported example on selective malonate reduction in the presence of a methyl carbamate, the selective reduction of an ester in the presence of a methyl carbamate is known (Scheme 121)91 and it provided a guideline for condition screening.

Scheme 120. Planned chemoselective reduction on malonate

133

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

Scheme 121. Chemoselective reduction of ester

5.2.2 Lithium-halogen exchange/lactamisation

The proposed formation of ring E, lithium-halogen exchange followed by intramolecular addition/substitution with the carbamate to form the lactam ring, is a known process and has been reported mainly for five-membered lactam ring synthesis. Treatment of bromide

552 with t-BuLi led to the lactam 553 formation (Scheme 122).92a A more complex example was reported for 555 where lithium-halogen exchange/lactamisation proceeded smoothly to give the highly caged skeleton 556.92b Although the lithum-halogen exchange/lactamisation has not been reported on six-membered lactam synthesis, we thought that this synthetic strategy should be feasible on 544.

134

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

Scheme 122. Lithium-halogenexchange/lactamisation

5.3 Synthetic studies

5.3.1 Optimisations on starting material preparation

In order to carried out the studies towards (+)-strempeliopine, more intermediate 535 was required and this gave us an opportunity for the optimisation of some of the early steps in the synthesis. Starting from indoline intermediate 499, oxidative cleavage of the alkene was performed first before lactam reduction. Following the protocol developed by

93 Niclaou et al, compound 499 was treated with catalytic OsO4 in the presence of NMO and 2,6-lutidine to perform the Upjohn dihydroxylation and the resulting diol was directly charged with PhI(OAc)2 in one pot to set up the oxidative cleavage to corresponding aldehyde 557 in 86% yield (Scheme 123).

Subsequently, aldehyde 557 was reduced to the alcohol, followed by silyl protection to 558 in 93% yield (Scheme 123, Route A). In our previous total synthesis of

(+)-vallesamidine, preparation of the piperidine from lactam used a two-step protocol that involved transforming the lactam to a thiolactam, followed by a MeI methylation/NaBH4 135

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

reduction sequence. On large-scale preparation this required a large amount of MeI as methylating agent and solvent, and we thought that this could be avoided by investigating the lactam reduction in a catalytic fashion.

Metal catalysed amide reduction has been extensively studied with various metal and silane combinations, including Rh, Ir, Ru, Fe, Zn, Cu, Pt, Mo and Ni.94 Considering the catalysts we had available in the lab, the NiCl2(dme) catalysed amide reduction, developed by Garg et al,94k was investigated first. Treatment of lactam 558 with

NiCl2(dme) (10 mol%) and PhSiH3 (2 eq.) in toluene at elevated temperature gave the desired piperidine 530 in excellent yield (75 to 90%) (Scheme 123). Following the previously developed route, compound 530 was then transformed to aldehyde 526 in two straightforward steps and in total 5 steps from 557. To simplify the synthetic route, aldehyde 557 was first protected as dioxolane 559 in quantitative yield without column purification. Lactam reduction of 559 using the NiCl2(dme)/PhSiH3 system afforded varied yield from 30 to 70% and a dioxolane ring opened side product was observed by

1H NMR analysis. This was probably due to the Lewis acidity of the nickel complex or oxidative addition to C-O bond.

Instead of using NiCl2(dme) catalyst, the Mo(CO)6/silane conditions, developed by Adolfsson et al, were investigated for the lactam reduction (Scheme 123).94j A survey of different silanes showed that PHMS gave a slow reaction. Trace amounts of reduced product 560 were observed from the 1H NMR after 24 hours and most of the starting material remained unreacted. Importantly no dixoxlane opened by-product was observed under these conditions. Changing the silane to PhSiH3 significantly improved the rate and yield of the reduction. Satisfactory yields (70 to 86%) were obtained consistently on different scales. The subsequent PMB group exchange with allyl chloroformate

136

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

performed well in 85% yield and the resulting 561 was then subjected to acid hydrolysis to give the corresponding aldehyde 526 in an overall more efficient sequence of reactions.

Decarboxylative allylation, followed by Knoevenagel condensation under previous conditions gave conjugate malonate 534 (Scheme 123).

At this stage, other Lewis acid were tested for the [1,4]-HT/cyclisation (Table 8).

The Lewis acid catalysed process performed better in toluene (entry 1,2) than 1,2- dichloroethane (entry 3), due to the different solubility of the catalysts in different solvents. It was observed that the reaction in toluene was heterogeneous and we speculated that there would be a lower amount of the tertiary amine complexed to the metal catalyst and that this may reduce degradation. The Lewis acid had a better solubility in 1,2-dichloroethane, that resulted in a faster HT/cyclisatio, but with an increased amount of degradation.

Instead of using Lewis acid, thiourea organocatalyst (Schreiner’s catalyst) was also investigated, but only a small conversion was observed after 24 h in refluxing toluene, probably due to insufficient activation.

137

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

Scheme 123. Optimised synthesis of 535

138

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

Table 8. Optimisation of 1,4-HT/cyclisation

Entry Condition Results

a 1 Yb(OTf)3 (10 mol%), PhMe, reflux 70% to 80%

2 Gd(OTf)3 (10 mol%), PhMe, reflux 70%

3 Yb(OTf)3 (10 mol%), ClCH2CH2Cl, reflux 57%

tiny conversion 4 (10 mol%) (checked by TLC) PhMe, reflux, 24 h

a. 80% yield was obtained when new Yb(OTf)3 was used.

5.3.2 Chemoselective reduction of a malonate

An investigation of the chemoselective malonate reduction of 535 commenced with

o temperature controlled LiAlH4 reduction (Table 9). At -78 C 535 was inert to LiAlH4 reduction. Conversion was only observed after raising the addition temperature to -40 oC and then warming. Pleasingly the carbamate was untouched. Based on the integration of the 1H NMR signal of C(21)-H [ 4.70 (1H, s)], an NMR yield in the crude material contained for the desired diol 546 could be determined (see Fig. 7 for a typical example of 535 ratio calculation from crude 1H NMR).

139

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

Table 9. Optimisation of LiAlH4 reduction

Entry Condition Yield of 546 from crude 1H NMR

LiAlH4 (3.0 eq.), THF (0.18 M), 1 No DP -78 oC， 1.5 h

LiAlH4 (3.0 eq.), THF (0.18 M), 2 40% -40 to rt; 1.5 h

LiAlH4 (3.0 eq.), THF (0.18 M), 50% 3 -40 to 0 oC; 3 h

LiAlH4 (4.0 eq.), THF (0.05 M), 4 56% 0 oC, 30 min

LiAlH4 (4.0 eq.), THF (0.05 M), 5 78% (61% isolated yield) 0 oC addition, then 10 min at rt

Ratio of 535: (0.62/1) x 100%=62% All allyl group containing products C(21)-H

Figure 7. A typical example of 535 NMR yield calculation from crude reduction products

The 1H NMR of crude reduction product (Fig. 7) showed the existence of aldehyde peaks (9-10 ppm). For those attempts at low temperature (< 0 oC), the total amount of aldehydes remained at ~40%, while the reduction at room temperature (entry 5, Table 9)

140

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

significantly reduced the amount of aldehydes to 10%. The mechanism of ester reduction using LiAlH4 is well known and normally involves a rapid collapse of a tetrahedral intermediate to the corresponding aldehyde and a second reduction to the desired alcohol

(Scheme 124). For malonate, we postulated that a chelate tetrahedral intermediate B could exist, similar to the stabilization of Weinreb amide tetrahedral intermediates, at relatively high reaction temperatures (e.g. 0 oC to rt). Intermediate B could collapse to the hydroxy aldehyde I, then a retro aldol reaction to II and further reduction to the diol

(Scheme 124). The high-resolution mass spectrum (HRMS) of the crude reduction mixture indicated the presence of several other products, including partially reduced compound 564, retro-aldol product 563 and completely reduced indoline compound 566

(Figure 8).

Figure 8. Mass spectroscopy (ESI-TOF) of crude product after LiAlH4 reduction

141

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

Scheme 124. Proposed reaction pathway of LiAlH4 reduction on malonates

It has been reported that malonates can be reduced to their corresponding diols with NaBH4 upon coordination of the metal to the 1,3-dicarbonyl group, effectively

95 forming an “activated ester”. Treatment of 535 with NaBH4 in methanol at elevated temperature afforded a new product, but 1H NMR of the crude material showed disappearance of the singlet peak of C(21)-H. It was suggested that the ring E opened during the reaction, leading to reduction products from a speculative intermediate 567

(Scheme 125). The results shown in Scheme 107 had revealed that the ring opening/closing process existed upon heating. Considering this reversible process on compound 535 could generate iminium ion 567 (Scheme 125), which could be quickly

. 96 reduced by NaBH4. Enhancing the reactivity of NaBH4 by adding catalytic CeCl3 7H2O resulted in reduction at room temperature, but 1H NMR of the crude product still suggested the formation of ring opened compounds. Commercially available LiBH4 was also tested and a similar result was observed.

142

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

Scheme 125. Malonate reduction using NaBH4

In addition to various borohydrides, DIBAL-H was also surveyed, but again, only ring opening products were observed. Super hydride afforded several compounds and the major product was identified as a mono-reduced compound, assigned as 568 (Scheme

126). The full reduction of the malonate could not be achieved with this reagent even by increasing the reaction temperature or the amount of reducing agent.

Scheme 126. Malonate reduction using Super-hydride

Ultimately, LiAlH4 reduction at room temperature for 10 to 15 min were the optimal conditions on large scale (100 to 500 mg) reductions, giving a constant isolated yield of 56% to 65%, which was acceptable for further synthesis.

5.3.3 First attempt towards (+)-strempeliopine

With a reliable route to diol 546 in hand, mono protection was performed. The more hindered TBDPSCl showed a much better selectivity and yield to 569 (55% over two steps, 38% over two steps for TBS protection, Scheme 127). The oxidation using IBX 143

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

proceeded in good yield as before but gave a worse ratio with epi-570 (1:1), the structure of which was confirmed by a NOESY experiment.

Interestingly, although the epimers were inseparable, we found that silica chromatography of the crude aldehyde gave a better ratio to 2:1 (570 : epi-570) and we suspected that a ring-opening/closing process occurred during the chromatography.

Extending this observation, we found that treatment of the 570 and epi-570 mixture with silica gel in dichloromethane with stirring at room temperature afforded the desired epimer with excellent ratio (9:1). The tandem Petasis olefination/ring closing metathesis reaction was the carried out to give the A/B/C/D/E ring intermediate 572 in 71% yield

(Scheme 127).

Scheme 127. Synthesis of A/B/C/D/E ring intermediate 572

To investigate the final lactam ring synthesis, the TBDPS group was removed

(Scheme 128). However, the desired alcohol product 575 was not isolated and 1H NMR analysis of the isolated product indicated the formation of cyclic carbamate 574 in 40%

144

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

vii yield. The mechanism was proposed that the resulting alcohol 573, generated from

TBDPS deprotection, cyclised onto the methyl carbamate to give a cyclic carbamate product.

At this stage, we speculated that we could use this unexpected cyclisation as a mono protection strategy which would negate the use of silyl protection. Therefore, diol

546 was treated with TBAF in THF at elevated temperature, but no cyclised product 576 was observed. Changing the reaction conditions to NaH in THF was also unsuccessful.

It was concluded that the cyclisation was sensitive to the conformation of the molecule.

From an inspection of a model of 573 it could be seen that the hydroxyl group was pushed down much closer to the methyl carbamate because of the fused ring skeleton, making cyclisation much more favourable than for 546.

Scheme 128. Unexpected formation of cyclic carbamate 574

An option to continue the synthesis from 574 would involve ring opening of the cyclic carbamate and an alternative functionalisation. We were interested in adopting

Qin’s method to construct final lactam ring that would involve attack of the free dihydro-

vii low yield may be due to insufficient extraction

145

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

indoline on to a cyano group (Scheme 129). We sought a decarboxylative substitution of

574 with cyanide (or halide) to give 580. Attempted direct decarboxylative substitution using TMSI and NaCN was unsuccessful. Harsher reaction conditions by mixing 574 with KOH in refluxed ethanol gave the hydrolysed product 581 efficiently.

Unfortunately, the transformation from alcohol to alkyl halide or mesylate 580 failed with degradation and/or rearrangement to an inseperable mixture of unknown products

(Scheme 129).

The biosynthesis of schizozygine type alkaloids provided a clue that the 1,2- diamine substructure could undergo rearrangement, especially when the indoline nitrogen was free from protection (Scheme 25, 37). We had also observed an intramolecular hydrogen bond between hydroxyl group and indoline nitrogen in compound 541 confirmed in a NOESY experiment (Scheme 130, phenomenon 1). Although this H-bond effect was not directly observed in the 1H NMR of 581, we speculated that this interaction still existed and served to sterically inhibit the reactivity of the hydroxyl group.

146

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

Scheme 129. Unsuccessful attempt towards strempeliopine

Scheme 130. Observation of an internal H-bond between hydroxyl group and indoline nitrogen

and large polarity change before and after the RCM reaction

147

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

Additionally, we observed a large polarity change before and after the ring closing metathesis reaction by thin layer chromatography (TLC) (Scheme 130, phenomenon 2).

This suggested that the tertiary amine in 572 was much more basic and would be more nucleophilic to trap the electrophile during the halogenation and mesylation reactions.

To solve these problems, the order of the late stage manipulations needed to be rearranged (Scheme 131). It was known that the tertiary amine in compound 571 was shrouded by sterics (i.e. non-basic) and could be inert towards reactive electrophiles.

Thus, the modification of the C-19 chain from hydroxyl group to alkyl halide might be feasible. In addition, the silyl group deprotection of 571 would probably not undergo the undesired cyclisation, based on the unsuccessful cyclisation of diol 546 (Scheme 128).

Therefore, the ring closing metathesis could be performed on the latent cyclisation precursor 583 and subsequent lithium-halogen exchange/lactamisation would close the ring F, encouraged by the formation of the ring fused skeleton (Scheme 131).

Scheme 131. Revised plan on late stage manipulations

148

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

5.3.4 Second attempt towards (+)-strempeliopine

The revised synthesis began from alcohol 569 (Scheme 132). The IBX oxidation yielded the aldehyde 570, which was directly charged with Petasis reagent to afford dialkene 571 after simple work-up by quenching the excess Petasis reagent with hot aqueous methanol.

The crude 571 was treated with excess TBAF in THF at elevated temperature to finally give the alcohol 582 in 65% yield. To our delight, no unexpected cyclised product 585

(see Scheme 131) was detected. This three-steps protocol was practical and efficient with only one chromatography purification after silyl group removal.

Scheme 132. Silyl group removal before ring closing metathesis

Although feasible, the conversion of the hydroxyl group in 582 to bromide or iodide was still challenging due to the steric hindrance caused by the adjacent quaternary centre, the neo-pentyl effect (Scheme 133). The first attempt using the Appel reaction failed to give any product. A phosphine-free method, developed by Paquin et al,97 using tetra-n-ethylammonium halide/[Et2NSF2]BF4 (XtalFluoro E), was adopted, but no desired bromide was detected. Alcohol 582 was successfully transformed to the mesylate 587

149

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

quantitatively, but the subsequent displacement using tetra-n-butylammonium bromide caused degradation. We proposed that the degradation at elevated temperature (65 to 100 oC) could come from the Grob type fragmentation and this instability would not be possible after the ring closing metathesis.

Scheme 133. Conversion of a hydroxyl group to a halide

Therefore, compound 587 was exposed to catalytic Hoveyda-Grubbs 2nd generation catalyst in toluene at 60 oC to give the cyclised product 588 (Scheme 134).

However, displacement of the mesylate of 588 with NaI and NaBr failed with slow degradation. Changing the nucleophile to KCN in hot DMSO (100 oC) cleanly yielded a mixture of two products. One was confirmed to be the cyclic carbamate 574 (32%). The other product, after careful analysis of 1D and 2D NMR and HRMS, was elucidated as compound 590 (31%) (Figure 9).

150

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

Scheme 134. Formation of unexpected unnatural products 574 and 590

Figure 9. 2D NMR (NOESY and HMBC) on compound 590.

A plausible mechanism was proposed for the formation of 574 and 590 (Scheme

135). As the mesylate was sterically hindered due to the adjacent quaternary centre, the direct displacement through an SN2 pathway was not favoured. Instead, cyclisation from the carbamate may occur and the resulting oxonium ion would then be demethylated to

574. The pathway to 590 could involve a demethylation/decarboxylation sequence to anion 592, which then cyclised to give 590. Unnatural compound 590 represented an unusual and highly rigid skeleton.

151

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

Scheme 135. Proposed mechanism of formation of 574 and 590.

5.3.5 Third attempt towards (+)-strempelipine

Based on the previous results, C-20 manipulations through an sp3 carbon were unsuccessful while oxidation of the C-20 alcohol to aldehyde and subsequent aldehyde manipulation to an alkene were applicable (i.e. 541 to 387, Scheme 117). Therefore, at an appropriate stage, oxidation of the C-19 alcohol to the corresponding aldehyde and then methylation would be an alternative method for a one-carbon extension.

Scheme 136. Ring closing metathesis of 582 with free hydroxyl group

Using 582, we first checked if the ring closing metathesis reaction in the presence of a hydroxyl group was possible (Scheme 136). Although metathesis reactions in the 152

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

presence of hydroxyl groups have been widely reported, the reaction on 582 under catalytic Hoveyda-Grubbs 2nd generation catalyst did not yield desired cyclised product and most of starting material was recovered. Addition of stoichiometric amount of Ti(Oi-

98 1 Pr)4 made the cyclisation occur but the H NMR indicated the structure of cyclic carbamate 574. This indicated that basic conditions (e.g. TBAF) were not essential for this cyclic carbamate formation.

Instead of conducting a ring closing metathesis, oxidation of the hydroxyl group in 582 was attempted. The Swern oxidation cleanly afforded the aldehyde but with a mixture of C-20 epimers (2.5:1), due to the fast ring E opening/closing epimerisation

(Scheme 137). At this stage, a Wittig reaction would be the best way of utilising both isomers and the single isomer of trialkene compound 594 could be obtained. To our delight, treatment of aldehyde 593 with concentrated and excess methylene- triphenylphosphorane (Ph3P=CH2) afforded desired trialkene 594 in 76% overall yield from 582.

Scheme 137. Oxidation/Olefination sequence to trialkene 594

With the trialkene 594 in hand, the ring closing metathesis was performed and cyclised compound 595 was obtained as a single diastereomer in 80% yield (Scheme 138).

As the metathesis reaction is reversible, cyclisation with ‘exo-olefin’ could be accessible but would be structurally disfavoured and equilibration via a series of [2+2]/retro [2+2]

153

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

process finally gives the thermodynamically favoured isomer 595. The stereochemistry of 595 was confirmed by NOESY experiment. This optimised route was closer to

Feldman’s example, with similar regioselectivity obtained in the ring closing metathesis reaction. Although the bis-oxidation/olefination strategy Feldman used was not applicable in our case (Scheme 138).

Scheme 138. Regioselective ring closing metathesis of trialkene 594 and comparison with

Feldman’s work

At this stage, compound 595 was recognised as a versatile intermediate that could be used to obtain both (+)-vallesamidine and (+)-strempeliopine (Scheme 139).

Reduction of 595 with LiAlH4 gave 387, the precursor that had been synthesised in our first route (Scheme 118) and gave (+)-vallesamidne (102) via catalytic hydrogenation.

154

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

Scheme 139. Alternative preparation of vallesamidine (102) from 595

With compound 595 in hand, the synthesis of (+)-strempeliopine (103) could be investigated further. Cross metathesis with vinyl boronate to install the hydroxyl or aldehyde group at the terminal carbon by simple oxidation was attempted. Two vinyl boronates, vinyl boronic acid pinacol ester 59899 and vinyl boronic acid MIDA ester

599100 were investigated but no metathesis reaction was observed, probably due to the hindered nature of the terminal alkene in 595 (Scheme 140).

Scheme 140. Unsuccessful cross metathesis with vinyl boronates

Movassaghi and co-workers reported a total synthesis of (-)-deoxoapodine 603 via an enantioselective ring closing metathesis (Scheme 141).101 After obtaining the

RCM product 601, a Wacker oxidation, followed by the reduction of the carbonyl on the tertiary vinyl alkene gave the desired primary alcohol 602 in 79% yield.

155

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

Scheme 141. Movassaghi’s total synthesis of (-)-deoxoapodine

This approach was then examined (Scheme 142). Considering the presence of basic nitrogen, 1.45 equivalents of HClO4 was used for amine protection. However, the reaction proceeded slowly and did not go to completion after 4 days. Among mostly unreacted starting material, one minor product was isolated and confirmed to be the methyl ketone 604 other than the desired aldehyde 597. No aldehyde signal was observed in the 1H NMR.

Scheme 142. Wacker oxidation on alkene 595

We then turned to the more general and classic hydroboration/oxidation sequence.

For the regioselective hydroboration, bulky borane (e.g. 9-BBN, c-Hex2BH, Sia2BH)

156

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

would be needed to react with the least hindered terminal alkene in the presence of internal and electronically more reactive alkene (Scheme 143).102 We chose the bulky dicyclohexylborane (c-Hex2BH) for our investigation and prepared the reagent directly before use, hoping to maximise our chances of achieving the desired selectivity. Our first attempt by treatment of compound 595 with c-Hex2BH, followed by oxidation using

H2O2/NaOH aq. system did not yield the alcohol 609. To our relief, changing the

- oxidation condition from H2O2/OH to a milder NaBO3/H2O system smoothly gave the desired product 609 in 75% yield.

Scheme 143. Regioselective hydroboration/oxidation

With alcohol 609 in hand, hydrolysis of methyl carbamate with aqueous KOH in methanol at elevated temperature gave free the indoline compound 610 which was used directly after simple work-up (Scheme 14). The final lactam ring F formation involved primary alcohol oxidation to an aldehyde, which the indoline nitrogen attacked to generate the aminol and a second oxidation took place in situ to give the desired amide.

During the condition screening (0.01 mmol scale, structure elucidated by 1H NMR and

HRMS), it was found that IBX oxidation in refluxed EtOAc gave a mixture of several products, one of which was elucidated as 611, determined by the presence of two additional olefin protons’ signals at 6.40 (1H, d, J = 7.0) and 5.13 (1H, d, J = 7.0) ppm.

This result illustrated that the alcohol oxidation and aminol formation occurred, but the subsequent oxidation did not proceed. Instead, an elimination process occurred, 157

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

presumably due to heating of the reaction, to give 611. A similar case in Scheme 34, reported by Padwa and co-workers, showed that their final oxidative lactamisation required sequential oxidation using Dess-Martin periodinane and PCC.

After that, the Griffiths-Ley oxidation (cat. TPAP/NMO) was tested. The initial test using ‘excess of the oxidant’ did cyclise the lactam ring F, but also oxidised the allylic position (C-3) to afford the 3-oxo-14,15-dehydrostrempeliopine 612 (Scheme 144).

Control of the amount of reagent usage using a stock solution (10 mol% TPAP, 2.1 eq.

NMO) furnished the cyclisation without over oxidation and desired product 543 was obtained in 40% yield over two steps. The 1H NMR data of 543 was compared with the literature reported data of schizozygine and the signals in the non-aromatic region, including the splitting patterns, fitted well and supported the assignment of 543 and therefore the total synthesis of (+)-14,15-dehydrostrempeliopine.

Scheme 144. Formation of lactam ring F

158

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

Table 10. Spectra comparison (1H and 13C) between 543 and schizozygine

1H NMR

H-19 2.70 (d, 18.0) 2.61 (d, 18.0) H-19’ 2.54 (dd, 18.0, 2.8) 2.45 (dd, 18.0, 2.8) H-21 2.27 (s) 2.25 (s) H-14 5.57 (ddd, 10.0, 4.5, 2.0) 5.57 (ddd, 10.0, 4.4, 2.0) H-15 5.73 (dt, 9.8, 2.2) 5.73 (dt, 9.9, 2.2) H-7 3.31 (t, 6.6) 3.20 (t, 6.6) H5 2.29 (ddd, 12.0, 7.1, 4.8), 2.27 (m) H5’ 3.06 (ddd, 11.5, 7.2, 5.2) 3.03 (brs) H6/6’ 2.16—2.06 (m) 2.04 (m) 13C NMR C3 53.5 53.5 C5 50.3 50.1 C6 25.5 25.8 C7 42.1 42.1 C2 71.9 72.6 C21 68.5 68.0 C20 44.7 44.7 C19 47.2 46.9 C18 169.8 168.9 C14 123.8 123.8 C15 130.2 130.2 C16 38.5 38.6 C17 37.7 37.6

159

Chapter 5. Studies towards the total synthesis of (+)-strempeliopine

5.4 Chapter conclusion

In summary, an optimised synthesis starting from the [1,4]-HT/cyclisation product 535 was developed and a divinyl intermediate 594 was prepared to conduct the regioselective ring closing metathesis. The resulting A/B/C/D/E ring compound 595 was then divergently applied to furnish the total synthesis of (+)-vallesamidine (102) and (+)-

14,15-dehydrostrempeliopine (543). Although the final step to strempeliopine (103) via catalytic hydrogenation was not attempted due to a lack of material, the product 543,

14,15-dehydrostrempeliopine, shares the identical core structure as schizozygine and suggests that the route is divergent and could be used for the synthesis of related alkaloids and unnatural derivatives.

160

Conclusions and future work

This research chose the monoterpene indole alkaloid schizozygine and related molecules as synthetic targets to investigate the application of nitro chemistry, especially nitro-

Mannich reaction to develop a divergent synthetic route to them.

In the initial investigation towards schizozygine, a Lewis acid mediated, nitro-

Mannich reaction on -branched nitroalkanes was developed to prepare complex nitroamines with three contiguous chiral centres and good diastereolectivity (5 to 7:1 dr, syn,anti predominant). The resulting nitroamine, after a reductive cyclisation using

. BF3 OEt2/Et3SiH condition, achieved the highly functionalised piperidine. However, subsequent manipulations on the ester group in 328 were unsuccessful and none of our attempts could transform the ester group to the corresponding vinyl ketone required, to explore the key phosphine catalysed cyclisation to ring E. Alternative approaches, including Dieckmann cyclisation, did not furnish the ring E synthesis.

The studies on schizozygine were discontinued due to the supply limitation of the starting material piperonal. Our efforts were then turned to the synthesis of vallesamidine and strempeliopine. The newly developed route to (+)-vallesamidine still used the nitro-

Mannich reaction as a key transformation to construct the ring B, and also made full use of nitro chemistry, including Michael addition, Tsuji-Trost allylation, nitro reduction/C-

N coupling reaction, to build the A/B/C ring intermediate in 7 steps with minimised purification. As another key reaction, the intramolecular C-H functionalisation method,

[1,4]-hydride transfer/cyclisation strategy was developed to synthesise ring E. This interesting transformation, although studied widely in methodology development, has been rarely adopted in the synthesis of complex molecules. Our research contributed a

161

Conclusions and future work good example of such a reaction and may provide potential application in other complex molecular synthesis.

The final divergent route commenced from the [1,4]-HT/cyclisation product 535 and the alkaloids (+)-vallesamidine and (+)-14,15-dehydrostrempeliopine were synthesised from the late stage manipulations of the intermediate 595, prepared from a regioselective ring closing metathesis of trialkene 594. Three other unnatural analogues,

581, 574 and 590, were also prepared during the studies. Most importantly, the C-14,15 double bond construction via ring closing metathesis provided a strategic route that could be used for the synthesis of the schizozygine type molecules.

Future work would focus on the synthesis of schizozygine and related molecules.

Other modifications would be investigated to shorten the synthetic route and make the route more efficient to scale-up. The molecules synthesised would also be evaluated for biological activity and to develop new biological applications.

162

References

[1] Carey, F. A., Sundberg, R. J. Advanced organic chemistry: Part B. New York, NY:

Springer, 2007.

[2] Noble, A.; Anderson, J. C. Nitro-Mannich Reaction. Chem. Rev. 2013, 113, 2887–2939.

[3] Weng, J.; Li, Y. B.; Wang, R. Bin; Li, F. Q.; Liu, C.; Chan, A. S. C.; Lu, G. A Practical

and Azide-Free Synthetic Approach to Oseltamivir from Diethyl d -Tartrate. J. Org. Chem.

2010, 75, 3125–3128.

[4] Hayashi, Y.; Itoh, T.; Ishikawa, H. Oxidative and Enantioselective Cross-Coupling of

Aldehydes and Nitromethane Catalyzed by Diphenylprolinol Silyl Ether. Angew. Chemie

Int. Ed. 2011, 50, 3920–3924.

[5] Tsuritani, N.; Yamada, K.-I.; Yoshikawa, N.; Shibasaki, M. Catalytic Asymmetric

Syntheses of ICI-199441 and CP-99994 Using Nitro-Mannich Reaction. Chem. Lett. 2002,

31, 276–277.

[6] Handa, S.; Gnanadesikan, V.; Matsunaga, S.; Shibasaki, M. Heterobimetallic Transition

Metal/Rare Earth Metal Bifunctional Catalysis: A Cu/Sm/Schiff Base Complex for Syn -

Selective Catalytic Asymmetric Nitro-Mannich Reaction. J. Am. Chem. Soc. 2010, 132,

4925–4934.

[7] Pei, Z.; Li, X.; Von Geldern, T. W.; Longenecker, K.; Pireh, D.; Stewart, K. D.; Backes,

B. J.; Lai, C.; Lubben, T. H.; Ballaron, S. J.; Beno, D. W. A.; Kempf-Grote, A. J.; Sham,

H. L.; Trevillyan, J.M. Discovery and Structure - Activity Relationships of Piperidinone-

and Piperidine-Constrained Phenethylamines as Novel, Potent, and Selective Dipeptidyl

Peptidase IV Inhibitors. J. Med. Chem. 2007, 50, 1983–1987.

[8] El-Desouky, S. K.; Ryu, S. Y.; Kim, Y. K. Piperazirum, a Novel Bioactive Alkaloid from

Arum Palaestinum Boiss. Tetrahedron Lett. 2007, 48, 4015–4017.

163

[9] Anderson, J. C.; Kalogirou, A. S.; Porter, M. J.; Tizzard, G. J. Synthesis of the Reported

Structure of Piperazirum Using a Nitro-Mannich Reaction as the Key Stereochemical

Determining Step. Beilstein J. Org. Chem. 2013, 9, 1737–1744.

[10] Hynes, P. S.; Stupple, P. A.; Dixon, D. J. Organocatalytic Asymmetric Total Synthesis of

(R)-Rolipram and Formal Synthesis of (3S,4R)-Paroxetine. Org. Lett. 2008, 10, 1389–

1391.

[11] Jakubec, P.; Cockfield, D. M.; Dixon, D. J. Total Synthesis of (-)-Nakadomarin A. J. Am.

Chem. Soc. 2009, 131, 16632–16633.

[12] Shen, B.; Johnston, J. N. A Formal Enantioselective Acetate Mannich Reaction: The Nitro

Functional Group as a Traceless Agent for Activation and Enantiocontrol in the Synthesis

of β-Amino Acids. Org. Lett. 2008, 10, 4397–4400.

[13] Kürti, L., & Czakó, B. Strategic applications of named reactions in organic synthesis:

Background and detailed mechanisms. Amsterdam: Elsevier Academic Press, 2005.

[14] Ballini, R.; Petrini, M. The Nitro to Carbonyl Conversion (Nef Reaction): New

Perspectives for a Classical Transformation. Adv. Synth. Catal. 2015, 357, 2371-2402

[15] Palomo, C.; Oiarbide, M.; Halder, R.; Laso, A.; López, R. Enantioselective Aza-Henry

Reactions Assisted by ZnIIand N-Methylephedrine. Angew. Chemie - Int. Ed. 2005, 45,

117–120.

[16] García-Muñoz, M. J.; Dema, H. K.; Foubelo, F.; Yus, M. Base-Promoted

Diastereoselective Addition of Nitromethane and Nitroethane to N-Tert-Butylsulfinyl

Imines: Synthesis of N-Protected α-Amino Acids and Amino Ketones. Tetrahedron

Asymmetry 2014, 25, 362–372.

[17] Jakubec, P.; Hawkins, A.; Felzmann, W.; Dixon, D. J. Total Synthesis of Manzamine A

and Related Alkaloids. J. Am. Chem. Soc. 2012, 134, 17482–17485.

[18] Seigler, D. S. Plant Secondary Metabolism; Springer, 1998.

164

[19] Saxton, J. E. Indoles, Part 4. New York: John Wiley, 1983, 1-882.

[20] O’Connor, S. E.; Maresh, J. J. Chemistry and Biology of Monoterpene Indole Alkaloid

Biosynthesis. Nat. Prod. Rep. 2006, 23, 532–547.

[21] Renner, U. Alkaloids of Schizozygia Caffaeoides ( Boj ) Baill . Structure Determinations

of Schizozygine Some Companion Alkaloids. Lloydia 1964, 27, 406–415.

[22] Omino, E. A.; Kokwaro, J. O. Ethnobotany of Apocynaceae Species in Kenya. J.

Ethnopharmacol. 1993, 40, 167–180.

[23] a) Kariba, R. M.; Houghton, P. J.; Yenesew, A. Antimicrobial Activities of a New

Schizozygane Indoline Alkaloid from Schizozygia Coffaeoides and the Revised Structure

of Isoschizogaline. J. Nat. Prod. 2002, 65, 566–569. b) Kariba, R. M.; Siboe, G. M.; Dossaji,

S. F. In vitro antifungal activity of Schizozygia coffaeoides Bail. (Apocynaceae) extracts

J. Ethnopharmacology 2001, 74, 41-44.

[24] Atilaw, Y.; Heydenreich, M.; Ndakala, A.; Akala, H. M.; Kamau, E.; Yenesew, A. 3-Oxo-

14α,15α-Epoxyschizozygine: A New Schizozygane Indoline Alkaloid from Schizozygia

Coffaeoides. Phytochem. Lett. 2014, 10, 28–31.

[25] Walser, A.; Djerassi, C. Alkaloid‐Studien LII. Die Alkaloide Aus Vallesia Dichotoma

Ruizet Pav. Helv. Chim. Acta 1965, 48 (2), 391–404.

[26] Laguna, A.; Novotny, L.; Dolejs, L.; Budesinsky, M. Alkaloids from Roots of

Strempeliopsis Strempelioides - Structures of Strempeliopine and Strempeliopidine. Planta

Med. 1984, 50, 285–288.

[27] Cavé, A.; Bruneton, J.; Ahond, A.; Bui, A. M.; Husson, H. P.; Kan, C.; Lukacs, G.; Potier,

P. Structure Analysis by 13C NMR Spectroscopy of Criophylline, a New Dimeric Indole

Alkaloid. Tetrahedron Lett. 1973, 14, 5081–5084.

165

[28] a) Floriani, F. An. Farm. Bioquim. 1931, 2, 215. b) Banerjee, J. N.; Lewis, J. J. A note on

the pharmacology of aspidospermine. J. Pharm. Pharmacol. 1955, 7, 46–49. c) Witkop, B.

Aspidospermine. I J. Am. Chem. Soc. 1948, 70, 3712–3716.

[29] Magadula, J. J.; Erasto, P. Bioactive natural products derived from the East African flora

Nat. Prod. Rep. 2009, 26, 1535.

[30] Atilaw, Y.; Heydenreich, M.; Ndakala, A.; Akala, H. M.; Kamau, E.; Yenesew, A. 3-Oxo-

14α,15α-Epoxyschizozygine: A New Schizozygane Indoline Alkaloid from Schizozygia

Coffaeoides. Phytochem. Lett. 2014, 10, 28–31.

[31] Brown, S. H.; Djerassi, C.; Simpson, P. G. Alkaloid Studies. LIX. Structure and Absolute

Configuration of Vallesamidine. J. Am. Chem. Soc. 1968, 90, 2445–2446.

[32] M. F. Grundon Alkaloids, volumn 6, Royal Society of Chemistry, ISBN 1847555616).

[33] Hajicek, J.; Trojanek, J. On Alkaloids. XLIV. Absolute Configuration of (-)-

Strempeliopine. Tetrahedron Lett. 1982, 23, 365–368.

[34] Miettinen, K.; Dong, L.; Navrot, N.; Schneider, T.; Burlat, V.; Pollier, J.; Woittiez, L.; Van

Der Krol, S.; Lugan, R.; Ilc, T.; et al. The Seco-Iridoid Pathway from Catharanthus Roseus.

Nat. Commun. 2014, 5 (1), 3606.

[35] Lévy, J.; Maupérin, P.; de Maindreville, M. D.; Le Men, J. Methylene-Indolines,

Indolenines et Indoleniniums-III (1). Action de Reactifs Reducteurs 3. Hemisynthese de La

(+)-Vallesamidine. Tetrahedron Lett. 1971, 12, 1003–1006.

[36] Dickman, D. A.; Heathcock, C. H. Total Synthesis of (±)-Vallesamidine. J. Am. Chem. Soc.

1989, 111, 1528–1530.

[37] Heathcock, C. H.; Norman, M. H.; Dickman, D. A. Total Synthesis of (±)-Vallesamidine.

J. Org. Chem. 1990, 55, 798–811.

166

[38] Antunes, O. A. C.; Bergter, L.; Costa, P. R. R.; Castro, R. N.; Farias, F. M. C. Formal

Synthesis of (-)-Vallesamidine A 2,2,3-Trialkylindoline Alkaloid. Tetrahedron:

Asymmetry 2002, 4 (7), 1499–1500.

[39] Harley-Mason, J.; Kaplan, M. A simple total synthesis of (±)-aspidospermidine. Chem.

Commun. 1967, 915-916.

[40] Tanino, H.; Fukuishi, K.; Ushiyama, M.; Okada, K. Total Syntheses of (-)-Vallesamidine

and Related Aspidosperma and Hunteria Type Indole Alkaloids from the Common

Intermediate. Tetrahedron 2004, 60, 3273–3282.

[41] Kobayashi, S., Jørgensen, K. A., Eds.; Cycloaddition Reactions in Organic Synthesis;

Wiley-VCH Verlag GmbH: Weinheim, FRG, 2001.

[42] Padwa, A.; Harring, S. R.; Semones, M. A. Cycloaddition Reaction of Mesoionic Betaines

as an Approach toward Trialkylindoline Alkaloids. J. Org. Chem. 1998, 63, 44–54.

[43] Wang, X.; Xia, D.; Qin, W.; Zhou, R.; Zhou, X.; Zhou, Q.; Liu, W.; Dai, X.; Wang, H.;

Wang, S.; et al. A Radical Cascade Enabling Collective Syntheses of Natural Products.

Chem 2017, 2, 803–816.

[44] Hajicek, J.; Trojanek, J. On Alkaloids. XL. Stereospecific Total Synthesis of (±)-

Strempeliopine. Tetrahedron Lett. 1981, 22, 2927–2928.

[45] Kuehne, M. E.; Huebner, J. A.; Matsko, T. H. Studies in Biomimetic Alkaloid Syntheses.

4. An Alternative Route to Secodine Intermediates Providing Syntheses of Minovine,

Vincadifformine, Ervinceine, and N(a)-Methylervinceine. J. Org. Chem. 1979, 44, 2477–

2480.

[46] Bobeck, D. R.; Lee, H. I.; Flick, A. C.; Padwa, A. Application of Cross-Conjugated

Heteroaromatic Betaines to the Synthesis of the Schizozygane Alkaloid (+/-)-

Strempeliopine. J. Org. Chem. 2009, 74, 7389–7402.

167

[47] Trost, B. M.; Braslau, R. A Convenient Chemoselective Semihydrogenation of Acetylenes

Using Homogeneous Catalysis. Tetrahedron Lett. 1989, 30, 4657–4660.

[48] Zhou, Q.; Dai, X.; Song, H.; He, H.; Wang, X.; Liu, X. Y.; Qin, Y. Concise Syntheses of

Eburnane Indole Alkaloids. Chem. Commun. 2018, 54, 9510–9512.

[49] Hajicek, J. Recent Developments in Syntheses of the Post-Secodine Indole Alkaloids. Part

III. Rearranged Alkaloid Types. Collect. Czech. Chem. Commun. 2011, 76 (12), 2023–

2083.

[50] Hubbs, J. L.; Heathcock, C. H. Total Synthesis of (±)-Isoschizogamine. Org. Lett. 1999, 1,

1315–1317.

[51] Examples on RCM reaction in alkaloids synthesis: For examples of ring closing metathesis

in indole alkaloids synthesis, see: a) Feldman, K. S.; Antoline, J. F. Synthesis studies on

the Melodinus alkaloid meloscine. Tetrahedron 2013, 69, 1434–1445. b) Hayashi, Y.;

Inagaki, F.; Mukai,C. Total synthesis of (±)-meloscine. Org. Lett. 2011, 13, 1778–1780.

c) Feldman,K. S.; Antoline, J. F. Allenyl azide cycloaddition chemistry: Application

to the total synthesis of (±)-meloscine. Org. Lett. 2012, 14, 934–937. d) Kozmin, S. A.;

Iwama, T.; Huang, Y.; Rawal, V. H. An Efficient Approach to Aspidosperma Alkaloids

via [4 + 2] Cycloadditions of Aminosiloxydienes: Stereocontrolled Total Synthesis of (()-

Tabersonine. Gram-Scale Catalytic Asymmetric Syntheses of (+)-Tabersonine and (+)-16-

Methoxytabersonine. Asymmetric Syntheses of (+)-Aspidospermidine and (-)-

Quebrachamine J. Am. Chem. Soc. 2002, 124, 4628–4641. e) Selig, P.; Bach, T. Angew.

Chem. Int. Ed. 2008, 47, 5082–5084. f) Selig, P.; Herdtweck, E.; Bach, T. Total Synthesis

of Meloscine by a [2+2]-Photocycloaddition/Ring-Expansion Route. Chem. Eur. J.

2009, 15, 3509–3525. g) Zhang, H.; Curran, D. P.; A Short Total Synthesis of (()-

Epimeloscine and (()-Meloscine Enabled by a Cascade Radical Annulation of a

168

Divinylcyclopropane. J. Am. Chem. Soc 2011, 133, 10376–10378. h) Zhang, H.; Jeon, K.

O.; Ben Hay, E.; Geib, S. J.; Curran, D. P.; Laporte, M. G. Radical [3 + 2]-Annulation

of Divinylcyclopropanes: Rapid Synthesis of Complex Meloscine Analogs. Org.

Lett 2014, 16, 94–97. i) Goldberg, A. F. G.; Craig, R. A.; O’Connor, N. R.; Stoltz, B. M.

Highly Functionalized Donor–Acceptor Cyclopropanes Applied toward the Synthesis of

the Melodinus Alkaloids. Tetrahedron Lett. 2015, 56, 2983–2990. j) Goli, N.; Kallepu, S.;

Mainkar, P. S.; Lakshmi, J. K.; Chegondi, R.; Chandrasekhar, S. Synthetic Strategy toward

the Pentacyclic Core of Melodinus Alkaloids. J. Org. Chem. 2018, 83 (4), 2244–2249.

[52] En, D.; Zou, G. F.; Guo, Y.; Liao, W. W. Nucleophilic Phosphine-Catalyzed Intramolecular

Michael Reactions of N-Allylic Substituted α-Amino Nitriles: Construction of

Functionalized Pyrrolidine Rings via 5-Endo-Trig Cyclizations. J. Org. Chem. 2014, 79,

4456–4462.

[53] Anderson, J. C.; Stepney, G. J.; Mills, M. R.; Horsfall, L. R.; Blake, A. J.; Lewis, W.

Enantioselective Conjugate Addition Nitro-Mannich Reactions: Solvent Controlled

Synthesis of Acyclic Anti- and Syn-β-Nitroamines with Three Contiguous Stereocenters.

J. Org. Chem. 2011, 76, 1961–1971.

[54] Gotoh, H.; Okamura, D.; Lshikawa, H.; Hayashl, Y. Diphenylprolinol Silyl Ether as a

Catalyst in an Asymmetric, Catalytic, and Direct Michael Reaction of Nitroethanol with

α,β-Unsaturated Aldehydes. Org. Lett. 2009, 11, 4056–4059.

[55] Mei, T. S.; Wang, X.; Yu, J. Q. Pd(II)-Catalyzed Amination of C-H Bonds Using Single-

Electron or Two-Electron Oxidants. J. Am. Chem. Soc. 2009, 131, 10806–1080.

[56] Tsunoda, T.; Suzuki, M.; Noyoro, R. Reaction of Acetals and Trialkylsilanes Catalyzed by

Trimethylsilyl Trifluoromethanesulfonate. A Simple Method for Conversion of Acetals to

Ethers. Tetrahedron Lett. 1979, 20, 4679–4680.

169

[57] Booth, H.; Mark Dixon, J.; Khedhair, K. A. Experimental Studies of the Anomeric Effect.

Part VI. Ring Inversion Equilibria in Cyclohexane, Tetrahydropyran and Piperidine Rings

Substituted by a Carbomethoxy or a Cyano Group. Tetrahedron 1992, 48, 6161–6174.

[58] Hoffmann, R. W. Allylic 1,3-Strain as a Controlling Factor in Stereoselective

Transformations. Chem. Rev 1989, 89, 1841–1860.

[59] Imashiro, R.; Uehara, H.; Barbas, C. F. One-Pot Enantioselective Syntheses of Iminosugar

Derivatives Using Organocatalytic Anti -Michael- Anti -Aza-Henry Reactions. Org. Lett.

2010, 12, 5250–5253.

[60] a) Pennell, M. N.; Turner, P. G.; Sheppard, T. D. Gold- and Silver-Catalyzed

Reactions of Propargylic Alcohols in the Presence of Protic Additives. Chem. Eur. J. 2012,

18, 4748–4758. b) Imagawa, H.; Asai, Y.; Takano, H.; Hamagaki, H.; Nishizawa, M.

Mercuric Triflate-Catalyzed Reaction of Propargyl Acetates with Water Leading Vinyl

Ketones. Org. Lett. 2006, 8, 447–450.

[61] Bonini, C.; Chiummiento, L.; Videtta, V.; Colobert, F.; Solladié, G. A Concise and

Efficient Stereoselective Synthesis of the C1-C11 Fragment of Macrolactin A. Synlett

2006, 2427–2430.

[62] Ware, O. PhD Thesis, University College London, 2018.

[63] For reviews on [1,n]-HT/cyclisation, see: a) Peng, B.; Maulide, N. The Redox-Neutral

Approach to C-H Functionalization. Chem. Eur. J. 2013, 19, 13274–13287. b) Haibach, M.

C.; Seidel, D. C-H Bond Functionalization through Intramolecular Hydride Transfer.

Angew. Chemie - Int. Ed. 2014, 53, 5010–5036. c) Wang, L.; Xiao, J. Hydrogen-Atom

Transfer Reactions. Top. Curr. Chem. 2016, 374, 17.

[64] Broeke, J. ten; Douglas, A. W.; Grabowski, E. J. J. A Symmetrical Diazaditwistane. 2,9-

Dicyano-5,11-Dimethyl-5,11-Diazatetracyclo[6.2.2.02.7.04.9]Dodecane. J. Org. Chem.

1976, 41, 3159–3163.

170

[65] a) Heathcock, C. H.; Hansen, M. M.; Ruggeri, R. B.; Kath, J. C. Daphniphyllum alkaloids.

11. Biomimetic total synthesis of methyl homosecodaphniphyllate. Development of the

tetracyclization reaction J. Org. Chem. 1992, 57, 2544 –2553; b) Heathcock, C. H. Nature

knows best: An amazing reaction cascade is uncovered by design and discovery Proc. Natl.

Acad. Sci. USA 1996, 93, 14323 – 14327.

[66] Verboom, W.; Hamzink, M. R. J.; Reinhoudt, D. N.; Visser, R. Novel applications of the

“t-amino effect” in heterocyclic chemistry. Synthesis of a pyrrolo[1,2-a]quinazoline and

5H-pyrrolo[1,2-a][3,1]benzothiazines Tetrahedron Lett. 1984, 25, 4309–4312;

[67] (a) Murarka, S.; Zhang, C.; Konieczynska, M. D.; Seidel, D. Lewis Acid Catalyzed

Formation of Tetrahydroquinolines via an Intramolecular Redox Process. Org. Lett. 2009,

11, 129–132. (b) Han, Y. Y.; Han, W. Y.; Hou, X.; Zhang, X. M.; Yuan, W. C. FeCl 3-

Catalyzed Stereoselective Construction of Spirooxindole Tetrahydroquinolines via

Tandem 1,5-Hydride Transfer/Ring Closure. Org. Lett. 2012, 14, 4054–4057.

[68] Zhou, G.; Zhang, J. Product-Selectivity Control by the Nature of the Catalyst: Lewis Acid-

Catalyzed Selective Formation of Ring-Fused Tetrahydroquinolines and

Tetrahydroazepines via Intramolecular Redox Reaction. Chem. Commun. 2010, 46, 6593–

6595.

[69] Ruble, J. C.; Hurd, A. R.; Johnson, T. A.; Sherry, D. A.; Barbachyn, M. R.; Toogood, P.

L.; Bundy, G. L.; Graber, D. R.; Kamilar, G. M. Synthesis of (-)-PNU-286607 by

Asymmetric Cyclization of Alkylidene Barbiturates. J. Am. Chem. Soc. 2009, 131, 3991–

3997.

[70] For another application utilising a HT/cyclisation using benzylic C-H as hydride donors,

see: a) Wölfling, J.; Frank, É.; Schneider, G.; Tietze, L. F. Synthesis of Unusual Bridged

Steroid Alkaloids by an Iminium Ion Induced 1,5-Shift of a Benzylic Hydride. Angew.

Chemie - Int. Ed. 1999, 38, 200–201. b) Wölfling, J.; Frank, É.; Schneider, G.; Tietze, L.

171

F. Synthesis of Novel D-Secoestrone Isoquinuclidines by an Unpredicted Iminium Ion-

Induced 1,5-Hydride Shift. Eur. J. Org. Chem. 2004, 90–100.

[71] Vadola, P. A.; Sames, D. C-H Bond Functionalization via Hydride Transfer: Direct

Coupling of Unactivated Alkynes and Sp3 C-H Bonds Catalyzed by Platinum Tetraiodide.

J. Am. Chem. Soc. 2009, 131, 16525–16528.

[72] Cambeiro, F.; Lõpez, S.; Varela, J. A.; Saá, C. Cyclization by Catalytic Ruthenium Carbene

Insertion into C Sp3 -H Bonds. Angew. Chemie - Int. Ed. 2012, 51, 723–727.

[73] Ono, N. The Nitro Group in Organic Synthesis; Wiley Series in Organic Nitro Chemistry;

John Wiley & Sons, Inc.: New York, USA, 2001.

[74] Manabe, A.; Ohfune, Y.; Shinada, T. Toward the Total Synthesis of Tetrodotoxin:

Stereoselective Construction of the 7-Oxanorbornane Intermediate. Tetrahedron Lett.

2014, 55, 6077–6080.

[75] Marivet, M. C.; Bourguignon, J. Jacques; Lugnier, C.; Mann, A.; Wermuth, C. Georges;

Stoclet, J. Claude. Inhibition of Cyclic Adenosine-3’,5’-Monophosphate Phospho-

diesterase from Vascular Smooth Muscle by Rolipram Analogues. J. Med. Chem. 1989,

32, 1450–1457.

[76] Jensen, K. L.; Poulsen, P. H.; Donslund, B. S.; Morana, F.; Jørgensen, K. A. Asymmetric

Synthesis of γ-Nitroesters by an Organocatalytic One-Pot Strategy. Org. Lett. 2012, 14,

1516–1519.

[77] Anderson, J. C.; Campbell, I. B.; Campos, S.; Shannon, J.; Tocher, D. A. Stereoselective

Synthesis of 1,2-Diamine Containing Indolines by a Conjugate Addition Nitro-Mannich

Reaction. Org. Biomol. Chem. 2015, 13, 170–177.

[78] Mangion, I. K.; Chen, C. Y.; Li, H.; Maligres, P.; Chen, Y.; Christensen, M.; Cohen, R.;

Jeon, I.; Klapars, A.; Krska, S.; et al. Enantioselective Synthesis of an HCV NS5a

Antagonist. Org. Lett. 2014, 16, 2310–2313.

172

[79] Tomás-Gamasa. M., Science of Synthesis: Cross Coupling and HeckType Reactions,

Thieme Chemistry, 2013, 2, 9–42.

[80] Zhang, H.; Cai, Q.; Ma, D. Amino Acid Promoted CuI-Catalyzed C-N Bond Formation

between Aryl Halides and Amines or N-Containing Heterocycles. J. Org. Chem. 2005, 70,

5164–5173.

[81] Hajra, S.; Aziz, S. M.; Maji, R. Organocatalytic Enantioselective Conjugate Addition of

Nitromethane to Alkylidenemalonates: Asymmetric Synthesis of Pyrrolidine-3-Carboxylic

Acid Derivatives. RSC Adv. 2013, 3, 10185–10188.

[82] Palomo, C.; Pazos, R.; Oiarbide, M.; García, J. M. Catalytic Enantioselective Conjugate

Addition of Nitromethane to -Hydroxy Enones as Surrogates of α,β-Unsaturated

Carboxylic Acids and Aldehydes. Adv. Synth. Catal. 2006, 348, 1161–1164.

[83] Evans, D. A.; Seidel, D. Ni(II)-Bis[(R,R)-N,N-Dibenzylcyclohexane-1,2-Diamine]Br2

Catalyzed Enantioselective Michael Additions of 1,3-Dicarbonyl Compounds to

Conjugated Nitroalkenes. J. Am. Chem. Soc. 2005, 127, 9958–9959.

[84] Asher, M.; Kumar, K. V.; Liach, H. Process for making (S)-pregabalin, U. S. Patent WO

2006110783, 19 October 2016.

[85] Liu, Y.; McWhorter, W. W. Synthesis of 8-Desbromohinckdentine A. J. Am. Chem. Soc.

2003, 125, 4240–4252.

[86] Lim, K. H.; Low, Y. Y.; Tan, G. H.; Kam, T. S.; Lim, T. M. Kopsine and Danuphylline

Alkaloids from Kopsia. Biomimetic Partial Synthesis of Danuphylline B. Helv. Chim. Acta

2008, 91, 1559–1566.

[87] Trost, B. M.; Osipov, M.; Krüger, S.; Zhang, Y. A Catalytic Asymmetric Total Synthesis

of (-)-Perophoramidine. Chem. Sci. 2015, 6, 349–353.

[88] More, J. D.; Finney, N. S. A Simple and Advantageous Protocol for the Oxidation of

Alcohols with O-Iodoxybenzoic Acid (IBX). Org. Lett. 2002, 4, 3001–3003.

173

[89] Petasis, N. A.; Morshed, M. M.; Ahmad, M. S.; Hossain, M. M.; Trippier, P. C.

Bis(Cyclopentadienyl)Dimethyltitanium. In Encyclopedia of Reagents for Organic

Synthesis; John Wiley & Sons, Ltd: Chichester, UK, 2012.

[90] Payack, J. F.; Huffman, M. A.; Cai, D.; Hughes, D. L.; Collins, P. C.; Johnson, B. K.;

Cottrell, I. F.; Tuma, L. D. Dimethyltitanocene: From Millimole to Kilomole. Org.

Process Res. Dev. 2004, 8, 256–259.

[91] (1) Arai, S.; Nakajima, M.; Nishida, A. A Concise and Versatile Synthesis of Alkaloids

from Kopsia Tenuis: Total Synthesis of (±)-Lundurine A and B. Angew. Chemie - Int. Ed.

2014, 53, 5569–5572; (2) Lè Ne Lebel, H.; Rie Paquet, V. Multicatalytic Processes Using

Diverse Transition Metals for the Synthesis of Alkenes. J. Am. Chem. Soc 2004, 126,

11152–11153.

[92] a) Affolter, O.; Baro, A.; Frey, W.; Laschat, S. Alkaloids from Alkaloids: Total Synthesis

of (±)-7a-Epi-Hyacinthacine A1 from Z-Protected Tropenone via Baeyer-Villiger

Oxidation. Tetrahedron 2009, 65, 6626–6634. b) Burns, T.; Helliwell, M.; Thomas, E. J.

An Asymmetric Synthesis of the Pentacyclic Core of Stemofoline. Tetrahedron Lett. 2013,

54, 2120–2123.

[93] Nicolaou, K. C.; Adsool, V. A.; Hale, C. R. H. An Expedient Procedure for the Oxidative

Cleavage of Olefinic Bonds with PhI(OAc)2, NMO, and Catalytic OsO4. Org. Lett. 2010,

12, 1552–1555.

[94] Selected references on metal catalysed amide reduction:

Rh (3o amide): a) Kuwano, R.; Takahashi, M.; Ito, Y. Reduction of Amides to Amines via

Catalytic Hydrosilylation by a Rhodium Complex. Tetrahedron Lett. 1998, 39, 1017–1020.

Ir (2o amide): b) Cheng, C.; Brookhart, M. Iridium-Catalyzed Reduction of Secondary

Amides to Secondary Amines and Imines by Diethylsilane. J. Am. Chem. Soc. 2012, 134,

174

11304–11307. For an application in total synthesis: c) He, C.; Stratton, T. P.; Baran, P. S.

Concise Total Synthesis of Herqulines B and C. J. Am. Chem. Soc. 2019, 141, 29–32.

Ru (2o and 3o amide): d) Cabrero-Antonino, J. R.; Alberico, E.; Junge, K.; Junge, H.;

Beller, M. Towards a General Ruthenium-Catalyzed Hydrogenation of Secondary and

Tertiary Amides to Amines. Chem. Sci. 2016, 7, 3432–3442.

Fe (3o amide): e) Zhou, S.; Junge, K.; Addis, D.; Das, S.; Beller, M. A Convenient and

General Iron-Catalyzed Reduction of Amides to Amines. Angew. Chemie - Int. Ed. 2009,

48, 9507–9510. f) Sunada, Y.; Kawakami, H.; Imaoka, T.; Motoyama, Y.; Nagashima, H.

Hydrosilane Reduction of Tertiary Carboxamides by Iron Carbonyl Catalysts. Angew.

Chemie - Int. Ed. 2009, 48, 9511–9514.

Cu (2o amide): g) Das, S.; Join, B.; Junge, K.; Beller, M. A General and Selective Copper-

Catalyzed Reduction of Secondary Amides. Chem. Commun. 2012, 48, 2683–2685.

Zn (3o amide): h) Das, S.; Addis, D.; Zhou, S.; Junge, K.; Beller, M. Zinc-Catalyzed

Reduction of Amides: Unprecedented Selectivity and Functional Group Tolerance. J. Am.

Chem. Soc. 2010, 132, 1770–1771.

Pt (3o amide): i) Nakatani, N.; Hasegawa, J. Y.; Sunada, Y.; Nagashima, H. Platinum-

Catalyzed Reduction of Amides with Hydrosilanes Bearing Dual Si-H Groups: A

Theoretical Study of the Reaction Mechanism. Dalt. Trans. 2015, 44, 19344–19356.

Mo (3o amide): j) Tinnis, F.; Volkov, A.; Slagbrand, T.; Adolfsson, H. Chemoselective

Reduction of Tertiary Amides under Thermal Control: Formation of Either Aldehydes or

Amines. Angew. Chemie - Int. Ed. 2016, 55, 4562–4566.

Ni (3o amide): k) Simmons, B. J.; Hoffmann, M.; Hwang, J.; Jackl, M. K.; Garg, N. K.

Nickel-Catalyzed Reduction of Secondary and Tertiary Amides. Org. Lett. 2017, 19, 1910–

1913.

175

[95] For examples of malonates reduction using NaBH4, see: a) Chen, W.; Xia, Y.; Lin, L.;

Yuan, X.; Guo, S.; Liu, X.; Feng, X. Asymmetric Synthesis of Furo[3,4-b]Indoles by

Catalytic [3+2] Cycloaddition of Indoles with Epoxides. Chem. - A Eur. J. 2015, 21,

15104–15107. b) Mahadevegowda, S. H.; Khan, F. A. Synthesis of the Tetrahydrofuran

Unit of Varitriol and γ- Butyrolactones from 5-Oxabicyclo[2.1.1]Hexane Derivative via

Oxidative Cleavage Reactions. Tetrahedron Lett. 2014, 55, 2266–2269.

[96] Xu, Y.; Wei, Y. CeCl 3 -Catalyzed Reduction of Methyl Esters of Carboxylic Acids to

Corresponding Alcohols with Sodium Borohydride. Synth. Commun. 2010, 40, 3423–3429.

[97] Pouliot, M. F.; Mahé, O.; Hamel, J. D.; Desroches, J.; Paquin, J.-F. F. Halogenation of

Primary Alcohols Using a Tetraethylammonium Halide/[Et2NSF2]BF4 Combination. Org.

Lett. 2012, 14, 5428–5431.

[98] Yang, Q.; Xiao, W. J.; Yu, Z. Lewis Acid Assisted Ring-Closing Metathesis of Chiral

Diallylamines: An Efficient Approach to Enantiopure Pyrrolidine Derivatives. Org. Lett.

2005, 7, 871–874.

[99] Enquist, J. A.; Stoltz, B. M. The Total Synthesis of (-)-Cyanthiwigin F by Means of Double

Catalytic Enantioselective Alkylation. Nature 2008, 453, 1228–1231.

[100] Uno, B. E.; Gillis, E. P.; Burke, M. D. Vinyl MIDA Boronate: A Readily Accessible and

Highly Versatile Building Block for Small Molecule Synthesis. Tetrahedron 2008, 65,

3130–3138.

[101] Kang, T.; White, K. L.; Mann, T. J.; Hoveyda, A. H.; Movassaghi, M. Enantioselective

Total Synthesis of (−)-Deoxoapodine. Angew. Chemie - Int. Ed. 2017, 56, 13857–13860.

[102] Examples of hydroboration/oxidation on terminal alkene in the presence of internal alkene,

see: a) Kim, J. H.; Kim, H. O.; Lee, K. M.; Chun, M. W.; Moon, H. R.; Jeong, L. S.

Asymmetric Synthesis of Homo-Apioneplanocin A from d-Ribose. Tetrahedron 2006, 62,

6339–6342; b) Jiao, L.; Lin, M.; Zhuo, L.-G.; Yu, Z.-X. Rh(I)-Catalyzed [(3+2)+1]

Cycloaddition of 1-Yne/Ene-Vinylcyclopropanes and CO: Homologous Pauson−Khand

Reaction and Total Synthesis of (±)-α-Agarofuran. Org. Lett. 2010, 12, 2528–2531. 176

Appendix Contents

Appendix ...... 1 Appendix 1: Experimental ...... 4 General Information ...... 4 Compound 307 ...... 5

Compound 282 ...... 6

Compound 309 ...... 7

Compound 310 ...... 8

Compound 319 ...... 9

Compound 320 ...... 10

Compound 321 ...... 11

Compound 322 ...... 12

Compound (+/-)-321 ...... 13

Compound (+/-)- 329 ...... 14

Compound (+/-) 329 ...... 14

Compound (+/-)-330 ...... 15

Compound (+/-)-328 and (+/-)-344 ...... 18

Compound (+/-)-348 ...... 20

Compound (+/-)-358 ...... 22

Compound (+/-)-359 and (+/-)-362 ...... 23

Compound (+/-)-382 ...... 25

Studies on (+)-vallesamidine and (+)-strempeliopine ...... 27 Compound (+/-)-468 ...... 27

Compound (+/-)-473 ...... 28

Compound (+/-)-477 ...... 29

Compound (+/-)-466 ...... 31

Compound (+/-)-471 ...... 32

Compound 395 ...... 33

Compound 394 ...... 34

Compound 470 ...... 35

Compound 472 ...... 36

Compound 478 ...... 38

Compound 479 ...... 39

Compound 499 ...... 40

Compound 501 ...... 41

Compound 502 ...... 42

Compound 503 ...... 44

Compound 526 ...... 45

Compound 530 ...... 46

Compound 531 ...... 48

Compound 533 ...... 49

Compound 534 ...... 50

Compound 535 ...... 51

Compound 536 ...... 52

Compound 537 ...... 53

Compound 538 ...... 55

Compound 539 ...... 56

Compound 540 ...... 57

Compound 541 ...... 59

Compound 542 ...... 60

Compound 387 ...... 62

(+)-Vallesamidine (102) ...... 63

Compound 499 (by triphosgene/MeOH) ...... 66

Compound 557 ...... 66

Compound 558 ...... 68

Compound 530 (by lactam reduction) ...... 69

Compound 526 (by Swern oxidation) ...... 70

Compound 559 ...... 70

Compound 560 ...... 72

Compound 526 (by acid hydrolysis) ...... 73

Compound 569 ...... 74

Compound 572 ...... 76

Compound 574 ...... 78

Compound 581 ...... 79

Compound 582 ...... 80

Compound 587 ...... 82

Compound 588 ...... 83

Compound 590 ...... 84

Compound 594 ...... 86

Compound 595 ...... 88

Compound 604 ...... 89

Compound 387 ...... 90

Compound 543 ...... 91

Appendix 2: Chiral HPLC analysis and X-ray diffraction data ...... 94 a. Chiral HPLC analysis on compound 394 ...... 94 b. X-ray diffraction data of single crystal on compound 499 ...... 96

Appendix 1: Experimental

General Information

Generally, glassware was flam-dried before use. All solvents and chemicals were used as received unless stated. The anhydrous solvents diethyl ether (Et2O), tetrahydrofuran

(THF), dichloromethane (DCM), toluene and hexane were obtained from solvent purification system. Thin layer chromatography (TLC) was performed using Merck silica-aluminium plates and visualised by UV light (254 nm) and potassium permanganate or anisaldehyde stains. For column chromatography, Merck Geduran® Si

60 silica gel was used.

All 1H NMR and 13C NMR data were recorded using Bruker AVANCE III 400

MHz, Bruker AVANCE III 600 MHz and Bruker AVANCE NEO 700 MHz machines at

400, 600 and 700 MHz for 1H NMR and 100, 125, 176 MHz for 13C NMR respectively.

Samples were prepared as dilute solutions of CDCl3, DMSO-d6 or MeOD-d4 and spectra were recorded at 298K, unless otherwise stated. Reference values for residual solvents

1 were taken as δ = 7.26 (CDCl3), 2.51 (DMSO-d6), 3.30 (MeOD-d4) ppm for H NMR and

13 δ = 77.2 (CDCl3), 39.5 (DMSO-d6), 49.0 (MeOD-d4) ppm for C NMR. Coupling constants (J) are given in Hz and are uncorrected and multiplicities for coupled signals were denoted as: s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet, br. = broad, apt. = apparent and dd = double doublet etc. COSY and DEPT experiments were carried out to aid assignment where appropriate.

Mass spectroscopy data was collected on a Thermo Finnigan Mat900xp (EI/CI),

Waters LCT Premier XE (ES) and Agilent 6510 Q TOF mass spectrometers (ESI).

Infrared data were collected using Bruker compact FTIR spectrometer. Melting points were uncorrected and recorded on a DigMelt MPA160-SRS machine. Optical rotations

were obtained using a Bellingham+Stanley ADP430 series polarimeter. Chiral HPLC was performed using a Chiralcel OD-H 15 cm analytical column.

Compound 307

To a stirred and cooled (0 oC) soulution of trimethyl phosphonoacetate (3.00 mL, 18.7 mmol) in THF (100 mL) was added potassium tert-butoxide (2.10 g, 18.7 mmol) portionwise and the resulting mixture was stirred at room temperature for 1 hour. Then a solution of 283 (4.28 g, 18.7 mmol) in THF (50.0 mL) was added through an addition funnel dropwise and the stirring was continued overnight. The reaction mixture was poured into a sat. NH4Cl aqueous solution (100 mL) and then extracted with CH2Cl2 (100 mL x 2). The combined organic layers were washed with brine (100 mL), dried (MgSO4) and concentrated by rotary evaporation. The resulting solid was further recrystalised in

EtOAc to give 307 (3.75 g, 75%) as a white solid.

1 Rf (10% EtOAc in hexane) 0.3; H NMR (500 MHz, CDCl3)  7.99 (1H, d, J = 16.0 Hz,

Ar-CH=CHCO2Me), 7.06 (1H, s, ArH), 7.05 (1H, s, ArH), 6.24 (1H, d, iJ = 16.0 Hz,

13 ArCH=CHCO2Me), 6.02 (2H, s, OCH2O), 3.81 (3H, s, CO2CH3); C NMR (151 MHz,

CDCl3)  167.1 (CO2Me), 150.2 (ArC), 148.0 (ArC), 143.1 (CH=CHCO2Me), 127.8

(ArC), 118.7 (CH=CHCO2Me), 117.9 (ArC), 113.3 (ArCH), 106.6 (ArCH), 102.4

1 (OCH2O), 51.9 (CO2CH3). Data was consistent with Lit.

Compound 282

o To a cooled (0 C) and stirred solution of ester 307 (3.00 g, 10.5 mmol) in CH2Cl2 (50.0 mL) under N2 atmosphere was added DIBAL-H (1.0 M in PhMe, 32.0 mL, 32.0 mmol) dropwise over 10 min and the resulting solution was stirred at room temperature for 2 h.

The reaction solution was cooled to 0 oC and Rochelle’s solution (200 mL) was added and the whole mixture was stirred until two layers were separated clearly. The mixture was extracted with CH2Cl2 (100 mL x 3) and the combined organic layers were washed with brine (200 mL), dried (MgSO4) and concentrated by rotary evaporation. The crude allyl alcohol was dissolved in CH2Cl2 (50.0 mL), and MnO2 (13.0 g, 150 mmol) was added. The mixture was stirred at room temperature overnight and then filtered through a pad of celite. The filtrate was concentrated by rotary evaporation to give a pale yellow solid and recrystallisation using EtOAc gave compound 282 (2.40 g, 90%) as a white solid.

1 Rf (30% EtOAc in hexane) 0.37; HNMR (500 MHz, CDCl3)  9.72 (1H, d, J = 7.8 Hz,

CHO), 7.82 (1H, d, J = 16.2 Hz, CH=CHCHO), 7.26 (1H, s, ArH), 7.11 (1H, s, ArH),

13 6.53 (1H, dd, J = 16.2, 7.8 Hz, CH=CHCHO), 6.06 (2H, s, OCH2O); C NMR (125 MHz,

CDCl3)  193.5 (CHO), 150.9 (ArC), 150.5 (CH=CHCHO), 148.2 (ArC), 129.0

(CH=CHCHO), 127.2 (ArC), 118.8 (ArC), 113.4 (ArCH), 106.6 (ArCH), 102.6

1 (OCH2O); Data was consistent with Lit.

Compound 309

To a stirred suspension of aldehyde 282 (2.00 g, 7.84 mmol), catalyst 308 (0.25 g, 0.784 mmol) and benzoic acid (0.20 g, 1.57 mmol) in methanol (18.0 mL), nitromethane (1.54 mL, 23.5 mmol) was added and the resulting mixture was stirred at room temperature for

24 h, during which time the reaction mixture became clear. Saturated NaHCO3 solution

(30 mL) was added to quench the reaction and the mixture was extracted with EtOAc (50 mL, three times). The combined organic layer was washed with brine, dried (MgSO4) and concentrated. Flash chromatography of the residue on silica gel, using 7:3 hexane :

EtOAc, gave the desired -nitroaldehyde 309 as a yellow sticky oil (2.0 g, 80%).

1 Rf (30% EtOAc in hexane) 0.32; H NMR (500 MHz, CDCl3)  9.72 (1H, s, CHO), 7.04

(1H, s, ArCH), 6.67 (1H, s, ArCH), 5.98 (2H, m, OCH2O), 4.71—4.63 (2H, m, CH2NO2),

4.51—4.47 (1H, m, CHCH2NO2), 3.04-2.99 (1H, ddd, J = 18.5, 7.5, 1.0 Hz, CH2CHO),

13 2.96-2.90 (1H, ddd, J = 18.0, 7.0, 1.5 Hz, CH2CHO); C NMR (151 MHz, CDCl3) δ

198.7 (CHO), 148.2 (ArC), 148.2 (ArC), 129.8 (ArC), 114.9 (ArC), 113.6 (ArCH), 107.7

(ArCH), 102.3 (OCH2O), 77.8 (CH2NO2), 45.6 (CH2CHO), 37.0 (CHCH2NO2). Data was consistent with Lit.1

Compound 310

To a cooled (-78 °C) and stirred solution of 309 (1.70 g, 5.38 mmol) in CH2Cl2 (30.0 mL) in N2 atmosphere, (1,2-bistrimethylsiloxy)-ethane (1.32 mL, 5.38 mmol) was added, followed by addition of TMSOTf (0.10 mL, 5.38 mmol). The resulting solution was stirred at -78 °C for 1 h then warmed to room temperature and the stirring was continued overnight. Reaction mixture was diluted with CH2Cl2 (30.0 mL) and quenched with sat.

K2CO3 aqueous solution (30 mL). The organic layer was washed with sat. NaHCO3 aqueous solution and brine, dried (MgSO4) and concentrated by rotary evaporation to give the titled product 310 as a yellow oil (1.70 g, 90%).

1 Rf = (30% EtOAc in Hexane) 0.32; H NMR (600 MHz, CDCl3)  7.27 (1H, s, ArCH),

6.71 (1H, s, ArCH), 5.98-5.97 (2H, m, OCH2O), 4.83 (1H, dd, J = 3.6, 5.4 Hz, OCHO),

4.75-4.71 (1H, dd, J = 6.6, 13.2 Hz, CHHNO2), 4.67-4.63 (1H, m, CHHNO2), 4.264.24

(1H, m, CHCH2NO2), 4.00-3.95 (2H, m, OCH2CH2O), 3.86-3.80 (2H, m, OCH2CH2O),

13 2.12-2.00 (2H, m, CHCH2CH); C NMR (151 MHz, CDCl3) δ 147.9 (ArC), 147.8 (ArC),

131.3 (ArC), 113.4 (ArCH), 102.4 (ArCH), 102.1 (OCH2O), 78.9 (CH2NO2), 65.1

1 (OCH2CH2O), 65.0 (OCH2CH2O), 36.4 (O,OCHCH2). Data was consistent with Lit.

Compound 319

To a stirred and cooled (0 °C) solution of trimethylphosphonoacetate (5.40 mL, 33.3 mmol) in THF (100 mL), t-BuOK (3.90 g, 35.0 mmol) was added portion-wise and the resulting mixture was stirred at room temperature for 1 h. Then a solution of piperonal

305 (5.00 g, 33.3 mmol) in THF (80.0 mL) was added through an addition funnel dropwise and the stirring was continued overnight. Half of the solvent was evaporated, and the reaction mixture was poured into a saturated NH4Cl aqueous solution (150 mL) and then extracted with CH2Cl2 (2 x 80.0 mL). The combined organic layers were washed with Brine (100 mL), dried (MgSO4) and concentrated. The resulting white solid was further washed with 20% Et2O in Hexane to give 319 (6.40 g, 93%) as a white powder.

1 Rf (30% EtOAc in Hexane) 0.43; H NMR (500 MHz, CDCl3)  7.59 (1H, d, J = 15.5

Hz, HC=CHCO2Me), 7.03- 6.99 (2H, m, ArCH), 6.81 (1H, d, J = 8 Hz, ArCH), 6.26 (1H,

13 d, J = 16 Hz, CH=CHCO2Me), 6.01 (2H, s, OCH2O), 3.79 (3H, s, CO2CH3); C NMR

(126 MHz, CDCl3) δ 167.7 (CH=CHCO2Me), 149.7 (ArC), 148.4 (ArC), 144.6

(CH=CHCO2Me), 128.9 (ArC), 124.5 (ArCH), 115.8 (CH=CHCO2Me), 108.6 (ArCH),

2 106.6 (ArCH), 101.6 (OCH2O), 51.7 (CO2CH3). Data was consistent with Lit.

Compound 320

Procedure: To a cooled (0 °C) and stirred solution of ester 319 (6.00 g, 29.0 mmol) in

CH2Cl2 (100 mL) under N2 atmosphere, DIBAL-H (1.0 M in PhMe, 87.0 mL, 87.0 mmol) was added dropwise over 10 min and the resulting solution was stirred for 2 h at room temperature. The reaction solution was cooled back to 0 °C and quenched with EtOAc

(20.0 mL). 10 min later, the whole mixture was poured into a cooled (0 °C) 1M HCl solution (200 mL) and extracted with CH2Cl2 (2 x 150 mL). The combined organic layers were washed with brine (200 mL), dried (MgSO4) and concentrated to give the crude allyl alcohol. This crude alcohol was re-dissolved in CH2Cl2 (50.0 mL) and MnO2 (36.0g, 290 mmol) and the resulting mixture was stirred overnight. MnO2 was filtered through a pad of celite and the yellowish solution was concentrated to give a pale-yellow solid.

Recrystallization using EtOAc afforded the desired aldehyde 320 (4.4 g, 86% over two steps) as a white powder.

1 Rf (30% EtOAc in Hexane) 0.34; H NMR (500 MHz, CDCl3)  9.63 (1H, d, J = 8 Hz,

CHO), 7.36 (1H, d, J = 15.5 Hz, CH=CHCHO), 7.06-7.04 (2H, m, ArCH), 6.84 (1H, d,

13 J = 8 Hz, ArCH), 6.54 (1H, dd, J = 8, 15.5 Hz, CH=CHCHO), 6.03 (2H, s, OCH2O); C

NMR (126 MHz, CDCl3) δ 193.6 (CHO), 152.7 (CH=CHCHO), 150.6 (ArC), 148.6

(ArC), 128.6 (ArC), 126.9 (CH=CHCHO), 125.4 (ArCH), 108.8 (ArCH), 106.8 (ArCH),

3 101.9 (OCH2O). Data was consistent with Lit.

S10

Compound 321

Procedure: To a stirred suspension of enal 320 (4.00 g, 22.7 mmol), catalyst 308 (0.70 g, 2.27 mmol) and benzoic acid (0.55 g, 4.54 mmol) in methanol (35.0 mL), nitromethane

(3.00 mL, 68.1 mmol) was added and the resulting mixture was stirred at room temperature for 24 h, during which time the reaction mixture became clear. Saturated

NaHCO3 solution (30.0 mL) was added to quench the reaction and the mixture was extracted with EtOAc (3 x 30.0 mL). The combined organic layer was washed with brine, dried (MgSO4) and concentrated. Flash chromatography of the residue on silica gel, using

30% EtOAc in hexane, gave -nitroaldehyde 321 as a yellow oil (3.70 g, 69%).

+. Rf (30% EtOAc in Hexane) 0.16; HRMS (EI, m/z) calcd for C11H11NO5 [M] 237.0632,

1 found 237.0632. H NMR (600 MHz, CDCl3)  9.70 (1H, s, CHO), 7.27—6.68 (3H, m,

ArCH), 5.96 (2H, s, OCH2O), 4.65—4.62 (1H, dd, J = 6.6, 12 Hz, CHHNO2), 4.57—4.54

(1H, dd, J = 7.8, 12 Hz, CHHNO2), 4.01—3.99 (1H, m, CHCH2NO2), 2.91—2.89 (2H,

13 m, CH2CHO); C NMR (151 MHz, CDCl3) δ 198.9 (CHO), 148.4 (ArC), 147.5 (ArC),

131.8 (ArC), 120.9 (ArCH), 108.9 (ArCH), 107.7 (ArCH), 101.4 (OCH2O), 79.7

(CH2NO2), 46.6 (CH2CHO), 37.9 (CHCH2CHO); The enantioselectivity was not available at this stage and its dimethyl acetal form (see preparation of compound 329) was more stable and suitable for HPLC experiment.

S11

Compound 322

To a cooled (-78 °C) and stirred solution of aldehyde 321 (3.20 g, 11.8 mmol) in CH2Cl2

(40.0 mL) in N2 atmosphere, (1,2-bistrimethylsiloxy)-ethane (2.90 mL, 11.8 mmol) was added, followed by addition of TMSOTf (0.20 mL, 1.18 mmol). The resulting solution was stirred at -78 °C for 1 h then warmed to room temperature and the stirring was continued overnight. Reaction mixture was quenched with sat. K2CO3 aqueous solution

(40 mL) and extracted with CH2Cl2 (40 mL). The combined organic layers were washed with brine, dried (MgSO4) and concentrated. Flash chromatography of the residue on silica, using 30% EtOAc in pet. ether, to give desired product 322 as a yellow oil (3.17 g,

96%).

-1 Rf (3:7 EtOAc : Hexane) 0.18; FTIR (neat, cm ) 2897, 1742, 1551, 1490, 1372, 1246,

+. 1 1131; HRMS (EI, m/z) calcd for C13H15NO6 [M] 281.0894, found 281.0894. H NMR

(600 MHz, CDCl3)  6.75—6.67 (3H, m, ArCH), 5.94 (2H, s, OCH2O), 4.74 (1H, dd, J

= 3.6, 7.8 Hz, CHHNO2), 4.73—4.69 (1H, dd, J = 7.8, 15 Hz, CHHNO2), 4.54—4.50

(1H, dd, J = 10.8, 15 Hz, OCHO), 3.99—3.93 (2H, m, OCH2CH2O), 3.85—3.78 (2H, m,

OCH2CH2O), 3.69—3.66 (1H, m, CHCH2NO2), 2.09—2.04 (1H, m, CHCH2CH), 1.95—

13 1.89 (1H, m, CHCH2CH); C NMR (151 MHz, CDCl3) δ 148.2 (ArC), 147.2 (ArC),

133.0 (ArC), 120.9 (ArCH), 108.7 (ArCH), 107.7 (ArCH), 102.3 (OCHO), 101.28

(OCH2O), 80.7 (CH2NO2), 65.1 (OCH2CH2O), 65.0 (OCH2CH2O), 39.9 (CHCH2CH),

37.5 (CHCH2NO2);

S12

Compound (+/-)-321

To a stirred solution of 319 (343 mg, 1.66 mmol) in nitromethane (5.00 mL) was added tetramethylguanidine (0.05 mL, 0.42 mmol) and the resulting mixture was heated to 70 oC and stirred for 24 h. Reaction mixture was cooled to room temperature and quenched with aq. HCl solution (1.0 M, 15.0 mL). The mixture was extracted with EtOAc (20.0 mL x 2) and the combined organic layers were dried (Na2SO4) and concentrated by rotary evaporation. Chromatogrphy of the residue on silica gel, using 30% EtOAc in hexane, gave compound 323 as a yellow oil (408 mg, 92%).

-1 Rf (30% EtOAc in hexane): 0.63; FTIR (neat, cm ) 2950, 1728, 1547, 1501, 1486, 1241,

.+ 1 1169, 1035; HRMS (EI) calcd. for C12H13NO6 [M] 267.0737, found 267.0738; H NMR

(400 MHz, CDCl3) d 6.76—6.67 (3H, m, ArH), 5.95 (2H, s, OCH2O), 3.91 (1H, m, Ar-

13 CH), 3.65 (3H, s, OCH3), 2.76—2.67 (2H, m, CH2CO2Me); C NMR (150 MHz, CDCl3) d 171.1 (CO2Me), 148.2 (ArC), 147.4 (ArC), 131.9 (ArC), 120.8 (ArCH), 108.8 (ArCH),

107.6 (ArCH), 101.4 (OCH2O), 79.7 (CH2NO2), 52.1 (OCH3), 40.1 (Ar-CH), 37.8

(CH2CO2Me).

o To a cooled (-78 C) and stirred solution of 323 (2.50 g, 9.36 mmol) in dry CH2Cl2 (70.0 mL) was added DIBAL-H (1.0 M in hexane, 11.3 mmol) via a syringe pump over 30 min.

The resulting solution was stirred at -78 oC for 2 h. The reaction solution was quenched with methanol (5.00 mL) and diluted with CH2Cl2 (100 mL). The whole mixture was poured into Rochelle’s salt solution (200 mL) and stirred for another 2 h. The organic layer was separated, and aqueous layer was xtracted with CH2Cl2 (100 mL x 2). The

S13

combined organic layer was dried (Na2SO4) and concentrated by rotary evaporation.

Chromatography of the residue on silica gel, using 30% EtOAc in hexane, gave compound 321 as a light-yellow oil (1.41 g, 63%, 80% brsm)

Compound (+/-)- 329

To a solution of aldehyde (+/-)-321 (11.10 g, 4.64 mmol) in methanol (15.0 mL) and

. CH2Cl2 (15.0 mL) was added TsOH H2O (88.0 mg, 0.464 mmol) and trimethylorthoformate (1.60 mL, 13.9 mmol) and the resulting solution was stirred at room temperature until the full consumption of starting material. The reaction solution was diluted with CH2Cl2 (50.0 mL) and washed with sat. NaHCO3 solution (50 mL). The organic layer was dried (Na2SO4) and concentrated by rotary evaporation to give the compound (+/-)-329 as a light-yellow oil (1.30 g, 100%).

Compound (+/-) 329

A solution of (+/-)-322 (1.20 g, 4.27 mmol), CSA (0.50 g, 2.14 mmol) and trimethylorthoformate (1.00 mL, 8.54 mmol) in methanol (10.0 mL) was refluxed for 2.5

S14

h. The reaction mixture was cooled to room temperature and concentrated. The residue was re-dissolved in CH2Cl2 (20.0 mL) and washed with sat. NaHCO3 aqueous solution

(20.0 mL). The organic layer was separated, dried (MgSO4) and concentrated. Flash chromatography of the residue on silica gel, using 1:2 acetone/Hexane, gave desired product (+/-)-329 as a yellow oil (1.14 g, 94%).

-1 Rf (20% EtOAc in hexnae) 0.34; FTIR (neat, cm ) 2937, 2834, 1743, 1550, 1443, 1247

-1 + cm ; HRMS (ESI) calcd for C13H18NO6 [M+H] 483.1050, found 283.1051.

Enantiomeric ratio (e.r.): 97.6:2.4 (Chiralcel OD-H column, 90:10 hexane:isopropanol,

1 1mL/min, 254 nm, major enantiomer tr 31.4 min, minor enantiomer tr 22.1 min); H NMR

(500 MHz, CDCl3)  6.77—6.66 (3H, m, ArCH), 5.95 (2H, s, OCH2O), 4.62—4.58 (1H, dd, J = 7, 12.5 Hz, CHHNO2), 4.51—4.47 (1H, dd, J = 8.5, 12.5 Hz, CHHNO2), 4.16

(1H, dd, J = 3.5, 7.5 Hz, OCHO), 3.58—3.54 (1H, m, CHCH2NO2), 3.29 (3H, s, OCH3),

3.25 (3H, s, OCH3), 2.01—1.96 (1H, m, CHCH2CH), 1.88—1.83 (1H, m, CHCH2CH);

13 C NMR (126 MHz, CDCl3) δ 148.2 (ArC), 147.2 (ArC), 132.7 (ArC), 120.9 (ArCH),

108.8 (ArCH), 107.6 (ArCH), 102.1 (MeOCHOMe), 101.3 (OCH2O), 80.6 (CH2NO2),

53.2 (OCH3), 53.0 (OCH3), 40.2 (CHCH2CH), 36.2 (CHCH2NO2);

Compound (+/-)-330

Method 1:

S15

To a cooled (-40 °C, dry ice/MeCN) and stirred solution of nitroalkane (+/-)-329 (140 mg, 0.50 mmol) in dry THF (4.00 mL), n-BuLi (1.60 M in hexane, 0.33 mL) was added dropwise. The cooling bath was left but not recharged and the stirring was stirred for 50 min. The resulting nitronate solution was then cooled to -78 °C and a solution of imine

312 (154 mg, 0.74 mmol) in THF (1.00 mL) was added dropwise. The mixture was stirred at -78 °C for 30 min and TFA (60.0 L, 0.74 mmol) was added dropwise. The resulting mixture was stirred at -78 °C for 1 h and slowly warmed to room temperature. NaHCO3 saturated solution (5.00 mL) was added to quench the reaction and the mixture was extracted with EtOAc (3 x 5.00 mL). The combined organic layers were dried (MgSO4) and concentrated. The flash chromatography of the residue on silica, using 1:4 EtOAc : hexane) gave the desired product (+/-)-330 as a pale yellow oil (140 mg, 58% yield).

-1 Rf (20% EtOAc in hexane) 0.26 and 0.33 (two isomers); FTIR (neat, cm ) 3370 (br),

+ 2904, 2833, 1738, 1550, 1510, 1243, 1036; HRMS (ESI) calcd for C24H31N2O9 [M+H]

1 491.2029, found 491.2031. Major isomer H NMR (600 MHz, CDCl3)  6.88—6.82

(3H, m, ArCH), 6.72—6.70 (2H, m, PMPCH), 6.30—6.29 (2H, m, PMPCH), 6.03 (2H, dd, J = 6.8, 1.5Hz, OCH2O), 4.98 (1H, dd, J =11.5, 2.7 Hz, CHNO2), 4.39—4.33 (1H, m,

CO2CH2CH3), 4.28—4.21 (2H, m, CO2CH2CH3 and NH), 4.15—4.10 (1H, m, OCHO),

4.01 (1H, dd, J = 8.8, 3.0 Hz, HNCHCO2Et), 3.83—3.77 (1H, m, CHCHNO2), 3.74 (3H, s, OCH3), 3.22 (3H, s, OCH3), 3.18 (3H, s, OCH3), 1.93—1.88 (1H, m, CHCH2CH),

13 1.82—1.76 (1H, m, CHCH2CH), 1.41 (3H, t, J = 7.1 Hz, CO2CH2CH3); C NMR (150

MHz, CDCl3)  168.7 (CO2Et), 153.5 (ArC), 148.4 (ArC), 147.6 (ArC), 138.6 (ArC),

131.1 (ArC), 122.1 (ArC), 115.9 (PMPCH), 115.0 (PMPCH), 109.2 (ArCH), 108.9

(ArCH), 101.5 (MeOCHOMe), 101.2 (OCH2O), 93.8 (CHNO2), 62.6 (CO2CH2CH3),

58.3 (HNCHCO2Et), 55.8 (OCH3), 53.1 (ArOCH3), 51.4 (OCH3), 41.3 (CHCHNO2), 35.4

(CHCH2CH), 14.2 (CO2CH2CH3);

S16

Methods 2:

To a cooled (-40 °C, dry ice/MeCN) and stirred solution of nitroalkane (+/-)-329 (181 mg, 0.64 mmol) in dry THF (4.00 mL) in N2 atmosphere, n-BuLi (1.6 M in hexane, 0.43 mL) was added dropwise. The cooling bath was left but not recharged and the stirring was stirred for 50 min and then cooled back to -78 °C (dry ice/acetone). In another N2 protected flask, a suspension of Cu(OTf)2 (350 mg, 0.96 mmol) in THF (2.00 mL) at -

40°C was added a solution of imine 312 (200 mg, 0.96 mmol) in THF (1 mL) dropwise and the resulting dark brown solution was transferred dropwise to the nitronate solution.

The dark green solution was stirred at -78 °C for 1 h and quenched with EtOH (5.00 mL) at same temperature. Solvent was then evaporated and the residue was directly put on column and flash chromatography on silica gel, using 1:4 EtOAc : Hexane, gave the desired product (+/-)-330 as a yellow oil (198 mg, 63%).

Method 3:

To a cooled (-40 oC, dry ice/MeCN) and stirred solution of (+/-)-329 (1.34 g, 4.77 mmol) in dry THF (30.0 mL) under N2 atmosphere was added n-BuLi (1.60 M, in hexane, 3.00 mL) dropwise. Cooling bath was left but not recharged and the stirring was continued for 50 min and then cooled back to -78 oC. The imine 312 (1.48 g, 7.16 mmol) in THF

(5.00 mL) was added dropwise and stirred for 10 min. To another N2 filled round bottom

o flask charged with THF (5.00 mL) at -40 C was added ZnEt2 (1.0 M in hexane, 7.20 mL) and then trifluoroacetic acid (1.00 mL) dropwise and resulting solution was stirred for 10 min. This Zn(TFA)2 solution was then transferred to the nitronate solution over 15 min at -78 oC via a syringe pump and the mixture was stirred at -78oC for 1 h and slowly warmed to 0 oC, during which time the starting material 329 was fully consumed. The reaction mixture was quenched with trifluoroacetic acid (1.00 mL), diluted with EtOAc

(30 mL) and washed with sat. NaHCO3 aq. (30 mL x 2). Aqueous layer was extracted

S17

with EtOAc (30 mL x 2) and combined organic layers were dried (NasSO4) and concentrated. Chromatography of the residue on silica gel, using 20% EtOAc in hexane, gave (+/-)-330 as a yellow oil (1.65 g, 70%).

Compound (+/-)-328 and (+/-)-344

To a cooled (-40 °C) and stirred solution of nitroamine (+/-)-330 (58.0 mg, 0.12 mmol) in CH2Cl2 (1.20 mL) in N2 atmosphere, BF3·OEt2 (47.0 L, 0.18 mmol) was added, followed by the addition of Et3SiH (38.0 mL, 0.24 mmol). Cooling bath was left but not recharged and the stirring was continued for 6 h. The reaction mixture was diluted with

CH2CH2 (5.00 mL) and quenched with sat. NaHCO3 aqueous solution (5.00 mL).

Separated organic layer was then dried (MgSO4) and concentrated. Flash chromatography of the residue on silica gel, using 20% EtOAc in hexane, gave (+/-)-344

(10.0 mg, 19%) as a yellow oil and (+/-)-328 (10.0 mg, 18%) as yellow oil.

Compound (+/-)-328:

-1 Rf (30% EtOAc in hexane): 0.38; FTIR (neat, cm ) 2904, 1731, 1551, 1511, 1487, 1238,

+ 1 1037; HRMS (ESI, m/z) calcd. for C22H25N2O [M+H] 429.1662, found 429.1660; H

NMR (600 MHz, CDCl3)  6.90—6.76 (7H, m, ArCH), 5.96 (2H, s, OCH2O), 5.37—5.35

(1H, m, CHNO2), 5.11 (1H, brs, PMPNCHCO2Et), 4.22 (2H, q, J = 7.2 Hz, CO2CH2CH3),

3.76 (3H, s, OCH3), 3.61—3.57 (1H, m, PMPNCH2), 3.45 (1H, dt, J = 12.0, 3.2 Hz,

PMPNCH2), 3.14 (1H, td, J = 13.2, 4.0 Hz, CH2CHCHNO2), 2.74 (1H, dq, J = 12.8, 5.2

S18

Hz, NCH2CH2), 1.91 (1H, d, J = 13.1 Hz, NCH2CH2), 1.27 (3H, t, J = 7.2 Hz,

13 CO2CH2CH3); C NMR (151 MHz, CDCl3) δ 169.2 (CO2Et), 154.3 (ArC), 148.0 (ArC),

146.8 (ArC), 144.2 (ArC), 133.4 (ArC), 120.7 (ArCH), 119.9 (ArCH), 114.4 (ArCH),

108.4 (ArCH), 108.1 (ArCH), 101.2 (OCH2O), 86.3 (CHNO2), 65.4 (PNPNCHCO2Et),

61.9 (CO2CH2CH3), 55.7 (OCH3), 45.2 (NCH2), 39.9 (CH2CHCHNO2), 24.4

(NCH2CH2), 14.4 (CO2CH2CH3);

Compound (+/-)-344:

+ Rf (30% EtOAc in hexane): 0.35; HRMS (ESI, m/z) calcd for C24H29N2O8 (M+H)

1 461.1928, found 461.1924. H NMR (400 MHz, CDCl3)  6.86 (m, 1 H), 6.79—6.81 (m,

2H), 6.71 (dd, J = 2.4, 8.8 Hz, 2H), 6.31 (dd, J = 2.4, 9.2 Hz, 2H), 6.02 (dd, J = 1.6, 6.0

Hz, 2H), 5.00 (dd, J = 2.4, 11.2 Hz, 1H), 4.36—4.22 (m, 2H), 4.12 (d, J = 2.8 Hz, 1H),

3.82 (dt, J = 3.6, 11.6, 11.6 Hz, 1H), 3.73 (s, 3H), 3.20 (s, 3H), 3.22—3.16 (m, 1H),

3.09—3.03 (m, 1H), 1.93—1.85 (m, 1H), 1.76—1.67 (m, 1H), 1.39 (t, J = 7.2, 7.2 Hz,

13 3H); C NMR (126 MHz, CDCl3) δ 168.7 (CO2Et), 153.4 (ArC), 148.3 (ArC), 147.4

(ArC), 138.6 (ArC), 131.2 (ArC), 122.1 (ArC), 115.8 (ArCH), 114.9 (ArCH), 109.2

(ArCH), 108.8 (ArCH), 101.4 (OCH2O), 94.0 (CHNO2), 69.1 (NCHCO2Et), 62.5

(CO2CH2CH3), 58.4 (OCH3), 55.8 (OCH3), 55.8 (OCH3), 41.8 (ArC-CH), 32.5

(MeOCH2CH2), 14.2 (CO2CH2CH3);

One-pot method to the synthesis (+/-)-328 using BBr3/Et3SiH

To a cooled (-78 °C) and stirred solution of (+/-)-330 (50.0 mg, 0.10 mmol) in CH2Cl2

(1.00 mL), BBr3 (1.0 M in CH2Cl2, 0.10 mL, 0.10 mmol) was added and the resulting mixture was stirred at -78 °C for 1.5 h and TLC showed full consumption of starting material. Then Et3SiH (25.0 L, 0.15 mmol) was added and the mixture was stirred for

1h during which time the temperature raised to -40°C. A saturated NaHCO3 (5.00 mL) was added to quench the reaction and the mixture was extracted with CH2Cl2 (2 x 5.00 S19

mL). The combined organic layers were washed with brine (5.00 mL), dried (MgSO4) and concentrated. Flash chromatography of the residue on silica gel, using 1:4 EtOAc : pet. Ether, to give the desired (+/-)-328 as a pale-yellow oil (23.0 mg, 50%).

. One-pot method to the synthesis (+/-)-328 using BF3 OEt2/Et3SiH

To a cooled (-40 °C) and stirred solution of (+/-)-330 (294 mg, 0.60 mmol) in CH2Cl2

. (10.0 mL), BF3 OEt2 (0.55 mL, 1.80 mmol) was added and the resulting mixture was stirred at -40 to 0 °C for 4 h and TLC showed full consumption of starting material. Then

Et3SiH (0.15 mL, 0.15 mmol) was added and the mixture was stirred for 1.5 h during which time the temperature raised to room temperature. A saturated NaHCO3 (10.0 mL) was added to quench the reaction and the mixture was extracted with CH2Cl2 (2 x 10.0 mL). The combined organic layers were washed with brine (10.0 mL), dried (MgSO4) and concentrated. Flash chromatography of the residue on silica gel, using 1:4 EtOAc : pet. Ether, to give the desired (+/-)-328 as a pale-yellow oil (162 mg, 63%).

Compound (+/-)-348

Method 1:

PPTS (33.0 mg, 0.16 mmol) was added to a stirred solution of nitroamine (+/-)-330 (120 mg, 0.25 mmol) in toluene (10.0 mL) and the resulting mixture was refluxed at 130 °C for 5 h. The dark brown solution was then cooled to room temperature and quenched with saturated NaHCO3 solution (5.00 mL). EtOAc was used to extract the quenched S20

reaction mixture (2 x 10.0 mL) and the combined organic layers were dried and concentrated. Flash chromatography of the residue on silica gel, using EtOAc : Hexane

1:3, gave the desired product (+/-)-348 as an orange oil (25.0 mg, 23% yield).

-1 Rf (20% EtOAc in hexane) 0.25; FTIR (neat, cm ) 2913, 1731, 1646, 1550, 1507, 1243,

-1 + 1036 cm ; HRMS (ESI, m/z) calcd for C22H22N2O7 (M+H) 427.14998, found 427.1501.

1 H NMR (600 MHz, CDCl3)  7.04—7.02 (m, 2H, PMPCH), 6.87—6.85 (m, 2H,

PMPCH), 6.80 (m, 1H, ArCH), 6.77—6.76 (m, 2H, ArCH), 6.62 (ddd, J = 0.8, 2.1, 8.0

Hz, 1H, PMPNCH=CH), 5.95 (m, 2H, OCH2O), 5.47 (ddd, J = 1.2, 3.6, 6.2 Hz, 1H,

CHNO2), 4.81 (ddd, J = 1.2, 2.8, 8.1 Hz, 1H, PMPNCH=CH), 4.75 (dd, J = 0.8, 3.6 Hz,

1 H, PMPNCHCO2Et), 4.35 — 4.25 (m, 2H, CO2CH2CH3), 4.04—4.01 (m, 1H,

13 CHCHNO2), 3.79 (s, 3H, OCH3), 1.31 (t, 7.1 Hz, 3H, CO2CH2CH3); C NMR (150 MHz,

CDCl3)  169.2 (CO2Et), 156.1 (ArC), 148.0 (ArC), 147.3 (ArC), 140.0 (ArC), 131.7

(PMPNCH=CH), 131.5 (ArC), 121.8 (ArCH), 121.5 (PMPCH), 114.7 (PMPCH), 108.9

(ArCH), 108.4 (ArCH), 101.3 (OCH2O), 97.8 (PMPNCH=CH), 83.0 (CHNO2), 62.7

(CO2CH2CH3), 61.5 (PMPNCHCO2Et), 55.7 (OCH3), 38.9 (CHCHNO2), 14.2

(CO2CH2CH3);

Method 2:

To a cooled (-78 °C) and stirred solution of nitroamine (+/-)-330 (50.0 mg, 0.10 mmol) in CH2Cl2 (1.00 mL) in N2 atmosphere, BBr3 (1.0 M in CH2Cl2, 0.10 mL) was added and the resulting solution was stirred for 1 h. The reaction mixture was then diluted with

CH2Cl2 (5.00 mL) quenched with sat. NaHCO3 aqueous solution (5.00 mL). The aqueous layer was separated and extracted with another portion of CH2Cl2 (5.00 mL). The combined organic layers were dried (Na2SO4) and concentrated. Flash chromatography

S21

of the residue on silica gel, using 1:4 EtOAc : Hexane, gave the desired product (+/-)-348

(15.0 mg, 35 %) as a yellow oil.

Compound (+/-)-358

To a cooled (0 oC) and stirred solution of ester (+/-)-328 (130 mg, 0.30 mmol) in THF

(5.00 mL) was added LiBH4 (20.0 mg, 0.91 mmol) in one portion and the resulting mixture was stirred at 0 oC for 10 min and then at room temperature until full consumption of starting material. The reaction mixture was cooled to 0 oC and quenched with sat.

NH4Cl aqueous solution (10.0 mL), followed by extraction with EtOAc (10.0 mL x 3).

The combined organic layers were dried (Na2SO4) and concentrated to give alcohol (+/-

)-358 (104 mg, 90%) as a white foam.

-1 Rf (33% EtOAc in hexane) 0.25; FTIR (neat, cm ) 3385, 2894, 1545, 1501, 1238, 1037;

HRMS (ESI, m/z) calcd. for C20H23N2O6 [M+H]+ 387.1566, found 387.1567; 1H NMR

(600 MHz, CDCl3)  6.91—6.77 (7H, m, ArH), 5.96—5.94 (2H, m, OCH2O), 5.12—5.10

(1H, m, CHNO2), 4.45—4.40 (1H, m, CHCH2OH), 3.92—3.89 (2H, m, CH2OH), 3.77

(3H, s, OCH3), 3.63—3.60 (1H, m, HCH-NPMP), 3.33 (1H, q, J = 4.2 Hz, Ar-CH),

3.31—3.29 (1H, m, HCH-NPMP), 2.86 (1H, dq, J = 12.6, 4.8 Hz, NCH2-HCH), 1.91—

13 1.89 (1H, m, NCH2-HCH), 1.67 (1H, t, J = 5.4 Hz, CH2OH); C NMR (151 MHz, CDCl3)

 154.0 (ArC), 148.0 (ArC), 146.9 (ArC), 144.0 (ArC), 133.2 (ArC), 120.8 (ArCH), 119.1

(ArCH), 114.7 (ArCH), 108.5 (ArCH), 108.1 (ArCH), 101.2 (OCH2O), 85.7 (CHNO2),

S22

63.3 (NCHCH2OH), 58.4 (CH2OH), 55.7 (OCH3), 43.7 (NCH2), 38.8 (ArC-CH), 24.2

(NCH2CH2).

Compound (+/-)-359 and (+/-)-362

Compound (+/-)-359

To a cooled (0 oC) and stirred solution of (+/-)-328 (25.0 mg, 0.058 mmol) in MeCN (1.00 mL) under N2 was added a solution of CAN (80.0 mg, 0.146 mmol) in H2O (1.00 mL) dropwise. The resulting mixture was stirred at 0 oC for 30 min during which time the colour of mixture changed from wine red to orange. The reaction mixture was diluted with EtOAc (10 mL) and quenched with 10% Na2S2O3 aq. solution (10 mL). Aqueous layer was extracted with another portion of EtOAc (10 mL) and combined organic layers were dried (Na2SO4) and concentrated to the free piperidine (+/-)-359 (16.8 mg, quant, contain small amount of hydroquinone) as a brown oil which was used directly into next

-1 step. Rf (70% EtOAc in hexane) 0.47; FTIR (neat, cm ) 3358, 2957, 2923, 1730, 1549,

1504, 1490, 1442, 1256, 1237, 1037; HRMS (ESI-TOF, m/z) calcd. for C15H19N2O6

+ 1 [M+H] 323.1243, found 323.1242; H NMR (600 MHz, CDCl3)  6.77 (1H, d, J = 8.0

Hz, ArH), 6.75 (1H, d, J = 1.8 Hz, ArH), 6.72—6.67 (1H, m, ArH), 5.95 (2H, q, J = 1.5

S23

Hz, OCH2O), 5.29 (1H, dd, J = 4.2, 2.1 Hz, CHNO2), 4.36 (1H, dq, J = 10.8, 7.2 Hz,

ABX3 system, CO2HCHCH3), 4.33 (1H, dq, J = 10.8, 7.1 Hz, ABX3 system,

CO2HCHCH3), 4.27 (1H, d, J = 2.2, N-CH-CO2Et), 3.26—3.15 (2H, m, ArC-CH and N-

HCH-CH2), 2.86 (1H, ddd, J = 13.4, 12.1, 3.2 Hz, N-HCH-CH2), 2.48 (1H, qd, J = 12.7,

4.5 Hz, N-CH2-HCH), 1.71 (1H, dq, J = 13.2, 3.3 Hz, N-CH2-HCH), 1.38 (3H, t, J = 7.1

13 Hz, CO2CH2CH3); C NMR (151 MHz, CDCl3) δ 170.1 (CCO2Et), 148.0 (C), 147.0 (C),

133.3 (C), 120.7 (CH), 108.5 (CH), 108.0 (CH), 101.3 (CH2), 85.7 (CHNO2), 62.4 (CO2-

CH2CH3), 59.5 (N-CHCO2Et), 42.4 (N-CH2CH2), 40.8 (ArC-CH), 24.8 (N-CH2-CH2),

14.4 (CO2CH2CH3).

Compound (+/-)-362

o To a cooled (-10 C) and stirred solution of above free piperidine (+/-)-359 in CH2Cl2

(1.00 mL) under N2 was added TFAA (10 L, 0.070 mmol) and pyridine (5.5 L, 0.070 mmol) and the resulting mixture was stirred at this temperature for 10 min. The reaction was quenched with 1 M HCl aq (5.0 mL) and extracted with CH2Cl2 (10 mL). Combined organic layers were dried (Na2SO4) and concentrated. Chromatography of reside on silica gel, using 30% EtOAc in hexane, gave (+/-)-362 (20.0 mg, 83% overall) as a colourless

-1 oil. Rf (30% EtOAc in hexane) 0.32; FTIR (neat, cm ) 2984, 2908, 1742, 1697, 1554,

1506, 1491, 1446, 1235, 1205, 1170, 1147; HRMS (ESI-TOF, m/z) calcd for

+ 1 C17H18N2O7F3 [M+H] 419.1066, found 419.1047; H NMR (400 MHz, CDCl3) a mixture of diastereomers, dr 85:15; major isomer:  6.77 (1H, d, J = 7.8, ArH), 6.70—6.59 (2H, m, ArH), 5.96 (2H, s, OCH2O), 5.89 (1H, t, J = 1.7 Hz, N-CH-CO2Et), 5.49—5.42 (1H, m, CH-NO2), 4.47—4.17 (3H, m, CO2CH2CH3 and N-HCH-CH2), 3.39 (1H, ddd, J =

14.3, 13.0, 2.9 Hz, N-HCH-CH2), 3.19 (1H, dt, J = 13.3, 3.4 Hz, ArC-CH), 2.67 (1H, qd,

J = 13.3, 4.4 Hz, N-CH2-HCH), 1.92—1.78 (1H, m, N-CH2-HCH), 1.36 (3H, t, J = 7.1

13 Hz, CO2CH2CH3); C NMR (101 MHz, CDCl3) δ 166.0 (CO2Et), 157.2 (q, J = 37.2 Hz,

S24

C(O)CF3), 148.3 (ArC), 147.6 (ArC), 130.9 (ArC), 120.5 (ArCH), 116.3 (q, J = 287.4 Hz,

CF3), 108.8 (ArCH), 107.5 (ArCH), 101.4 (OCH2O), 84.5 (CHNO2), 63.5

(CO2CH2CH3), 55.5 (NCHCO2Et), 43.4 (NCH2CH2), 41.2 (ArC-CH), 23.8 (NCH2CH2),

19 14.2 (CO2CH2CH3); F NMR (282 MHz, CDCl3) δ -67.90 (minor isomer), -68.97 (major isomer)

Compound (+/-)-382

To a stirred solution of (+/-)-328 (50.0 mg, 0.117 mmol) in MeCN (1.00 mL) was added methyl acrylate (0.03 mL, 0.035 mmol) and Triton-B (40% w/w in MeOH, 2 drops) and the resulting solution was stirred at room temperature overnight. The reaction mixture was quenched with sat. NH4Cl aqueous solution (10.0 mL) and extracted with EtOAc

(10.0 mL x 2). The combined organic layers were dried (Na2SO4) and concentrated.

Chromatography of the residue on silica gel, using 20% EtOAc in hexane, gave (+/-)-382

(21.0 mg, 35%) as a yellow solid.

-1 Rf (20% EtOAc in hexane) 0.14; FTIR (neat, cm ) 2958, 1731, 1542, 1508, 1487, 1440,

+ 1239, 1191, 1037; HRMS (ESI-TOF, m/z) calcd. For C26H31N2O9 [M+H] 515.2029,

1 found 515.2023; H NMR (600 MHz, CDCl3) δ 6.94 (2H, d, J = 9.0 Hz, ArPMPH), 6.84

(2H, d, J = 9.0 Hz, ArPMPH), 6.75 (1H, d, J = 8.0 Hz, ArH), 6.71 (1H, d, J = 1.8 Hz, ArH),

6.64 (1H, dd, J = 8.0, 1.9 Hz, ArH), 6.02 – 5.91 (2H, m, OCH2O), 4.45 (1H, s,

NCHCO2Et), 4.31 (1H, dd, J = 13.3, 4.1 Hz, Ar-CH), 4.07 (2H, qd, J = 7.2, 0.9 Hz,

S25

CO2CH2CH3), 3.89 (1H, td, J = 12.2, 3.9 Hz, NHCHCH2), 3.78 (3H, s, OCH3), 3.67 (3H, s, OCH3), 3.28 (1H, ddd, J = 11.9, 5.7, 1.8 Hz, N-HCHCH2), 2.86–2.75 (1H, m, CH2-

HCHCO2Me), 2.53–2.28 (4H, m, CH2-HCHCO2Me and NCH2-HCH), 2.04–1.98 (m, 1H,

13 NCH2-HCH); C NMR (151 MHz, CDCl3) δ 172.7 (CO2), 168.8 (CO2), 155.2 (ArC),

147.7 (ArC), 147.3 (ArC), 143.1 (ArC), 131.4 (ArC), 122.9 (ArCH), 120.5 (ArPMPCH),

114.6 (ArPMPCH), 109.6 (ArCH), 108.3 (ArCH), 101.3 (OCH2O), 94.8 (C-NO2), 64.2 (N-

CHCO2Et)), 60.9 (CO2CH2CH3), 55.7 (OCH3), 52.1 (OCH3), 44.3 (N-CH2CH2), 43.7

(ArC-CH), 28.8 (CH2CO2Me), 28.5 (NCH2CH2), 24.81 (CH2CH2CO2Me), 14.1

(CO2CH2CH3).

S26

Studies on (+)-vallesamidine and (+)-strempeliopine

Compound (+/-)-468

A mixture of ester (+/-)-323 (1.0 g, 3.75 mmol), (HCHO)n (116 mg, 3.86 mmol) and

o BnNH2 (0.82 mL, 7.49 mmol) in EtOH (10.0 mL) was heated to 90 C in a sealed tube and stirred for 16 h. After cooling to room temperature, the reaction mixture was diluted with EtOAc (30 mL) and washed with HCl aq. (2M, 30 mL). The aqueous layer was extracted with EtOAc (30 mL x 2) and combined organic layers were dried (Na2SO4) and concentrated to give intermediate (+/-)-467 (1.32 g, quant.) as a mixture of diastereomers

(dr 77:23) which was used in next step without purification.

To a round-bottom flask charged with (+/-)-467 (220 mg, 0.62 mmol), Pd(PPh3)2Cl2 (22 mg, 0.031 mmol) and PPh3 (16.3 mg, 0.062 mmol). The system was degassed and backed filled with nitrogen. Then DMF (4.00 mL), DBU (0.23 mL, 1.55 mmol) and allyl acetate

(0.17 mL, 1.55 mmol) was added successively and then stirred at 80 oC for 1 h. After cooling to room temperature, the reaction mixture was quenched with sat. NH4Cl aq. (15 mL) at 0 oC and extracted with EtOAc (15 mL x 2). Combined organic layers were dried

(Na2SO4) and concentrated. Chromatography of residue on silica gel, using 30% EtOAc in hexane, gave (+/-)-468 as a white foam (133 mg, 55%).

S27

-1 Rf (50% EtOAc in hexane) 0.4; FTIR (neat, cm ) 2924, 2899, 1647, 1542, 1489, 1254,

1 1238, 1038, 932; H NMR (400 MHz, CDCl3)  7.39—7.31 (5H, m, CH2-Ph), 6.69 (1H, d, J = 8.0 Hz, ArH), 6.52 (1H, d, J = 1.9 Hz, ArH), 6.48 (1H, dd, J = 8.0, 1.9 Hz, ArH),

5.95 (2H, d, J = 1.1 Hz, OCH2O), 5.27 (1H, dddd, J = 16.8, 10.1, 7.9, 6.6 Hz, CH2-

CH=CH2), 5.04 (1H, dd, J = 10.1, 1.0 Hz, CH2-CH=HCH), 4.92 (1H, d, J = 14.2 Hz, N-

HCH-Ph), 4.81 (1H, dd, J = 16.9, 1.4 Hz, CH2-CH=HCH), 4.46 (1H, d, J = 14.2 Hz, N-

HCH-Ph), 3.72 (1H, d, J = 13.3 Hz, C(=O)N-HCH-C), 3.58—3.52 (1H, m, Ar-CH), 3.39

(1H, d, J = 13.3 Hz, C(=O)N-HCH-C), 2.99 (1H, dd, J = 18.6, 6.9 Hz, NC(=O)-HCH),

2.88—2.77 (2H, m, NC(=O)-HCH and HCH-CH=CH2), 2.54 (1H, dd, J = 14.1, 7.9 Hz,

13 HCH-CH=CH2); C NMR (151 MHz, CDCl3)  167.5 (C(=O)N), 148.2 (ArC), 147.8

(ArC), 136.2 (Ar(Bn)C), 130.5 (ArC), 129.0 (Ar(Bn)CH), 129.0 (Ar(Bn)CH), 128.9

(CH2-CH=CH2), 128.3 (Ar(Bn)CH), 122.5 (CH2-CH=CH2), 121.8 (ArCH), 108.6

(ArCH), 108.5 (ArCH), 101.5 (OCH2O), 89.5 (C-NO2), 50.6 (N-CH2Ph), 46.8 (C(=O)N-

CH2-C), 46.0 (ArC-CH), 40.4 (CH2-CH=CH2), 36.1 (NC(=O)CH2); NC(=O)CH2

Compound (+/-)-473

To a solution of (+/-)-468 (100 mg, 0.254 mmol) in i-PrOH (5.00 mL) was added HCl aq.

Solution (1.0 M, 2.50 mL) and zinc dust (330 mg, 5.08 mmol) and the resulting mixture

o was stirred vigorously at 50 C for 30 min. After cooling to room temperature, sat. K2CO3 aq. Solution (8.00 mL) was added and the mixture was stirred for another 20 min. The precipitates were filtered, and the filtrate was extracted with CHCl3 (20 mL x 3). The

S28

combined organic layers were dried (MgSO4) and concentrated to give (+/-)-473 (92 mg, quant.) which was pure to go to next step.

-1 Rf (50% EtOAc in hexane) 0.1; FTIR (neat, cm ) 3367, 3064, 2905, 1632, 1486, 1441,

+ 1353, 1330, 1037, 925, 729, 705; HRMS (ESI-TOF, m/z) calcd. for C22H26N2O3 [M+H]

1 365.1865, found 365.1840; H NMR (600 MHz, CDCl3)  7.37—7.28 (5H, m, Bn), 6.76

(1H, d, J = 1.8 Hz, ArH), 6.75 (1H, d, J = 7.8 Hz, ArH), 6.65 (1H, dd, J = 7.8, 1.8 Hz,

ArH), 5.96 (2H, d, J = 1.2 Hz, OCH2O), 5.65 (1H, ddt, J = 17.4, 10.1, 7.4 Hz, CH2-

CH=CH2 ) 5.08 (1H, dd, J = 10.2, 2.0 Hz, CH2-CH=HCH), 4.98 (1H, dd, J = 17.0, 1.7

Hz, CH2-CH=HCH), 4.66 (1H, d, J = 14.5 Hz, N-HCHPh), 4.60 (1H, d, J = 14.5 Hz, N-

HCHPh), 3.15 (1H, d, J = 12.6 Hz, N-HCH-C), 2.98—2.86 (3H, m, N-HCH-C, C(O)HCH and Ar-CH), 2.74 (1H, dd, J = 18.0, 6.0 Hz, C(O)HCH), 2.09 (1H, dd, J = 14.0, 7.4 Hz,

13 HCH-CH=CH2), 1.99 (1H, dd, J = 13.9, 7.7 Hz, HCH-CH=CH2); C NMR (151 MHz,

CDCl3)  169.3 (C(=O)N), 147.9 (ArC), 147.0 (ArC), 137.0 (Ar(Bn)C), 133.3 (ArC),

132.4 (CH2-CH=CH2), 128.8 (Ar(Bn)CH), 128.4 (Ar(Bn)CH), 127.7 (Ar(Bn)CH), 122.0

(ArCH), 119.9 (CH2-CH=CH2), 109.1 (ArCH), 108.3 (ArCH), 101.2 (OCH2O). 57.1

(C(=O)N-CH2-C), 52.2 (C-CH2CH=CH2), 50.3 (N-CH2Ph), 47.5 (ArC-CH), 44.3 (CH2-

CH=CH2), 36.6 (NC(=O)CH2);

Compound (+/-)-477

S29

o To a cooled (-78 C) and stirred solution of (+/-)-473 (100 mg, 0.25 mmol) in CH2Cl2

(5.00 mL) was added i-PrNEt2 (0.12 mL, 0.63 mmol) and trifluoroacetic anhydride (0.08 mL, 0.63 mmol). The resulting mixture was stirred at -78 oC for 5 min and then warmed to room temperature. The reaction mixture was quenched with aq. HCl solution (1.0 M,

5 mL) and extracted with CH2Cl2 (10 mL x 2). Combined organic layers were dried

(Na2SO4) and concentrated by rotary evaporation and chromatography of residue on silica gel, using 30% to 40% EtOAc in hexane, gave compound 477 as a yellow oil (68 mg,

60%).

Rf (50% EtOAc in hexane) 0.66; HRMS (ESI-TOF, m/z) calcd. for C24H25N2O4F3

+ 1 [M+H] 461.1688, found 461.1651; H NMR (600 MHz, CDCl3)  7.43—7.29 (5H, m,

CH2Ph), 6.76 (1H, d, J = 7.9 Hz, ArH), 6.60—6.47 (2H, m, ArH), 5.98 (1H, q, J = 1.4

Hz, OCH2O), 5.85 (1H, s, TFA-NH), 5.52 (1H, ddt, J = 17.4, 10.1, 7.4 Hz, CH2-

CH=CH2), 5.14 (1H, dd, J = 10.2, 1.7 Hz, CH2-CH=HCH), 5.03 (1H, qd, J = 17.0, 1.4

Hz, CH2-CH=HCH), 4.74 (1H, d, J = 14.2 Hz, N-HCH-Ph), 4.60 (1H, d, J = 14.2 Hz, N-

HCH-Ph), 3.49—3.44 (2H, m, C(=O)N-CH2-C), 3.37 (1H, dd, J = 7.1, 4.9 Hz, Ar-CH),

2.90 (1H, dd, J = 18.4, 7.1 Hz, NC(=O)-HCH), 2.78 (1H, dd, J = 18.4, 4.9 Hz, NC(=O)-

13 HCH), 2.75—2.65 (2H, m, CH2-CH=CH2); C NMR (151 MHz, CDCl3)  168.2

(C(=O)N), 156.8 and 156.5 (C(=O)N), 148.9 (ArC), 147.8 (ArC), 136.4 (Ar(Bn)C), 131.8

(ArC), 130.7 (CH2-CH=CH2), 129.0 (Ar(Bn)CH), 128.8 (Ar(Bn)CH), 128.1 (Ar(Bn)CH),

121.4 (ArCH), 121.0 (CH2-CH=CH2), 116.3 and 114.4 (COCF3), 108.9 (ArCH), 108.3

(ArCH), 101.6 (OCH2O), 57.1 (C-NHTFA), 50.6 (C(=O)N-CH2-C), 50.3 (CH2Ph), 44.4

19 (ArC-CH), 37.9 (CH2-CH=CH2), 34.8 (NC(=O)-CH2); F NMR (300 MHz, CDCl3)  -

76.1;

S30

Compound (+/-)-466

To a solution of (+/-)-465 (3.00 g, 12.45 mmol) in MeNO2 (15.0 mL) was added tetramethylguanidine (TMG) (0.40 mL, 3.12 mmol) and the resulting mixture was stirred

o at 70 C for 16 h. After cooling to room temperature, MeNO2 was removed by rotary evaporation and aq. HCl solution (1.0 M, 15 mL) was added. The mixture was extracted with EtOAc (15 mL x 3) and combined organic layers were dried and concentrated by rotary evaporation. Chromatography of residue on silica gel, using 20% EtOAc in hexane, gave compound (+/-)-466 as pale-yellow oil (3.75 g, quant.)

-1 Rf (20% EtOAc in hexane) 0.34; FTIR (neat, cm ) 3006, 2953, 1735, 1552, 14721436,

79 + 1375, 1202, 1172; HRMS (ESI-TOF, m/z) calcd. for C11H12 BrNO4Na [M+Na]

1 323.9847, found 323.9845; H NMR (400 MHz, CDCl3) d 7.61 (1H, dd, J = 8.0, 1.3 Hz,

ArH), 7.31 (1H, td, J = 7.6, 1.3 Hz, ArH), 7.21 (1H, dd, J = 7.9, 1.7 Hz, ArH), 7.16 (1H, ddd, J = 8.0, 7.3, 1.7 Hz, ArH), 4.80 (1H, dd, J = 15.9, 6.9 Hz, HCHNO2, AB system),

4.76 (1H, dd, J = 16.1, 6.8 Hz, HCHNO2, AB system), 4.48 (1H, p, J = 7.0 Hz, Ar-CH),

3.66 (3H, s, OCH3), 2.89 (1H, dd, J = 16.6, 7.4 Hz, HCHCO2Me, AB system), 2.86 (1H, dd, J = 16.6, 7.4 Hz, HCHCO2Me, AB system). The data was consistent with the literature. 4

S31

Compound (+/-)-471

To a solution of ester (+/-)-466 (3.50 g, 11.6 mmol) in EtOH (30.0 mL) was added BnNH2

(2.50 mL, 23.2 mmol) and paraformaldehyde (365 mg, 12.2 mmol) and the resulting mixture was stirred at 80 oC overnight. Reaction mixture was cooled to room temperature and quenched with HCl aq. Solution (1.0 M, 50 mL) and extracted with CH2Cl2 (50 mL x 2). Combined organic layers were dried (Na2SO4) and concentrated to give (+/-)-469 as a brown foam (a mixture of two diastereomers, dr 16:9) which was used directly in next step.

To a solution of (+/-)-469 (3.0 g, crude, ca. 7.70 mmol) in CH2Cl2 was added Pd(PPh3)4

(446 mg, 0.39 mmol), allyl acetate (1.25 mL, 11.6 mmol) and DBU (1.73 mmol, 11.6 mmol) under N2 atmosphere. The resulting solution was stirred at room temperature until full consumption of starting material (30 min). Solvent was removed by rotary evaporation and chromatography of the residue on silica gel, using 30% EtOAc in hexane, gave (+/-)-471 (1.60 g, 48% over two steps from (+/-)-466) as a white foam.

FTIR (neat, cm-1) 3064, 2926, 1650, 1540, 1493, 1430, 1353, 1298, 1261; HRMS (ESI-

79 + 1 TOF, m/z) calcd. for C21H21 BrN2O3Na [M+Na] 451.0633, found 451.0593; H NMR

(600 MHz, CDCl3)  7.62 (1H, dd, J = 8.0, 1.3 Hz, ArH), 7.42—7.28 (6H, CH2Ph and

ArH), 7.20 (1H, dd, J = 7.7, 1.6 Hz, ArH), 7.00 (1H, dd, J = 7.9, 1.6 Hz, ArH), 5.41 (1H, dddd, J = 16.5, 10.2, 8.2 6.1 Hz, CH2-CH=CH2), 5.11 (1H, d, J = 9.9 Hz, CH2-CH=HCH),

5.05 (1H, dq, J = 16.9, 1.4 Hz, CH2-CH=HCH), 4.83 (1H, d, J = 14.7 Hz, N-HCH-Ph),

4.64 (1H, d, J = 14.7 Hz, N-HCH-Ph), 4.26 (1H, dd, J = 11.6, 5.6 Hz, Ar-CH), 3.82 (1H, S32

d, J = 14.3 Hz, C(=O)N-HCH-C), 3.50 (1H, dd, J = 14.3 Hz, C(=O)N-HCH-C), 2.90—

2.81 (2H, m, HCH-CH=CH2 and NC(=O)-HCH), 2.77 (1H, dd, J = 17.8, 5.6 Hz, NC(=O)-

13 HCH), 2.63 (1H, dd, J = 14.7, 8.3 Hz, HCH-CH=CH2); C NMR (151 MHz, CDCl3) 

167.7 (C(=O)), 136.0 (Ar(Bn)C), 135.3 (ArC), 133.5 (ArCH), 130.1 (ArCH), 129.2 (CH2-

CH=CH2), 129. (ArCH), 128.9 (Ar(Bn)CH), 128.4 (Ar(Bn)CH), 128.2 (Ar(Bn)CH),

128.0(ArCH), 126.1 (ArC), 122.2 (CH2-CH=CH2), 90.3 (C-NO2), 52.7 (C(=O)N-CH2-

C), 50.2 (N-CH2Ph, 44.0 (ArC-CH), 40.3 (CH2-CH=CH2), 35.1 (NC(=O)-CH2);

Compound 395

To a mixture of NH4OAc (7.50 g, 97.0 mmol) in AcOH (150 mL) was added 2- bromobenzaldehyde (6.30 mL, 54.0 mmol) and the resulting mixture was stirred at room temperature for 10 min before nitromethane (10.0 mL) was added slowly. After addition the reaction solution was flushed with nitrogen and heated to 130 oC for 4 h with stirring.

Then the orange solution was cooled to room temperature and poured into ice water (200 mL) and extracted with CH2Cl2 (150 mL x 2). The combined organic layers were washed with saturated aq. NaHCO3 (150 mL x 2), dried (Na2CO3) and concentrated by rotary evaporation to give crude product as a yellow solid. Hexane (150 mL) was added to the crude product and heated to reflux and during which time diethyl ether was added slowly until most of the solid dissolved. Then the solution was cooled to room temperature and furtherly to 0 oC to afford the recrystalised compound as a microcrystalline yellow solid

(9.40 g, 78%) and data was consistent with the literature.5

S33

o o 1 Rf (1:2 CH2Cl2/hexane): 0.2; m.p. 83.5 – 84.5 C (lit. 87 – 88 C); H NMR (600 MHz,

CDCl3)  8.41 (1H, d, J = 13.6, CH=CHNO2), 7.70 (1H, dd, J = 7.9, 1.3, ArH), 7.58 (1H, dd, J = 7.7, 1.8, ArH), 7.54 (1H, d, J = 13.6, CH=CHNO2), 7.39 (1H, td, J = 7.6, 1.3,

13 ArH), 7.35 (1H, td, J = 7.7, 1.8, ArH); C NMR (151 MHz, CDCl3)  139.0

(CH=CHNO2), 137.7 (CH=CHNO2), 134.2 (ArCH), 133.0 (ArCH), 130.5 (ArC), 128.6

(ArCH), 128.2 (ArCH), 126.5 (ArC).

Compound 394

To a solution of nitroalkene (3.80 g, 16.5 mmol) in toluene (8.5 mL) was added diethylmalonate (2.50 mL, 16.5 mmol) and catalyst (133 mg, 0.16 mmol). The resulting mixture was stirred at room temperature for 3 days and then concentrated by rotary evaporation. Chromatography of the residue on silica gel, using 1:4 EtOAc/hexane, gave the titled compound as a yellow oil (5.83 g, 91%).

25 Rf (1:4 EtOAc/hexane): 0.29; []D +6.42 (c 0.1, CHCl3); Enantiomeric ratio (e.r.): 95:5

(Chiralcel OD-H column, 90:10 hexane:isopropanol, 1mL/min, 215 nm, major enantiomer tr 7.0 min, minor enantiomer tr 12.7 min); absolute stereochemistry was determined by analogy with literature;6 FTIR (neat, cm-1): 2979, 2934, 1726, 1551, 1369,

79 + 1285, 1225, 1022; HRMS (ESI-TOF, m/z) calcd. for C15H18NO6 BrNa [M+Na]

1 410.0210, found 410.0212; H NMR (600 MHz, CDCl3)  7.61 (1H, dd, J = 8.0, 1.2 Hz,

S34

ArH), 7.38 – 7.22 (2H, m, ArH), 7.16 (ddd, 1H, J = 8.0, 7.1, 1.9 Hz, ArH), 5.12 (1H, dd,

J = 13.6, 8.3 Hz, HCHNO2), 4.95 (1H, dd, J = 13.6, 4.4 Hz, HCHNO2), 4.76 (1H, td, J =

8.3, 4.4 Hz, Ar-CH), 4.21 (2H, qd, J = 10.8, 7.2 Hz, CO2CH2), 4.12 – 4.08 (3H, m,

CO2CH2 and CH(CO2Et)2), 1.24 (3H, t, J = 7.1 Hz, CO2CH2CH3), 1.13 (3H, t, J = 7.1 Hz,

13 CO2CH2CH3); C NMR (151 MHz, CDCl3)  167.5 (CO2Et), 167.0 (CO2Et), 135.5

(ArCH), 134.0 (ArCH), 129.8 (ArCH), 128.6 (ArC), 128.0 (ArCH), 125.0 (ArC), 75.9

(CH2NO2), 62.3 (CO2CH2CH3), 62.2 (CO2CH2CH3), 53.4 (CH(CO2Et)2), 41.6 (Ar-CH),

14.1 (CO2CH2CH3), 13.9 (CO2CH2CH3).

Compound 470

To a solution of nitro malonate 394 (13.0 g, 33.5 mmol) in ethanol (90.0 mL) was added p-methoxybenzylamine (8.70 mL, 67.0 mmol) and paraformaldehyde powder (1.20 g,

36.9 mmol) and the resulting mixture was heated to 80 oC under nitrogen atmosphere for

2 hours. Reaction mixture was then cooled to room temperature and quenched with 1 M

HCl aq. solution (100 mL) followed by extraction with EtOAc (150 mL x 3). Combined organic layers were dried (Na2SO4) and concentrated by rotary evaporation to give the crude product 495 which was put into next step directly without further purification.

Sodium chloride (11.0 g) was added to the mixture of above crude product in DMSO/H2O

(36.0 mL/12.00 mL) and the resulting mixture was heated to 150 - 160 oC (heating block) under nitrogen atmosphere overnight. The reaction mixture was cooled to room temperature and diluted with water (150 mL) and extracted with EtOAc (150 x 3 mL). S35

The combined organic layers were washed with brine (300 mL x 2), dried (Na2SO4) and concentrated by rotary evaporation. Chromatography of the crude product on silica gel, using 4:5 EtOAc/ hexane, gave titled compound 470 as a yellow foam (11.8 g, 84%). The proton NMR of product shows a mixture of inseparable diastereomers at C-2 (dr 5:1, determined by 1H NMR). A small fraction of major diastereomer was obtained for data collection.

-1 Rf (50% EtOAc in hexane): 0.34; FTIR (neat, cm ) 2927, 2833, 1642, 1549, 1509, 1242,

79 +. 1024; HRMS (EI, m/z) calcd. for C19H19N2O4 Br [M] 418.05227, found 418.05232.

1 Major isomer: H NMR (700 MHz, CDCl3)  7.59 (1H, dd, J = 8.0, 1.3 Hz, ArH), 7.27

(1H, td, J = 7.6, 1.3 Hz, ArH), 7.22 (2H, d, J = 8.6 Hz, ArPMBH), 7.16 (1H, td, J = 7.7,

1.6 Hz, ArH), 7.13 (1H, dd, J = 7.7, 1.6 Hz, ArH), 6.87 (2H, d, J = 8.6 Hz, ArPMBH), 4.99

(1H, q, J = 5.2 Hz, CHNO2), 4.85 (1H, d, J = 14.4 Hz, N-HCH), 4.43 (1H, q, J = 6.4 Hz,

ArC-CH), 4.38 (1H, d, J = 14.3 Hz, N-HCH), 3.85 (1H, dd, J = 13.6, 5.1 Hz, HCH-

CHNO2), 3.80 (3H, s, OMe), 3.41 (1H, dd, J = 13.7, 4.6 Hz, HCH-CHNO2), 2.97 (1H, dd, J = 18.2, 6.7 Hz, NC(O)HCH), 2.71 (1H, dd, J = 18.2, 6.1 Hz, NC(O)HCH); 13C NMR

(176 MHz, CDCl3)  167.1 (C(O)N), 159.6 (ArPMBC), 137.3 (ArC), 134.0 (ArCH), 130.0

(ArPMBC), 129.9 (ArPMBCH x 2), 128.6 (ArCH), 127.8 (ArCH), 127.7 (ArCH), 124.4

(ArC), 114.3 (ArPMBCH x 2), 81.7 (CHNO2), 55.4 (OCH3), 49.8 (N-CH2PMP), 46.3 (N-

CH2), 41.0 (ArC-CH), 34.2 (NC(O)CH2).

Compound 472

S36

o To a cooled (0 C) and stirred solution of lactam 470 (11.8 g, 28.2 mmol) in CH2Cl2 (95.0 mL) under nitrogen was added Pd(PPh3)4 (330 mg, 0.28 mmol), allyl acetate (4.56 mL,

42.3 mmol) and DBU (6.31 mL, 42.3 mmol). The resulting solution was stirred at 0 oC for 1 hour before it was concentrated by rotary evaporation. Chromatography of the residue on silica gel, using 40% EtOAc in hexane, gave compound 472 as a yellow foam

(11.1 g, 86%).

25 -1 Rf (50% EtOAc in hexane): 0.4; []D +45.3 (c 0.22, CHCl3); FTIR (neat, cm ) 3064,

79 2928, 1647, 1538, 1510, 1243, 1174, 1026; HRMS (EI, m/z) cald for C22H23N2O4 Br

+. 1 [M] 458.08357, found 458.08358; H NMR (700 MHz, CDCl3)  7.60 (1H, dd, J = 8.0,

1.2 Hz, ArH), 7.30 – 7.26 (1H, m, ArH), 7.27 (2H, d, J = 8.5 Hz, ArPMBH), 7.18 (1H, td,

J = 7.7, 1.6 Hz, ArH), 6.98 (1H, dd, J = 8.0, 1.6 Hz, ArH), 6.89 (2H, d, J = 8.5 Hz,

ArPMBH), 5.41 (1H, dddd, J = 16.6, 10.2, 8.2, 6.1 Hz, CH2CH=CH2), 5.11 (1H, d, J = 10.1

Hz, CH=HCH), 5.05 (1H, dd, J = 16, 1.6 Hz, CH=HCH), 4.76 (1H, d, J = 14.4 Hz, C(O)N-

HCH), 4.54 (1H, d, J = 14.4 Hz, C(O)N-HCH), 4.23 (1H, dd, J = 11.5, 5.6 Hz, Ar-CH),

3.81 (3H, s, OMe), 3.79 (1H, d, J = 14.3 Hz, N-HCHPMP), 3.47 (1H, d, J = 14.3 Hz, N-

HCHPMP), 2.86 – 2.78 (2H, m, NC(O)HCH and HCH-CH=CH2), 2.74 (1H, dd, J = 17.8,

13 5.7 Hz, NC(O)HCH), 2.62 (1H, dd, J = 14.7, 8.2 Hz, HCH-CH=CH2); C NMR (176

MHz, CDCl3)  167.6 (C(O)N), 159.4 (ArPMBC), 135.4 (ArC), 133.5 (ArCH), 130.1

(ArCH), 129.9 (ArPMBCH x 2), 129.3 (CH=CH2), 129.0 (ArCH), 128.2 (ArCH), 128.0

(ArPMBC), 126.1 (ArC), 122.1 (CH=CH2), 114.3 (ArPMBCH x 2), 90.3 (CNO2), 55.4

(OCH3), 52.5 (CH2-PMP), 49.6 (C(O)NCH2), 44.1 (ArC-CH), 40.4 (allyl CH2), 35.2

(NC(O)CH2).

S37

Compound 478

To a cooled (0 oC) and stirred solution of nitro compound 472 (3.18 g, 6.93 mmol) in

EtOAc/EtOH (80.0/80.0 mL) was added zinc dust (4.5 g, 69.3 mmol) and 6M HCl aq. solution (23.0 mL, 138.6 mmol). The resulting mixture was stirred at 0 oC until all the starting material was fully consumed (typically around 30 min, checked by TLC).

o Reaction mixture was then poured into saturated NaHCO3 aq. Solution (150 mL) at 0 C and then the whole mixture was filtered through a pad of celite, followed by washing with

EtOAc (200 mL). The organic layer was separated, and aqueous layer was extracted by another portion of EtOAc (150 mL). Combined organic layers were dried (Na2SO4) and concentrated by rotary evaporation to give the primary amine 478 as a grey foam (2.90 g,

98%) without further purification.

25 -1 Rf (66% EtOAc in hexane): 0.11; []D +33.3 (c 0.27, CHCl3); FTIR (neat, cm ): 3372,

79 +. 3310, 3064, 2923, 1636, 1509, 1243; HRMS (EI, m/z) calcd. for C22H25N2O2 Br [M]

1 428.10939, found 428.10940; H NMR (600 MHz, CDCl3) δ 7.57 (1H, dd, J = 8.0, 1.3

Hz, ArCH), 7.45 (1H, dd, J = 7.9, 1.7 Hz, ArCH), 7.28 – 7.25 (3H, m, ArPMBCH2 and

ArCH), 7.12 (1H, td, J = 7.6, 1.7 Hz, ArCH), 6.88 (2H, d, J = 8.6 Hz, ArPMBCH2), 5.65

(1H, ddt, J = 17.5, 10.1, 7.5 Hz, CH2-CH=CH2), 5.09 (1H, dd, J = 10.2, 1.9 Hz, CH2-

CH=HCH), 5.04 – 4.96 (1H, m, CH=HCH), 4.74 (1H, d, J = 14.2 Hz, HCH-PMP), 4.42

(1H, d, J = 14.2 Hz, HCH-PMP), 3.82 (3H, s, OCH3), 3.77 (1H, dd, J = 9.3, 6.3 Hz, ArC-

CH), 3.22 (1H, d, J = 12.6 Hz, CHHNC=O), 2.93 (1H, d, J = 12.7 Hz, CHHNCO), 2.83

(1H, d, J = 9.3 Hz, HCHC(O)N), 2.73 (1H, dd, J = 18.2, 6.3 Hz, HCHC(O)N), 2.27 (1H,

S38

dd, J = 14.0, 7.3 Hz, HCH-CH=CH2), 2.02 (1H, dd, J = 14.0, 7.7 Hz, HCH-CH=CH2);

13 C NMR (151 MHz, CDCl3) δ 169.0 (NC=O), 159.2 (ArPMBC), 139.4 (ArC), 133.3

(ArCH), 132.1 (CH=CH2), 129.9 (ArPMBCH x 2), 129.0 (ArPMBC), 129.0 (ArCH), 128.8

(ArCH), 128.0 (ArCH), 126.4 (ArC), 120.0 (CH=CH2), 114.2 (ArPMBCH x 2), 57.5

(CH2NC(O)), 55.4 (OCH3), 53.0 (C-NH2), 49.8 (CH2-PMP), 45.0 (ArC-CH), 43.3

(CH2CH=CH2), 35.8 (CH2C(O)N).

Compound 479

A flame-dried, argon filled round bottom flask was added CuI (78.0 mg, 0.41 mmol), L- proline (94.0 mg, 0.82 mmol) and K3PO4 (1.70 g, 8.18 mmol), followed by addition of amine 478 (1.75 g, 4.09 mmol) in anhydrous DMSO (10.0 mL). The mixture was degassed and back filled with argon (repeated 3 times) and then heated to 95 oC (heating block) with stirring for 1 hour (TLC indicated all consumption of starting material). After cooling to room temperature, the reaction mixture was added water (50.0 mL) and EtOAc

(50.0 mL). Aqueous layer was separated and extracted with EtOAc (50.0 mL x 2). The combined organic layers were washed with brine (100 mL x 3), dried (Na2SO4) and concentrated to give compound 479 as a brown oil (1.42 g, quant.) which was pure enough to go to next step. A small fraction was purified by chromatography (50% EtOAc/hexane) for data collection.

S39

25 -1 Rf (50% EtOAc in hexane): 0.21; []D -95.1 (c 0.10, CHCl3); FTIR (neat, cm ): 3330,

3079, 2908, 1646, 1608, 1511, 1486, 1244; HRMS (ESI-TOF, m/z): calcd. for

+ 1 C22H25N2O2 [M+H] 349.1916, found 349.1911; H NMR (600 MHz, CDCl3)  7.14 –

6.94 (4H, m, ArPMBCH and ArCH), 6.84 (2H, d, J = 8.6, ArPMBCH), 6.71 (1H, td, J = 7.4,

0.9 Hz, ArCH), 6.41 (1H, d, J = 7.8 Hz, ArCH), 5.66 (1H, ddt, J = 17.3, 10.2, 7.2 Hz,

CH2-CH=CH2), 5.16 – 4.99 (2H, m, CH2-CH=CH2), 4.79 (1H, d, J = 14.5 Hz, HCH-

PMP), 4.13 (1H, d, J = 14.5 Hz, HCH-PMP), 3.82 (3H, s, OMe), 3.45 (1H, t, J = 5.9 Hz,

ArC-CH), 3.35 (1H, d, J = 13.6 Hz, C(O)N-HCH), 3.30 (1H, s, N-H), 3.10 (1H, d, J =

13.6 Hz, C(O)N-HCH), 2.75 (1H, dd, J = 15.1, 6.5 Hz, NC(O)-HCH), 2.66 (1H, dd, J =

13 15.0, 5.4 Hz, NC(O)-HCH), 2.26 (2H, d, J = 7.3 Hz, CH2-CH=CH2); C NMR (151 MHz,

CDCl3) δ 171.0 (NC=O), 159.2 (ArPMBC), 149.8 (ArC), 132.3 (CH=CH2), 130.1

(ArPMBCH x 2), 129.9 (ArC), 129.1 (ArPMBC), 128.4 (ArCH), 124.5 (ArCH), 119.9

(CH=CH2), 119.2 (ArCH), 114.1 (ArPMBCH x 2), 109.3 (ArCH), 65.1 (C-allyl), 55.4

(OCH3), 53.9 (C(O)NCH2), 49.1 (PMP-CH2), 46.0 (ArC-CH), 44.4 (CH2CH=CH2), 37.4

(CH2C(O)N).

Compound 499

To a solution of indoline 479 (3.60 g, 10.3 mmol) in methyl chloroformate (80 mL) was

o added Na2CO3 (11.0 g, 103 mmol) and the resulting mixture was heated to 60 C (oil bath) under N2 atmosphere and stirred overnight. The reaction mixture was cooled to

S40

room temperature and solid was filtered. The filtrate was concentrated via rotary evaporation and chromatography of the residue on silica gel, using 60% EtOAc in hexane, gave compound 499 as a yellow foam (3.70 g, 88%).

25 o Rf (2:1 EtOAc:hexane): 0.26; []D -41.7 (c 0.12, CHCl3); m.p. 107.7 – 109.5 C.

(crystals obtained from over saturated toluene solution) FTIR (neat, cm-1) 3003, 2953,

1696, 1660, 1511, 1484, 1365, 1202; HRMS (ESI-TOF, m/z) calcd. for C24H27N2O4

+ 1 [M+H] 407.1971, found 407.1962; H NMR (600 MHz, CDCl3)  7.87 – 6.78 (8H, m,

ArH), 5.57 – 5.47 (1H, m, CH2-CH=CH2), 5.13 – 5.00 (2H, m, CH2-CH=CH2), 4.90 –

4.74 (1H, m, HCH), 4.13 (1H, d, J = 13.9 Hz, C(O)N-HCH), 3.91 (1H, d, J = 14.6 Hz,

HCH-PMP), 3.81 (3H, s, OCH3), 3.66 – 3.58 (4H, m, ArC-CH and OCH3), 3.54 – 3.37

(1H, m, C(O)N-HCH), 3.02 (1H, brs, HCH-CH=CH2), 2.75 (1H, dd, J = 15.1, 6.0 Hz,

NC(O)-HCH), 2.69 (1H, dd, J = 15.1, 5.2 Hz, NC(O)-HCH), 2.33 – 2.22 (1H, m, HCH-

13 CH=CH2); C NMR (151 MHz, CDCl3)  170.6 (NC=O), 159.0 (ArPMBC), 153.2

(NCO2Me), 142.3 (ArC), 131.5 (CH=CH2), 130.8 (ArC), 130.0 (ArPMBCH X 2), 129.2

(ArC), 128.4 (ArCH), 124.2 (ArCH), 123.5 (ArCH), 120.0 (CH=CH2), 115.3 (ArCH),

113.9 (ArPMBCH X 2), 69.7 (C-Allyl), 55.4 (OCH3), 52.4 (NCO2CH3), 51.2 (C(O)NCH2),

49.0 (CH2-PMP), 44.0 (ArC-CH), 40.2 (CH2-CH=CH2), 37.1 (NC(O)CH2).

Compound 501

S41

To a solution of lactam 499 (2.50 g, 5.90 mmol) in toluene (120 mL) was added

Lawesson’s reagent (1.20 g, 2.96 mmol) and the resulting mixture was heated to reflux and stirred for 1 hour. The clear yellow solution was cooled to room temperature and concentrated by rotary evaporation. Chromatography of the residue on silica gel, using

30% EtOAc in hexane, gave compound 501 as a white foam (2.32 g, 93%).

25 -1 Rf (30% EtOAc in hexane): 0.49; []D -140.0 (c 0.22, CHCl3); FTIR (neat, cm ): 3073,

2953, 1697, 1485, 1438, 1299, 1250, 1238; HRMS (ESI-TOF, m/z): calcd. for

+ 1 C24H27N2O3S [M+H] 423.1742, found 423.1745; H NMR (700 MHz, CDCl3)  7.84 –

6.79 (8H, m, ArH), 5.65 – 5.43 (2H, m, HCH-PMP and CH2-CH=CH2), 5.13 – 4.98 (2H, m, CH2-CH=CH2), 4.69 – 4.42 (1H, m, C(S)N-HCH), 4.32 – 3.92 (1H, m, HCH-PMP),

3.82 (3H, s, OCH3), 3.71 – 3.63 (4H, m, C(S)N-HCH and NCO2CH3), 3.60 (1H, t, J = 5.6

Hz, Ar-CH), 3.42 – 3.11 (2H, m, CH2C(S)N), 3.05 – 2.70 (1H, m, HCH-CH=CH2), 2.29

13 – 2.14 (1H, m, HCH-CH=CH2); C NMR (176 MHz, CDCl3)  200.0 (C=S), 159.4

(ArPMBC), 153.2 (NCO2Me), 142.1 (ArC), 131.0 (CH=CH2), 130.1 (ArPMBCH x 2), 129.9

(ArPMBC), 128.5 (ArCH), 127.6 (ArPMBC), 124.5 (ArCH), 123.6 (ArCH), 120.3

(CH=CH2), 115.4 (ArCH), 114.1 (ArPMBCH X 2), 69.7 (C-allyl), 56.4 (PMP-CH2), 55.4

(OCH3), 53.9 (C(S)NCH2), 52.5 (NCO2CH3), 47.1 (CH2C(S)N), 44.3 (ArC-CH), 40.2

(CH2-CH=CH2).

Compound 502

S42

A solution of thiolactam 501 (696 mg, 1.65 mmol) in methyl iodide (10.0 mL) was heated to reflux and stirred until the complete consumption of starting material (appx. 1 hour, monitored by TLC). Methyl iodide was distilled off (recovered) and the residue was dissolved in methanol (25.0 mL) and cooled to 0 oC and sodium borohydride (0.62 g,16.5 mmol) was added portion wise. The resulting mixture was stirred at the same temperature for and then quenched with saturated NaHCO3 aq. solution (15.0 mL) and water (15.0 mL). The mixture was extracted with CH2Cl2 (30 mL x 3) and the combined organic layers were dried over Na2SO4 and concentrated by rotary evaporation. Chromatography of residue on silica gel, using 20% EtOAc in hexane, gave compound 502 as a colourless oil (560 mg, 87%).

25 -1 Rf (30% EtOAc in hexane): 0.43; []D -54.8 (c 1.04, CHCl3); FTIR (neat, cm ) 3001,

2950, 1700, 1510, 1460, 1376, 1241; HRMS (ESI-TOF, m/z) calcd. for C24H29N2O3

+ 1 [M+H] 393.2178, found 393.2176; H NMR (400 MHz, CDCl3)  7.72 (1H, brs, ArH),

7.22 – 7.14 (3H, m, ArPMBH and ArH), 7.09 (1H, dt, J = 7.4, 1.4 Hz, ArH), 6.99 (1H, td,

J = 7.4, 1.0 Hz, ArH), 6.84 (2H, d, J = 8.5 Hz, ArPMBH), 5.78 (1H, dddd, J = 16.6, 10.1,

8.4, 6.1 Hz, CH2-CH=CH2), 5.14 – 4.96 (2H, m, CH2-CH=CH2), 3.81 (3H, s, OCH3), 3.78

(3H, s, NCO2CH3), 3.40 (1H, d, J = 13.1 Hz, HCH-PMP), 3.38 – 3.35 (1H, m, ArC-CH),

3.31 (1H, d, J = 13.2 Hz, HCH-PMP), 3.13 – 2.94 (2H, m, HCH-CH=CH2 and N-HCH-

C), 2.84 (1H, dd, J = 14.5, 8.4 Hz, HCH-CH=CH2), 2.64 – 2.56 (1H, m, N-HCHCH2),

2.28 (1H, d, J = 11.6 Hz, N-HCH-C), 2.08 – 2.02 (3H, m, N-HCHCH2 and N-CH2CH2);

13 C NMR (101 MHz, CDCl3) δ 158.7 (ArPMBC), 154.1 (NCO2CH3), 142.7 (ArC), 134.2

(CH2CH=CH2), 132.5 (ArC), 130.8 (ArPMBC), 129.9 (ArPMBCH x 2), 127.6 (ArCH),

122.8 (ArCH x 2), 118.2 (CH2CH=CH2), 115.5 (ArCH), 113.6 (ArPMBCH x 2), 69.5 (C-

CH2CH=CH2), 62.0 (PMP-CH2), 58.4 (N-CH2-C), 55.3 (OCH3), 52.2 (NCO2CH3), 49.4

(N-CH2CH2), 41.2 (ArC-CH), 39.2 (CH2CH=CH2), 24.3 (N-CH2CH2).

S43

Compound 503

o To a cooled (0 C) and stirred solution of alkene 502 (2.20 g, 5.60 mmol) in CH2Cl2 (100 mL) was added TFA (1.30 mL, 16.8 mmol) and the resulting solution was stirred at this temperature for 10 min before it was cooled to -78 oC. Then ozone/oxygen flow was bubbled into the solution until a light blue colour appeared and persisted (TLC indicated all consumption of starting material). Ozone generator was stopped, and oxygen was bubbled until the blue colour disappeared and then dimethyl sulfide (4.00 mL, 56.0 mmol) was added. The mixture was stirred for 1.5 hour and triethylamine (4.00 mL, 28.0 mmol) was added to neutralise the acid. The mixture was concentrated by rotary evaporation and chromatography of the residue on silica gel, using 30% EtOAc/hexane, gave titled compound 503 as a colourless oil (1.95 g, 89%).

25 -1 Rf (30% EtOAc/hexane); []D -104.4 (c 0.12, CHCl3); FTIR (neat, cm ) 2948, 2803,

1697, 1509, 1478, 1439, 1364, 1241; HRMS (ESI-TOF, m/z) calcd. for C23H27N2O4

+ 1 [M+H] 395.1971, found 395.1957; H NMR (700 MHz, CDCl3)  9.80 (1H, t, J = 2.6

Hz, CHO), 7.66 (1H, brs, ArH), 7.21 (1H, t, J = 7.8 Hz, ArH), 7.17 (2H, d, J = 8.4 Hz,

ArPMBH), 7.13 (1H, dt, J = 7.4, 1.4 Hz, ArH), 7.04 (1H, td, J = 7.4, 1.0 Hz, ArH), 6.85

(2H, d, J = 8.6 Hz, ArPMBH), 3.81 (6H, s, CO2CH3 and OCH3), 3.43 – 3.31 (3H, m, ArC-

CH and CH2-PMP), 3.25 – 3.09 (3H, m, C-HCH-N and CH2CHO), 2.72 – 2.64 (1H, m,

CH2-HCH-N), 2.25 (1H, d, J = 11.8 Hz, C-HCH-N), 2.13 – 2.02 (3H, m, CH2-CH2-N and

S44

13 CH2-HCH-N); C NMR (176 MHz, CDCl3)  200.7 (CHO), 158.9 (ArPMBC), 153.9

(NCO2Me), 141.8 (ArC), 131.6 (ArC), 130.2 (ArPMBC), 130.1 (ArPMBCH x 2), 128.1

(ArCH), 123.4 (ArCH), 123.3 (ArCH), 115.9 (ArCH), 113.8 (ArPMBCH x 2), 67.8 (C-

CH2-N), 61.9 (CH2-PMP), 57.4 (C-CH2-N), 55.4 (OCH3), 52.7 (NCO2CH3), 49.2 (CH2-

CH2-N), 48.7 (CH2CHO), 43.7 (ArC-CH), 24.4 (CH2-CH2-N);

Compound 526

To a solution of 503 (1.95 g, 4.95 mmol) in 1,2-dichloroethane (10.0 mL) was added allylchloroformate (5.3 mL, 49.5 mmol). The resulting solution was stirred at 65 oC for

24h and then warmed to reflux for another 24h. The reaction mixture was cooled to room temperature and concentrated by rotary evaporation. Chromatography of the residue on silica gel, using 30% EtOAc/hexane, gave compound 526 as a colourless oil (1.00 g, 56% yield, 77% yield brsm).

25 -1 Rf (30% EtOAc/hexane): 0.12; []D -73.3 (c 0.69, CHCl3); FTIR (neat, cm ) 2953,

+. 1697, 1649, 1551, 1461, 1365, 1116, 1076; HRMS (EI, m/z) calcd. for C19H22N2O5 [M]

1 358.1523233, found 358.1522814; H NMR (700 MHz, CDCl3)  9.70 – 9.64 (1H, m,

CHO), 8.08 – 7.33 (1H, m, ArH), 7.23 – 7.18 (1H, m, ArH), 7.12 (1H, d, J = 7.4 Hz,

ArH), 7.02 (1H, t, J = 7.4 Hz, ArH), 5.95 – 5.80 (1H, m, CH2-CH=CH2), 5.35 – 5.10 (2H, m, CH2-CH=CH2), 4.45 (2H, brs, OCH2-CH=CH2), 4.13 – 3.74 (5H, m, N-C-CH2-N and

NCO2Me), 3.49 (1H, dt, J = 2.1, 5.9 Hz, N-HCHCH2), 3.46 – 3.37 (1H, m, ArC-CH),

S45

3.34 – 2.79 (3H, N-HCH-CH2 and CH2CHO), 2.28 – 2.11 (1H, m, N-CH2-HCH), 2.05 –

13 1.87 (1H, m, N-CH2-HCH); C NMR (175 MHz, CDCl3)  199.5 (CHO), 156.1 (NCO2),

153.5 (NCO2), 141.1 (ArCH), 133.0 (CH2CH=CH2), 130.4 (ArC), 128.7 (ArCH), 124.0

(ArCH), 123.8 (ArCH), 117.7 (CH2CH=CH2), 115.5 (ArCH); 67.4 (N-C-CH2N), 66.3

(CO2CH2CH=CH2), 53.0 (CO2CH3), 49.9 (CH2CHO), 45.0 (N-CH2-C), 44.6 (ArC-CH),

40.2 (NCH2CH2), 25.1 (NCH2CH2);

Compound 530

To a cooled (0 oC) and stirred solution of aldehyde 526 (1.44 g, 3.65 mmol) in MeOH

(35.0 mL) was added NaBH4 (270 mg, 7.30 mmol) in one portion. The resulting mixture was stirred at 0 oC for 15 min and concentrated by rotary evaporation to around 10 mL and quenched with sat. NH4Cl aqueous solution (30.0 mL). The mixture was extracted with CH2Cl2 (30.0 mL x 2) and combined organic layers were dried (Na2SO4) and concentrated by rotary evaporation to crude alcohol 529 which was directly used into next step.

To a cooled (-78 oC) and stirred solution of above crude alcohol and 2,6-lutidine (0.70 mL, 5.48 mmol) in CH2Cl2 (35.0 mL) was added TESOTf (1.20 mL, 5.48 mmol) S46

dropwise and the resulting solution was stirred at -78 oC until the alcohol was fully consumed (checked by TLC). A sat. NaHCO3 aqueous solution (20.0 mL) was added to quench the reaction and aqueous layer was separated and extracted with CH2Cl2 (20.0 mL). The combined organic layers were dried (Na2SO4) and concentrated by rotary evaporation. Chromatography of the residue on silica gel, using 30% EtOAc in hexane, gave compound 530 as a colourless oil (1.37 g, 74% over two steps).

25 -1 Rf (30% EtOAc in hexane) 0.57; []D -39.8 (c 0.9, CHCl3); FTIR (neat, cm ): 2949,

2870, 1700, 1510, 1479, 1243, 1076, 747; HRMS (ESI-TOF, m/z) calcd. for

+ 1 C29H43N2O4Si [M+H] 511.2992, found 511.2987; H NMR (700 MHz, CDCl3)  7.75

(1H, brs, ArH), 7.18 (1H, t, J = 7.8 Hz, ArH), 7.15 (2H, d, J = 8.6 Hz, ArPMBH), 7.09 (1H, d, J = 7.3 Hz, ArH), 6.99 (1H, td, J = 7.38, 0.8 HZ, ArH), 6.82 (2H, d, J = 8.6 Hz,

ArPMBH), 3.80 (3H, s, OCH3), 3.78 – 3.68 (5H, m, NCO2CH3 and CH2OTES), 3.46 (1H, brs, Ar-CH), 3.38 (1H, d, J = 13.4 Hz, HCH-PMP), 3.28 (1H, d, J = 13.4. HCH-PMP),

2.94 (1H, d, J = 11.2 Hz, N-HCH-C), 2.61 – 2.55 (1H, m, N-HCH-CH2), 2.47 (1H, brs,

HCH-CH2OTES), 2.38 (1H, ddd, J = 14.3, 8.1, 6.2 Hz, HCH-CH2OTES), 2.28 (1H, d, J

= 11.8 Hz, N-HCH-C), 2.12 – 1.99 (3H, m, N-HCH-CH2 and N-CH2CH2); 0.91 (9H, t, J

13 = 7.9 Hz, Si(CH2CH3)3), 0.55 (6H, q, J = 7.8 Hz, Si(CH2CH3)3); C NMR (175 MHz,

CDCl3)  158.7 (ArPMBC), 154.3 (NCO2Me), 142.8 (ArC), 132.5 (ArC), 131.0 (ArPMBCH x 2), 129.9 (ArCH), 127.7 (ArCH), 122.9 (ArCH), 115.6 (ArCH), 113.7 (ArPMBCH x 2),

69.0 (N-CH2-C), 62.0 (CH2PMP), 59.5 (CH2OTES), 58.8 (N-CH2-C), 55.4 (OCH3), 52.3

(NCO2CH3), 49.3 (N-CH2CH2), 42.2 (Ar-CH), 37.8 (CH2CH2OTES), 24.6 (N-CH2CH2),

6.9 (Si(CH2CH3)3), 4.5 (Si(CH2CH3)3);

S47

Compound 531

To a stirred solution of 530 (1.37 g, 2.69 mmol) in DCE (30.0 mL) was added NaHCO3

(2.25 g, 26.9 mmol) and allyl chloroformate (2.82 mL, 26.9 mmol). The mixture was heated to 80 oC and stirred for 30 min (TLC indicated all consumption of starting material). After cooling to room temperature, NaHCO3 was filtered and the filtrate was concentrated y rotary evaporation and chromatography of the residue on silica gel, using

20% EtOAc in hexane, gave compound 531 as a colourless oil (1.12 g, 88%).

25 -1 Rf (30% EtOAc in hexane) 0.35; []D -32 (c 0.5, CHCl3); FTIR (neat, cm ) 2953, 2876,

1698, 1671, 1484, 1462, 1440, 1078, 750; HRMS (ESI-TOF, m/z) calcd. for

+ 1 C25H39N2O5Si [M+H] 475.2628, found 475.2631; H NMR (700 MHz, CDCl3)  7.90

(1H, brs, ArH), 7.20 – 7.14 (1H, m, ArH), 7.07 (1H,d, J = 7.3 Hz, ArH), 7.26 (1H, t, J =

7.3 Hz, ArH), 5.94 – 5.80 (1H, m, CH=CH2), 5.34 – 5.10 (2H, m, CH=CH2), 4.68 – 4.47

(2H, m, CO2CH2CH=CH2), 4.24 – 4.03 (1H, m, N-HCH-C), 3.90 – 3.68 (4H, m,

NCO2CH3 and N-HCH-C), 3.68 – 3.51 (3H, m, Ar-CH and CH2OTES), 3.45 – 3.36 (1H, m, N-HCH-CH2), 3.22 – 3.10 (1H, m, N-HCH-CH2), 2.42 (1H, brs,NCH2-HCH), 2.20 –

2.08 (1H, m, HCH-CH2OTES), 2.01 – 1.83 (2H, m, NCH2-HCH and HCH-CH2OTES),

13 0.84 (9H, t, J = 8.2 Hz, Si(CH2CH3)3), 0.48 – 0.40 (6H, m, Si(CH2CH3)3); C NMR (176

MHz, CDCl3)  156.1 (NCO2), 153.7 (NCO2), 143.36 (ArC), 133.2 (CH2-CH=CH2),

131.5 (ArC), 128.2 (ArCH), 123.6 (ArCH), 123.2 (ArCH), 117.1 (CH2-CH=CH2), 115.4

(ArCH), 68.8 (NCH2-C), 66.1 (NCO2CH2CH=CH2), 58.5 (CH2OTES), 52.6 (NCO2CH3),

S48

45.3 (N-CH2-C), 44.5 (ArC-CH), 40.2 (NCH2CH2), 26.1 (CH2CH2OTES), 25.2

(NCH2CH2), 6.82 (Si(CH2CH3)3), 4.28 (Si(CH2CH3)3).

Compound 533

To a solution of 531 (3.06 g, 8.54 mmol) in dichloromethane (80.0 mL) under N2 atmosphere was added Pd(PPh3)4 (246 mg, 2.50 mol%) and the resulting yellow solution was stirred at room temperature for 1 hour. The solution was then concentrated by rotary evaporation and flash chromatography of the residue on silica gel, using 30%

EtOAc/hexane, gave compound 533 as a yellow oil (2.47 g, 92% yield).

25 -1 Rf (3:7 EtOAc/hexane): 0.28; []D -36.7 (c 0.15, CHCl3); FTIR (neat, cm ) 2914, 2853,

2806, 1711, 1479, 1462, 1154, 1074; HRMS (ESI-TOF, m/z) calcd. for C18H23N2O3

+ 1 [M+H] 315.1703, found 315.1707; H NMR (700 MHz, CDCl3)  9.86 (1H, brs, CHO),

7.68 (1H, brs, ArH), 7.21 (1H, t, J = 7.8 Hz, ArH), 7.13 (1H, d, J = 7.4 Hz, ArH), 7.03

(1H, t, J = 7.5 Hz, ArH), 5.77 (1H, ddd, J = 14.7, 11.0, 7.0 Hz, CH2-CH=CH2), 5.15 –

5.09 (2H, m, CH2-CH=CH2), 3.85 (3H, s, NCO2Me), 3.38 (1H, brs, ArC-CH), 3.29 – 3.10

(3H, m, CH2CHO and N-C-HCH-N), 2.96 (1H, dd, J = 13.7, 6.3 Hz, HCH-CH=CH2),

2.90 (1H, dd, J = 13.7, 6.6 Hz, HCH-CH=CH2), 2.75 – 2.69 (1H, m, N-HCH-CH2), 2.21

(1H, d, J = 11.9 Hz, N-C-HCH-N), 2.13 – 2.02 (3H, m, N-HCH-CH2 and N-CH2-CH2);

13 C NMR (176 MHz, CDCl3)  200.4 (CHO), 154.0 (NCO2Me), 141.9 (ArC), 134.9

(CH2-CH=CH2), 131.6 (ArC), 128.2 (ArCH), 123.5 (ArCH), 123.3 (ArCH), 118.1 (CH2-

S49

CH=CH2), 115.9 (ArCH), 67.8 (N-C-CH2N), 61.2 (N-CH2CH=CH2), 57.3 (N-C-CH2N),

52.7 (NCO2CH3), 49.4 (NCH2CH2), 48.7 (CH2CHO), 43.6 (ArC-CH), 24.5 (NCH2CH2).

Compound 534

To a solution of aldehyde 533 (725 mg, 2.31 mmol) in DMSO (25.0 mL) was added dimethyl malonate (0.80 mL, 6.93 mmol) and L-proline (400 mg, 4.46 mmol). The resulting mixture was stirred at room temperature overnight. Water (80.0 mL) was added and the mixture was extracted with EtOAc (80.0 mL x 2). Combined organic layers were washed with brine (100 mL x 2), dried (Na2SO4) and concentrated by rotary evaporation.

Flash chromatography of the residue on silica gel, using 30% EtOAc in hexane, gave compound 534 as a colourless oil (915 mg, 93% yield).

25 -1 Rf (30% EtOAc/hexane) 0.12; []D -54.2 (c 0.12, CHCl3); FTIR (neat, cm ) 2952, 1707,

+ 1481, 1439, 1363, 1249; HRMS (ESI-TOF, m/z) calcd. for C23H29N2O6 [M+H]

1 429.2020, found 429.2017; H NMR (700 MHz, CDCl3) δ 7.72 (1H, brs, ArH), 7.19 (1H, t, J = 7.8 Hz, ArH), 7.13 – 7.10 (2H, m, CH=C(CO2Me)2 and ArH), 7.02 (1H, t, J = 7.4

Hz, ArH), 5.78 (1H, ddt, J = 16.7, 10.2, 6.3 Hz, CH2CH=CH2), 5.14 – 5.06 (2H, m,

CH2CH=CH2), 3.85 (6H, brs, CO2CH3 and N-CO2CH3), 3.73 (3H, s, CO2CH3), 3.34 –

3.15 (3H, m, ArC-CH and CH2-CH=C(CO2Me)2), 3.09 (1H, d, J = 11.1 Hz, N-HCH-C),

2.94 – 2.86 (2H, m, CH2-CH=CH2), 2.69 – 2.66 (1H, m, N-HCHCH2), 2.19 (1H, d, J =

13 11.7, N-HCH-C), 2.09 – 1.97 (3H, m, N-HCH-CH2 and N-CH2CH2); C NMR (176

MHz, CDCl3)  166.0 (CO2Me), 164.2 (CO2Me), 154.2 (N-CO2Me), 145.5

S50

(CH=C(CO2Me)2), 142.3 (ArC), 135.3 (CH2CH=CH2), 131.9 (ArC), 129.8

(CH=C(CO2Me)2), 128.0 (ArCH), 123.3 (ArCH), 123.1 (ArCH), 117.7 (CH2CH=CH2),

115.9 (ArCH), 69.3 (N-CCH2N), 61.4 (N-CH2CH=CH2), 58.2 (N-CCH2N), 52.7, 52.53 and 52.48 (CO2Me x 2 and N-CO2Me), 49.6 (N-CH2CH2), 42.2 (ArC-CH), 35.2

(CH2CH=C(CO2Me)2), 24.3 (N-CH2CH2).

Compound 535

To a solution of 534 (3.24 g, 7.57 mmol) in anhydrous toluene (150 mL) under N2 atmosphere was added Yb(OTf)3 (469 mg, 0.76 mmol) and the resulting mixture was heated to 110 oC and stirred for 4 h. The reaction mixture was cooled to room temperature and concentrated by rotary evaporation. Flash chromatography of the crude material on silica gel, using 30% EtOAc in hexane, gave compound 535 as a colourless oil (2.59 g,

80% yield).

25 -1 Rf (30% EtOAc/hexane) 0.39; []D +34.4 (c 0.1, CHCl3); FTIR (neat, cm ) 2951, 1730,

+ 1705, 1484, 1438, 1370, 1260; HRMS (ESI-TOF, m/z) calcd. for C23H29N2O6 [M+H]

1 429.2020, found 429.2018; H NMR (700 MHz, CDCl3)  7.78 (1H, s, ArH), 7.18 (1H, td, J = 7.9, 0.8 Hz, ArH), 7.11 (1H, d, J = 7.4 Hz, ArH), 6.99 (1H, td, J = 7.4, 1.0 Hz,

ArH), 5.63 (1H, dddd, J = 17.2, 10.1, 7.2, 5.8 Hz, NCH2CH=CH2), 5.42 (1H, s, N-C-CH-

C), 5.09 – 4.99 (2H, m, NCH2CH=CH2), 3.86 (3H, s, CO2CH3), 3.78 (3H, s, CO2CH3),

3.72 (3H, s, CO2CH3), 3.30 – 3.24 (2H, m, NCH2CH=CH2), 3.21 (1H, dd, J = 6.6, 3.3 Hz,

ArC-CH), 2.69 (1H, td, J = 13.6, 6.4 Hz, C-HCHCH2C), 2.52 (1H, ddd, J = 11.1, 5.4, 3.5 S51

Hz, N-HCHCH2), 2.40 (1H, td, J = 12.8, 6.6 Hz, C-HCHCH2C), 2.27 – 2.18 (2H, m, C-

CH2-HCH-C and N-HCHCH2), 2.16 – 2.07 (1H, m, N-CH2-HCH), 1.68 (1H, ddt, J =

13.7, 5.6, 2.9 Hz, N-CH2-HCH), 1.61 (1H, ddd, J = 12.6, 6.4, 1.9 Hz, C-CH2-HCH-C);

13 C NMR (176 MHz, CDCl3)  173.0 (CO2Me), 171.1 (CO2Me), 153.7 (NCO2Me), 142.4

(ArC), 135.4 (NCH2CH=CH2), 133.4 (ArC), 127.9 (ArCH), 124.0 (ArCH), 123.1

(ArCH), 117.3 (NCH2CH=CH2), 115.4 (ArCH), 74.8 (C(CO2Me)2), 65.4 (N-C-CHN),

64.9 (N-C-CHN), 60.0 (N-CH2CH=CH2), 53.0 (CO2CH3), 52.6 (CO2CH3), 52.4

(CO2CH3), 45.5 (ArC-CH), 42.1 (N-CH2CH2), 35.0 (N-CH2CH2), 30.0 (C-CH2CH2-C),

29.9 (C-CH2CH2-C);

Compound 536

To a cooled (0 oC) and stirred solution of 535 (555 mg 1.30 mmol) in anhydrous THF

(13.0 mL) under N2 atmosphere was added LiAlH4 (1.0 M in THF, 5.80 mL, 5.80 mmol).

The cooling bath was removed, and the reaction mixture was heated to 65 oC for 30 min.

Heating was then stopped, and the reaction mixture was cooled to 0 oC and quenched with slow addition of H2O (0.35 mL) and 15% NaOH aq. (0.35 mL). The resulting milky suspension was diluted with EtOAc (20.0 mL) and stirred for 10 min at room temperature.

The mixture was then directly dried (Na2SO4) and stirred for another 1 hour. Solids were filtered and the filtrate was concentrated by rotary evaporation to give titled compound

536 as a white foam without further purification (332 mg, 78% yield).

S52

25 -1 Rf (30% acetone in hexane) 0.13; []D +67.3 (c 0.41, CHCl3); FTIR (neat, cm ) 3390,

2938, 2868, 1605, 1481, 1035, 1020; HRMS (ESI-TOF, m/z) calcd. for C20H29N2O2

+ 1 [M+H] 329.2224, found 329.2220; H NMR (600 MHz, CDCl3)  7.10 (1H, td, J = 7.6,

1.3 Hz, ArH), 7.07 (1H, d, J = 7.2 Hz, ArH), 6.67 (1H, td, J = 7.3, 1.0 Hz, ArH), 6.41

(1H, d, J = 7.7 Hz, ArH), 5.84 (1H, ddt, J = 17.6, 9.6, 6.7 Hz, CH2CH=CH2), 5.20 – 5.06

(2H, m, CH2CH=CH2), 4.08 (1H, dd, 11.8, 1.7 Hz, HCHOH), 3.77 – 3.62 (3H, m,

HCHOH and CH2OH), 3.61 – 3.52 (2H, m, HCHCH=CH2 and N-CH-C-N), 3.28 (1H, ddt, J = 13.4, 6.8, 1.3 Hz, HCHCH=CH2), 2.97 (1H, ddd, J = 13.1, 9.1, 3.9 Hz, N-

HCHCH2), 2.91 (1H, t, J = 7.6 Hz, Ar-CH), 2.80 (1H, ddd, J = 13.2, 5.6, 3.6 Hz, N-

HCHCH2), 2.70 (3H, s, N-CH3), 1.75 – 1.69 (2H, m, NCH2CH2), 1.68 – 1.60 (1H, m, C-

HCHCH2-C), 13.8 (1H, ddd, J = 12.9, 13.1, 6.6 Hz, C-HCHCH2-C), 1.29 – 1.23 (1H, m,

13 C-CH2-HCH-C), 1.20 (1H, td, J = 13.1, 6.6 Hz, C-CH2-HCH-C); C NMR (151 MHz,

CDCl3)  151.0 (ArC), 136.0 (CH2CH=CH2), 133.3 (ArC), 127.8 (ArCH), 123.8 (ArCH),

118.5 (ArCH), 117.7 (CH2CH=CH2), 106.7 (ArCH), 74.21 (N-C-CH-N), 72.0 (CH2OH),

70.44 (CH2OH), 63.7 (N-CH), 60.2 (N-CH2CH=CH2), 47.1 (C(CH2OH)2), 46.0 (N-

CH2CH2), 44.2 (Ar-CH), 28.3 (NCH3), 27.6 (NCH2CH2), 27.2 (C-CH2CH2-C), 25.8 (C-

CH2CH2-C).

Compound 537

o To a cooled (0 C) and stirred solution of 536 (439 mg, 1.34 mmol) in CH2Cl2 (10.0 mL) was added triethylamine (0.30 mL, 2.01 mmol) and DMAP (10.0 mg), followed by slow

S53

addition of a solution of TBSCl (242 mg, 1.61 mmol). The reaction solution was slowly warmed to room temperature and stirred overnight. Water (10.0 mL) was added, and the mixture was extracted with CH2Cl2 (10.0 mL x 2) and the combined organic layers were dried (Na2SO4) and concentrated. Chromatography of the residue on silica gel, using

10% EtOAc in hexane, gave titled compound 537 as a colourless oil (342 mg, 58% yield).

25 -1 Rf (10% EtOAc in hexane): 0.15; []D +11.0 (c 0.10, CHCl3); FTIR (neat cm ) 3397,

2951, 2928, 2856, 1606, 1482, 1254, 1081; HRMS (ESI-TOF, m/z) calcd. for

+ 1 C26H43N2O2Si [M+H] 443.3088, found 443.3087; H NMR (600 MHz, CDCl3)  7.09

(1H, td, J = 7.6, 1.3 Hz, ArH), 7.05 (1H, d, J = 7.1 Hz, ArH), 6.67 (1H, td, J = 7.3, 1.0

Hz, ArH), 6.41 (1H, d, J = 7.7 Hz, ArH), 5.89 (1H, dddd, J = 17.9, 9.4, 7.3, 6.0 Hz,

CH2CH=CH2), 5.18 – 5.07 (2H, m, CH2CH=CH2), 4.33 (1H, brs, OH), 4.05 (1H, dd, J =

9.6, 0.8 Hz, HCHOH), 3.78 (1H, d, J = 10.2 Hz, HCHOTBS), 3.72 (H, d, J = 9.6 Hz,

HCHOH), 3.62 (1H, d, J = 10.2 Hz, HCHOTBS), 3.51 (1H, dd, J = 13.5, 7.3 Hz, N-

HCHCH=CH2), 3.31 (1H, s, N-CH-C-N), 3.24 (1H, ddt, J = 13.5, 6.1, 1.4 Hz, N-

HCHCH=CH2), 2.88 (1H, dd, J = 10.2, 6.3 Hz, Ar-CH), 2.74 (1H, dt, J = 13.7, 4.0 Hz,

N-HCHCH2), 2.71 – 2.65 (4H, m, N-CH3 and N-HCHCH2), 1.79 (1H, ddd, J = 14.0, 8.2,

4.5 Hz, C-HCHCH2-C), 1.68 (1H, ddd, J = 12.5, 7.6, 4.5 Hz, C-CH2-HCH-C), 1.61 – 1.38

(3H, m, N-CH2CH2 and C-HCHCH2-C), 1.24 (1H, dt, J = 12.9, 8.6 Hz, C-CH2-HCH-C);

13 C NMR (151 MHz, CDCl3)  151.0 (ArC), 137.1 (CH2CH=CH2), 133.8 (ArC), 127.6

(ArCH), 123.4 (ArCH), 117.9 (ArCH), 117.7 (CH2CH=CH2), 107.2 (ArCH), 71.9 (N-C-

CH-N), 70.9 (CH2O), 65.6 (CH2O), 64.8 (N-CH), 60.3 (N-CH2CH=CH2), 50.3

(OCH2CCH2O), 48.2 (N-CH2CH2), 46.7 (Ar-CH), 30.4 (N-CH3), 28.4 (CH2CH2), 27.5

(CH2CH2), 26.1 (SiMe2C(Me)3), 24.9 (NCH2CH2), 18.4 (SiC), -5.3 and -5.4 (SiMe2).

S54

Compound 538

To a solution of alcohol 537 (191 mg, 0.43 mmol) in EtOAc (10.0 mL) was added IBX

(363 mg, 1.29 mmol) in one portion and the resulting suspension was heated to 80 oC and stirred for 1.5 h. After cooling to room temperature, the precipitates were filtered, and the filtrate was concentrated by rotary evaporation to give crude product 538 as a yellow oil which was used directly without further purification (190 mg, quant. yield, 5:1 with

C20 epimer).

-1 Rf (10% EtOAc in hexane): 0.6; FTIR (neat, cm ) 2952, 2930, 2856, 1722, 1606, 1483,

+ 1470, 1253, 1093; HRMS (ESI-TOF, m/z) calcd. for C26H41N2O2Si [M+H] 441.2932,

1 found 441.2939; Major (desired): H NMR (400 MHz, CDCl3)  9.74 (1H, s, CHO),

7.12 – 7.00 (2H, m, ArH x 2), 6.65 (1H, td, J = 7.4, 1.0 Hz, ArH), 6.38 (1H, d, J = 7.8

Hz, ArH), 5.85 (1H, dddd, J = 17.2, 10.1, 7.0, 5.9 Hz, CH2CH=CH2), 5.21 – 5.04 (2H, m,

CH2CH=CH2), 3.93 (1H, d, J = 9.7 Hz, HCHOTBS), 3.65 (1H, d, J = 9.7 Hz,

HCHOTBS), 3.53 (1H, s, N-CH), 3.45 (1H, ddt, J = 13.9, 7.0, 1.3 Hz, HCH-CH=CH2),

3.25 (1H, ddt, J = 13.9, 5.9, 1.5 Hz, HCH-CH=CH2), 2.89 (1H, t, J = 7.8 Hz, Ar-CH),

2.67 (3H, s, N-Me), 2.55 – 2.51 (1H, m, N-HCHCH2), 2.24 – 2.16 (1H, m, N-HCHCH2),

1.71 – 1.35 (6H, m, N-CH2CH2 and C-CH2CH2-C), 0.90 – 0.88 (9H, m, SiMe2t-Bu), 0.09

13 – 0.07 (6H, m, SiMe2t-Bu); C NMR (101 MHz, CDCl3)  205.5 (CHO), 151.1 (ArC),

137.3 (CH2CH=CH2), 133.3 (ArC), 127.5 (ArCH), 123.6 (ArCH), 117.5 (ArCH), 117.1

(CH2CH=CH2), 106.4 (ArCH), 73.1 (N-C-CH-N), 66.4 (N-CH), 64.0 (N-CH2CH=CH2),

59.9 (C-CHO), 58.4 (CH2OTBS), 45.4 (N-CH2CH2), 44.5 (Ar-CH), 28.2 (N-CH3), 27.6

S55

(CH2CH2), 26.8 (CH2CH2), 25.9 (SiMe2C(CH3)3), 23.4 (NCH2CH2), 18.3 (Si-C), -5.5 (Si-

CH3 x 2).

Compound 539

To a round bottom flask charged with 538 (190 mg, 0.43 mmol) in toluene (5.30 mL) under N2 was added a solution of Cp2TiMe2 in toluene (ca. 17% w/w, 898 mg, 4.32 mmol) and the resulting solution was heated to 65 oC for 4 h. The reaction was cooled to room temperature and hexane (20.0 mL) was added. The precipitated were filtered through a short pad of celite and washed with ether. Silica gel (2.00 g) was added to the filtrate and then concentrated. The resulting dry powder was transferred to the silica column and purified with 5% EtOAc in hexane to afford the titled compound 539 as a yellow oil as a mixture of C-20 epimers (110 mg, 60% from 537) with dr of 5:1 (a second column was needed in some cases to completely remove the Petasis reagent residues). Note: To optimise the purification, in later Petasis olefination experiments, a solution of

o MeOH/H2O (9:1) was added and stirred at 65 C for 2-3 hours to quench the reaction and precipitate titanium residues. After simple filtration, the solution was concentrated to give crude alkene product that could be used directly into next step. The yield of crude product was remained at 55%.

-1 Rf (10% EtOAc in hexane): 0.66; FTIR (neat, cm ) 2952, 2930, 2856, 1607, 1482, 1253,

+ 1085; HRMS (ESI-TOF, m/z) calcd. for C27H43N2OSi [M+H] 439.3139, found

1 439.3143; H NMR (400 MHz, CDCl3)  7.06 (1H, td, J = 7.6, 1.4 Hz, ArH), 7.01 (1H,

S56

d, J = 7.2 Hz, ArH), 6.62 (1H, td, J = 7.4, 1.0 Hz, ArH), 6.37 (1H, d, J = 7.7 Hz, ArH),

6.18 (1H, dd, J = 17.9, 11.0 Hz, C-CH=CH2), 5.90 – 5.79 (1H, m, CH2CH=CH2), 5.19 –

4.98 (4H, m, C-CH=CH2 and N-CH2CH=CH2), 3.62 (1H, s, N-CH), 3.55 (1H, d, J = 9.3

Hz, HCHOTBS), 3.53- 3.48 (1H, m, N-HCH-CH=CH2), 3.32 – 3.26 (1H, m, N-HCH-

CH=CH2), 3.22 (1H, d, J = 9.3 Hz, HCHOTBS), 2.86 (1H, ddd, J = 13.6, 11.0, 2.9 Hz,

N-HCHCH2), 2.79 (1H, dd, J = 9.6, 6.3 Hz, ArC-CH), 2.70 (3H, s, NMe), 2.57 (1H, dt, J

= 13.2, 4.2 Hz, N-HCHCH2), 1.90 – 1.34 (6H, m, NCH2CH2 and C-CH2CH2-C), 0.91

13 (9H, s, SiMe2t-Bu), 0.05 and 0.06 (6H, s, SiMe2t-Bu); C NMR (101 MHz, CDCl3) 

151.5 (ArC), 142.8 (CH=CH2), 138.5 (CH2-CH=CH2), 134.2 (ArC), 127.4 (ArCH),

123.42 (ArCH), 117.1 (ArCH), 116.2 (CH2-CH=CH2), 113.1 (CH=CH2), 106.4 (ArCH),

72.8 (N-C-CH-N), 69.3 (N-CH), 61.8 (N-CH2CH=CH2), 60.0 (CH2OTBS), 51.3 (C-

CH=CH2), 46.2 (N-CH2CH2), 45.9 (ArC-CH), 28.7 (N-CH3), 28.4 (CH2CH2), 27.8

(CH2CH2), 26.1 (SiC(CH3)3, 25.4 (NCH2CH2), 18.5 (SiC), -5.3 (Si(CH3)2);

Compound 540

To a solution of 539 (120 mg, 0.27 mmol) in PhMe (10.0 mL) under nitrogen was added

Hoveyda-Grubbs 2nd catalyst (12.0 mg, 0.014 mmol) and the resulting solution was heated to 60 oC and stirred for 1 hour. After cooling to room temperature, the reaction mixture was concentrated by rotary evaporation and chromatography of the residue on silica gel, using 10 to 30% EtOAc/hexane, gave 540 as a yellow oil (59.6 mg, 53% yield) and epi-540 (7 mg, 6% yield) as a yellow oil.

S57

25 Compound 540: Rf (30% EtOAc in hexane): 0.28; []D +67.6 (c 1.5, CHCl3);

FTIR (neat, cm-1) 2950, 2927, 1607, 1482, 1470, 1252, 1092; HRMS (ESI-TOF, m/z)

+ 1 calcd. for C25H39N2OSi [M+H] 411.2826, found 411.2828; H NMR (700 MHz, CDCl3)

 7.08 (1H, td, J = 7.7, 1.4 Hz, ArH), 7.03 (1H, d, J = 7.2 Hz, ArH), 6.69 (1H, td, J = 7.3,

1.0 Hz, ArH), 6.49 (1H, d, J = 7.7 Hz, ArH), 5.84 – 5.75 (2H, m, CH=CH), 3.51 (1H, d,

J = 9.4 Hz, HCH-OTBS), 3.46 (1H, d, J = 9.5 Hz, HCH-OTBS), 3.12 – 2.96 (3H, m, N-

CH and N-CH2-CH=CH), 2.92 (1H, dd, J = 9.2, 6.5 Hz, ArC-CH), 2.78 (3H, s, N-Me),

2.74 (1H, ddd, J = 12.6, 8.1, 4.6 Hz, N-HCH-CH2), 2.33 (1H, ddd, J = 11.5, 6.9, 4.6 Hz,

N-HCH-CH2), 2.01 – 1.51 (6H, m, N-CH2CH2 and C-CH2CH2-C), 0.89 (9H, s, SiMe2t-

13 Bu), 0.04 (6H, s, SiMe2t-Bu); C NMR (176 MHz, CDCl3)  152.1 (ArC), 134.6 (ArC),

133.6 (CH=CH), 127.5 (ArCH), 124.4 (CH=CH), 123.1 (ArCH), 118.2 (ArCH), 108.9

(ArCH), 76.9 (N-C-CH-N), 69.3 (CH2OTBS), 65.3 (N-CH), 52.5 (N-CH2-CH=CH), 48.8

(N-CH2CH2), 46.9 (C-CH2OTBS), 45.0 (ArC-CH), 31.5 (C-CH2CH2-C), 31.0 (N-Me),

28.2 (C-CH2CH2-C), 27.4 (N-CH2CH2), 26.1 (SiC(Me)3), 18.5 (SiC), -5.2 and -5.3

(SiMe2t-Bu).

25 Compound epi-540: Rf (30% EtOAc in hexane) 0.36; []D -5.7 (c 0.17, CHCl3); FTIR

(neat, cm-1) 2952, 2928, 2856, 1607, 1483, 1253, 1083, 837; HRMS (ESI-TOF, m/z) calcd

+ 1 for C25H39N2OSi [M+H] 411.2832, found 411.2830; H NMR (400 MHz, CDCl3) 

7.12—7.03 (2H, m, ArH), 6.66 (1H, td, J = 7.3, 1.0 Hz, ArH), 6.42 (1H, d, J = 7.7 Hz,

ArH), 6.27 (1H, ddd, J = 9.8, 2.9, 1.8 Hz, CH=CH), 5.64 (1H, ddd, J = 9.9, 3.6, 2.5 Hz,

CH=CH), 3.90 (1H, ddd, J = 18.6, 3.6, 1.9 Hz, N-HCH-CH=CH), 3.80—3.69 (2H, m, C-

CH2-OTBS), 3.40—3.30 (2H, m, N-CH and N-HCH-CH=CH), 2.79 (1H, dd, J = 10.8,

6.4 Hz, N-HCH-CH2), 2.73 (3H, s, N-Me), 2.69 (1H, dd, J = 7.9, 3.5 Hz, ArC-CH), 2.06

(1H, dd, J = 12.3, 7.7 Hz, C-HCH-CH2-C), 1.94 (1H, dd, J = 14.6, 8.4 Hz, C-CH2-HCH-

C), 1.84 (1H, ddt, J = 13.5, 6.7, 3.4 Hz, N-CH2-HCH), 1.70 (1H, ddd, J = 14.7, 12.3, 7.8

S58

Hz, C-CH2-HCH-C), 1.55—1.42 (2H, m, N-HCH-CH2 and N-CH2-HCH), 1.20 (1H, td,

J = 12.5, 8.7 Hz, C-HCH-CH2-C), 0.90 (9H, s, SiMe2-t-Bu), 0.21—0.18 (6H, m, SiMe2-

13 t-Bu); C NMR (101 MHz, CDCl3)  151.2 (ArC), 137.0 (CH=CH), 134.7 (ArC), 127.5

(ArCH), 126.5 (CH=CH), 123.0 (ArCH), 117.7 (ArCH), 107.8 (ArCH), 71.9 (N-CH-C-

N), 67.0 (CH2OTBS), 64.5 (N-CH-C-N), 53.8 (N-CH2-CH=CH), 49.7 (C-CH2OTBS),

48.9 (ArC-CH), 47.3 (N-CH2CH2), 31.9 (N-CH2-CH2), 30.9 (C-CH2-CH2-C ), 28.6 (C-

CH2-CH2-C), 28.1 (N-Me), 26.0 (SiC(Me)3), 18.5 (SiC(CH3)3), -5.3 (SiMe2).

Tandem olefination/ring closing metathesis protocol from 538：Aldehyde 538 (48 mg, 0.11 mmol) in a round bottom flask under nitrogen was added Petasis reagent (15% w/w in PhMe, 1.50 mL) and the resulting solution was heated to 65 oC in dark until TLC indicated all consumption of starting material. PhMe (3.50 mL) was added, followed by addition of Hoveyda-Grubbs 2nd catalyst (6.27 mg, 0.01 mmol) and the resulting mixture was stirred at 65 oC fo1.5 hours. The reaction was then cooled to 40 oC and added aq.

MeOH (MeOH/H2O 9:1, 0.20 mL) and the whole mixture was stirred at this temperature overnight. Precipitates were filtered and the filtrated was concentrated. Chromatography of residue on silica gel, using 10 to 30% EtOAc in hexane, gave titled compound 540 as a yellow oil (20.3 mg, 62% yield based on single diastereomer).

Compound 541

To a solution of 540 (34.0 mg, 0.083 mmol) in THF (2.00 mL) was added TBAF (1.0 M in THF, 0.88 mL) and the resulting solution was heated to 65 oC for 1 hour. After cooling

S59

to room temperature, the solution was concentrated by rotary evaporation and chromatography of the residue on silica gel, using 100% acetone, gave the titled compound 541 as a colourless oil (22.6 mg, 92% yield).

25 -1 Rf (100% acetone): 0.23; []D +34.3 (c 0.21, CHCl3); FTIR (neat, cm ) 3363, 2937,

+ 2861, 1607, 1481, 1144, 1132; HRMS (ESI-TOF, m/z) calcd. for C19H25N2O [M+H]

1 297.1961, found 297.1960; H NMR (700 MHz, CDCl3)  7.19 (1H, tt, J = 7.5, 1.2 Hz,

ArH), 7.10 (1H, dt, J = 7.4, 1.4 Hz, ArCH), 7.04 (1H, td, J = 7.4, 1.0 Hz, ArCH), 6.99

(1H, d, J = 7.7 Hz, ArCH), 6.48 (1H, brs, OH) 5.90 (1H, ddd, J = 9.8, 6.2, 1.8 Hz, N-

CH2CH=CH), 5.56 (1H, ddd, J = 9.8, 2.7, 0.9 Hz, N-CH2CH=CH), 3.50 (1H, t, J = 3.5

Hz, ArC-CH), 3.35 (1H, dd, J = 10.6, 8.5 Hz, HCHOH), 3.27 (HCHOH), 2.95 (1H, dd, J

= 15.7, 6.2 Hz, N-HCH-CH=CH), 2.71 (3H, s, NMe), 2.65 – 2.56 (2H, m, N-HCH-

CH=CH and N-HCHCH2), 2.37 (1H, ddd, J = 13.7, 11.0, 9.1 Hz, C-HCHCH2-C), 2.32

(1H, s, N-CH), 2.20 – 2.14 (3H, m, N-HCHCH2 and N-CH2-HCH and C-CH2-HCH-C),

2.12 – 2.07 (1H, m, N-CH2-HCH), 1.95 (1H, ddt, J = 13.7, 7.2, 2.0 Hz, C-HCHCH2-C),

13 1.73 (1H, ddd, J = 13.3, 9.1, 2.3 Hz, C-CH2-HCH-C); C NMR (176 MHz, CDCl3) 

153.0 (ArC), 137.3 (ArC), 132.7 (N-CH2CH=CH), 128.0 (ArCH), 125.8 (N-

CH2CH=CH), 123.9 (ArCH), 123.0 (ArCH), 117.4 (ArCH), 76.9 (N-C-CH-N), 72.8 (N-

C-CH-C), 69.6 (CH2OH), 52.0 (N-CH2CH=CH), 49.4 (N-CH2CH2), 49.4 (C-CH2OH),

41.5 (ArC-CH), 37.3 (N-CH3), 30.5 (C-CH2CH2-C), 28.5 (C-CH2CH2-C), 24.5 (N-

CH2CH2).

Compound 542

S60

Procedure: To a solution of alcohol 541 (36.0 mg, 0.122 mmol) in EtOAc (2.00 mL) was added IBX (102 mg, 0.365 mmol) in one portion and the mixture was heated to 80 oC and stirred for 2 hours. After cooling to room temperature, the solids were filtered through a short pad of silica gel and the the filtrate was concentrated to give titled compound 542 as a yellow oil (34.5 mg, 96% yield) which was used directly into next step.

-1 Rf (100% EtOAc): 0.65; FTIR (neat, cm ) 2934, 2806, 1713, 1476, 1462; HRMS (ESI-

+ 1 TOF, m/z) calcd. for C19H23N2O [M+H] 295.1805, found 295.1807; H NMR (700 MHz,

CDCl3)  9.20 (1H, s, CHO), 7.14 (1H, tt, J = 7.6, 1.2 Hz, ArCH), 7.05 (1H, dt, J = 7.3,

1.5 Hz, ArCH), 6.90 (1H, td, J = 7.4, 1.0 Hz, ArCH), 6.75 (1H, d, J = 7.8 Hz, ArCH),

5.94 (1H, ddd, J = 9.8, 5.6, 1.7 Hz, CH2CH=CH), 5.34 (1H, ddd, J = 9.8, 2.9, 1.3 Hz,

CH2CH=CH), 3.42 (1H, brs, ArC-CH), 3.19 (1H, ddd, J = 16.5, 5.6 1.3 Hz, N-HCH-

CH=CH), 2.78 (1H, dt, J = 13.7, 8.7 Hz, C-HCHCH2-C), 2.75 – 2.69 (1H, m, N-

HCHCH2), 2.63 (1H, ddd, J = 16.6, 2.9, 1.8 Hz, N-HCH-CH=CH), 2.54 (3H, s, NMe),

2.49 (1H, s, N-CH), 2.34 (1H, ddd, J = 14.3, 11.5, 8.2 Hz, C-CH2-HCH-C), 2.22 – 2.13

(3H, m, N-HCHCH2 and N-CH2CH2), 1.94 (1H, ddd, J = 14.3, 9.2, 3.0 Hz, C-CH2-HCH-

13 C), 1.72 (1H, ddd, J = 13.6, 10.9, 3.0 Hz, C-HCH-CH2-C); C NMR (176 MHz, CDCl3)

 197.8 (CHO), 154.3 (ArC), 135.4 (ArC), 128.5 (NCH2CH=CH), 128.0 (ArCH), 125.2

(NCH2CH=CH), 122.8 (ArCH), 121.6 (ArCH), 115.5 (ArCH), 74.8 (N-C-CH-N), 71.0

(N-C-CH-N), 60.5 (C-CHO), 53.2 (N-CH2CH=CH), 49.8 (N-CH2CH2), 42.5 (ArC-CH),

35.1 (N-CH3), 28.3 (C-CH2CH2-C), 27.0 (C-CH2CH2-C), 23.5 (N-CH2CH2).

S61

Compound 387

To a round bottom flask charged with 542 (29 mg, 0.10 mmol) under N2 was added

o Cp2TiMe2 solution (15% w/w, 1.40 mL) and the resulting solution was warmed to 60 C and stirred in dark for 5 hours. The aq. methanol (MeOH/H2O 9:1, 1.00 mL) was added and the mixture was stirred at 60 oC for another 2 hours to decompose the Petasis reagent.

After cooling to room temperature, the precipitates were filtered, and the solution was concentrated by rotary evaporation. Chromatography of the residue on silica gel, using

20% EtOAc/hexane, gave titled compound 387 as a colourless oil (13.0 mg, 45% yield).

25 -1 Rf (20% EtOAc/hexane): 0.2; []D +130.8 (c 0.13, CHCl3); FTIR (neat, cm ) 2954,

+ 2922, 2851, 1668, 1483, 1462; HRMS (ESI-TOF, m/z) calcd. for C20H25N2[M+H]

1 293.2012, found 293.2008; H NMR (700 MHz, CDCl3)  7.07 (1H, td, J = 7.6, 1.3 Hz,

ArH) 7.02 (1H, d, J = 7.1 Hz, ArH), 6.67 (1H, td, J = 7.3, 1.0 Hz, ArH), 6.42 (1H, d, J =

7.7 Hz, ArH), 5.89 (1H, dd, J = 17.5, 10.6 Hz, CH=CH2), 5.76 (1H, ddd, J = 9.9, 5.8, 2.0

Hz, CH2CH=CH), 5.56 (1H, ddd, J = 9.9, 2.7, 0.8 Hz, N-CH2CH=CH), 5.12 (1H, dd, J =

17.5, 0.9 Hz, CH=HCH), 5.09 (1H, dd, J = 10.6, 0.9 Hz, CH=HCH), 3.13 (1H, ddd, J =

16.5, 5.8, 0.8 Hz, N-HCH-CH=CH), 3.00 (1H, t, J = 8.0 Hz, ArC-CH), 2.87 – 2.0 (2H, m, HCH-CH=CH2 and N-HCHCH2), 2.78 (1H, s, N-CH), 2.71 (3H, s, NMe), 2.29 (1H, ddd, J = 11.4, 7.5,6.1 Hz, N-HCHCH2), 2.06 (1H, dq, J = 14.1, 7.1 Hz, NCH2-HCH), 2.00

(1H, dddd, J = 12.6, 6.7, 3.0, 1.2 Hz, C-HCHCH2-C), 1.88 (1H, ddd, J = 12.6, 10.6, 6.5,

C-CH2-HCH-C), 1.81 (1H, ddd, J = 12.6, 10.6, 7.3 C-HCHCH2-C), 1.72 – 1.65 (2H, m,

13 NCH2-HCH and C-CH2-HCH-C); C NMR (176 MHz, CDCl3)  152.0 (ArC), 145.6

(CH=CH2), 134.6 (ArC), 133.5 (CH2CH=CH), 127.5 (ArCH), 122.9 (ArCH), 122.9

S62

(CH2CH=CH), 117.8 (ArCH), 112.6 (CH=CH2), 108.0 (ArCH), 77.9 (N-C-CH-N), 72.1

(N-C-CH-N), 52.9 (CH2CH=CH), 49.8 (C-CH=CH2), 49.7 (N-CH2CH2), 44.6 (ArC-CH),

33.4 (C-CH2CH2-C), 31.3 (N-CH3), 28.2 (C-CH2CH2-C), 27.3 (N-CH2CH2).

(+)-Vallesamidine (102)

To a round bottom flask charged with 387 (8 mg, 0.027 mmol) was added Pd/C (10% w/w, 5 mg) and methanol (4.0 mL) under N2. The system was degassed and back-filled with H2 (balloon) and the mixture was stirred in H2 atmosphere for 3 hours. The Pd/C was filtered through a pad of celite and the filtrate was concentrated by rotary evaporation.

Chromatography of the residue on silica gel, eluted with acetone, to give vallesamidine

(102) as a slightly yellow oil (6.2 mg, 77%).

25 Rf (50% EtOAc/hexane): 0.18; []D +72.7 (c 0.22, CHCl3), Literature data for (-)-

20 7 20 8 vallesamidine: []D = −81.5 (c 0.33, CHCl3) ; []D −76.6 (c 0.25, CHCl3) ; FTIR

(neat, cm-1) 2936, 2855, 2796, 2750 (Bohlman-Wenkert bands), 1607, 1481; HRMS (ESI-

+ 1 TOF, m/z) calcd. for C20H29N2 [M+H] 297.2325, found 297.2325; H NMR (700 MHz,

CDCl3)  7.06 (1H, t, J = 7.6 Hz, ArH), 7.02 (1H, d, J = 7.1 Hz, ArH), 6.65 (1H, t, J =

7.3 Hz, ArH), 6.42 (1H, d, J = 7.7 Hz, ArH), 2.91 – 2.84 (2H, m, ArC-CH and N-

HCHCH2), 2.83 – 2.76 (4H, m, NMe and N-HCHCH2CH2), 2.29 – 2.17 (3H, m, N-CH,

N-HCHCH2 and N-HCHCH2CH2), 1.93 – 1.86 (1H, m, N-CH2-HCH), 1.84 – 1.74 (2H, m, N-CH2-HCH-CH2 and C-HCHCH2-C), 1.67 – 1.59 (3H, m, C-HCHCH2-C, C-CH2-

HCH-C and HCH-CH3), 1.58 – 1.44 (5H, m, N-CH2CH2CH2, N-CH2-HCHCH2 and N-

S63

CH2-HCH and HCHCH3), 1.43 – 1.38 (1H, m, C-CH2-HCH-C), 0.89 (3H, t, J = 7.4 Hz,

13 CH2CH3); C NMR (176 MHz, CDCl3)  151.5 (ArC), 135.0 (ArC), 127.3 (ArCH), 123.0

(ArCH), 117.7 (ArCH), 107.7 (ArCH), 79.1 (N-C-CH-N), 73.2 (N-C-CH-C), 50.6 (N-

CH2CH2), 50.0 (N-CH2CH2CH2), 44.6 (C-CH2CH3), 44.4 (ArC-CH), 35.6 (C-CH2-CH2-

C), 31.4 (N-CH3), 31.2 (CH2CH3), 30.4 (C-CH2CH2-C), 27.7 (N-CH2CH2), 26.6 (N-

CH2CH2CH2), 18.5 (N-CH2CH2CH2), 9.3 (CH2CH3).

Comparison of 1H NMR data

Error Lit. data 1 (600 MHz, Lit. data 2 (600 MHz, Our data (700 MHz,  /ppm 8 7 CDCl3) CDCl3) CDCl3) (1/2)

7.07 (1H, t, 7.3) 7.07 (1H, t, 7.2) 7.06 (1H, t, 7.6) -0.01/-0.01

7.03 (1H, d, 7.3) 7.02 (1H, d, 7.2) 7.02 (1H, d, 7.1) -0.01/0

6.67 (1H, t, 7.3) 6.66 (1H, t, 7.2) 6.65 (1H, t, 7.3) -0.02/-0.01

6.44 (1H, d, 7.3) 6.42 (1H, d, 7.8) 6.42 (1H, d, 7.7) -0.02/0

3.00 – 2.70 (3H, m) 2.94 – 2.78 (3H, m) 2.91 – 2.76 (3H, m) -

2.78 (3H, s) 2.78 (3H, s) 2.78 (3H, s) 0

2.50 – 2.25 (3H, m) 2.32 -2.24 (3H, m) 2.29 – 2.17 (3H, m) -

2.15 – 1.35 (12H, m) 1.95 – 1.38 (12H, m) 1.93 – 1.38 (12H, m) -

0.90 (3H, t, 7.3) 0.90(3H, t, 7.8) 0.89 (3H, t, 7.4) -0.01/-0.01

Comparison of 13C NMR data

Lit. data 1 (150 Lit. data 2 (150 Our data (176 Error  /ppm (1/2)

8 7 MHz, CDCl3) MHz, CDCl3) MHz, CDCl3)

151.3 151.3 151.5 0.2/0.2

S64

134.6 134.8 135.0 0.4/0.2

127.3 127.1 127.3 0/0.2

122.9 122.8 123.0 0.1/0.2

117.7 117.5 117.7 0/0.2

107.6 107.5 107.7 0.1/0.2

78.6 78.9 79.1 0.5/0.2

72.6 72.9 73.2 0.6/0.3

50.1 50.3 50.6 0.5/0.3

49.6 49.8 50.0 0.4/0.2

44.4 44.4 44.6 0.2/0.2

44.0 44.2 44.4 0.4/0.2

35.4 35.4 35.6 0.2/0.2

31.2 31.2 31.4 0.2/0.2

31.0 31.1 31.2 0.2/0.1

30.1 30.2 30.4 0.3/0.2

27.3 27.4 27.7 0.4/0.3

26.4 26.5 26.6 0.2/0.1

18.1 18.3 18.5 0.4/0.2

9.1 9.1 9.3 0.2/0.2

S65

Compound 499 (by triphosgene/MeOH)

To a cooled (-10 oC) and stirred solution of indoline 479 (2.27 g, 6.52 mmol) and pyridine

(1.60 mL, 19.6 mmol) in CH2Cl2 (100 mL) was added a solution of triphosgen (1.94 g,

6.52 mmol) in CH2Cl2 (30.0 mL) dropwise over 30 min under nitrogen atmosphere. After addition, TLC indicated all consumption of starting material and the reaction mixture was quenched with methanol (2.00 mL) and then concentrated by rotary evaporation. The residue was dissolved in methanol (30.0 mL) and heated to reflux and stirred overnight.

After cooling to room temperature, the reaction mixture was concentrated by rotary evaporation and chromatography of the residue on silica gel, using 60% EtOAc in hexane, gave compound 499 as a light-yellow oil (2.32 g, 87%).

Compound 557

To a solution of 499 (5.10 g, 12.6 mmol) in acetone/water (10:1, 110 mL) was added

NMO (2.20 g, 18.8 mmol), 2,6-lutidine (3.00 mL, 25.1 mmol) and OsO4 (4% w/w in water, 1.60 mL, 0.25 mmol). The resulting solution was stirred at room temperature for

24 h (TLC indicated all consumption of alkene) and PhI(OAc)2 (6.0 g, 18.8 mmol) was

S66

added and the mixture was stirred for 0.5 h. Acetone was removed by rotary evaporation and the residue was added EtOAc (150 mL) and quenched with sat. Na2SO3 aq. (150 mL).

The aqueous layer was separated and extracted with EtOAc (150 mL x 2). Combined organic layers were washed with brine (150 mL), dried (Na2SO4) and concentrated.

Chromatography of the crude product on silica gel, using 3:2 to 4:1 EtOAc/hexane, gave aldehyde 557 as a colourless oil (4.01 g, 78%).

25 -1 Rf (100% EtOAc): 0.38; []D -14.4 (c 1.22, CHCl3); FTIR (neat, cm ) 2951, 2833,

1690, 1655, 1608, 1510, 1483, 1438, 1241, 751, 727; HRMS (ESI-TOF, m/z) calcd. for

+ 1 C23H25N2O5 [M+H] 409.1764, found 409.1748; H NMR (700 MHz, CDCl3)  9.59 (1H, s, CHO), 7.30 (1H, brs, ArH), 7.18 (1H, brs, ArH), 7.15 (1H, d, J = 7.5 Hz, ArH), 7.02

(1H, t, J = 7.4 Hz, ArH), 6.93 (2H, d, J = 8.2 Hz, ArPMBH), 6.78 - 6.77 (2H, m, ArPMBH),

4.78 (1H, brs, HCH-PMP), 4.10 (1H, brs, C(O)N-HCH), 3.90 (1H, brs HCH-PMP), 3.80

(3H, s, OCH3), 3.70 (1H, t, J = 4.3 Hz, ArC-CH), 3.66 – 3.60 (4H, m, OCH3 and C(O)N-

HCH), 3.25 (1H, brs, HCH-CHO), 3.05 (1H, brs, HCH-CHO), 2.98 (1H, dd, J = 15.2, 6.1

Hz, NC(O)-HCH), 2.70 (1H, dd, J = 15.2, 4.5 Hz, NC(O)-HCH); 13C NMR (176 MHz,

CDCl3)  199.0 (CHO), 170.5 (NC=O), 159.1 (ArPMBC), 153.3 (NCO2Me), 141.5 (ArC),

130.3 (ArC), 130.0 (ArPMBCH X 2), 129.0 (ArPMBC), 128.7 (ArCH), 124.4 (ArCH), 124.0

(ArCH), 115.3 (ArCH), 114.0 (ArPMBCH X 2), 67.6 (C-CH2CHO), 55.4 (OCH3), 52.6

(OCH3), 50.1 (NCO2CH3), 49.5 (C(O)NCH2), 48.9 (CH2-PMP), 45.5 (ArC-CH), 37.0

(CH2CHO).

S67

Compound 558

To a cooled (-10 oC) stirred solution of aldehyde 557 (1.59 g, 3.88 mmol) in methanol

(40.0 mL) was added NaBH4 (0.29 g, 7.77 mmol) portion-wise and the resulting mixture was stirred at -10 to 0 oC for 30 min. Methanol was removed by rotary evaporation and the residue was quenched with sat. NH4Cl aq. (40.0 mL) and extracted with CH2Cl2 (40.0 mL x 3). Combined organic layers were dried (Na2SO4) and concentrated by rotary evaporation to give crude alcohol (1.56 g, 98%) which was used directly.

o To a cooled (0 C) and stirred solution of crude alcohol (1.56 g, 3.80 mmol) in CH2Cl2

(40.0 mL) was added i-Pr2NEt (1.00 mL, 5.82 mmol), DMAP (50.0 mg) and TESCl (0.80 mL, 4.65 mmol). The resulting solution was stirred at 0 oC for 30 min and quenched with water (50.0 mL). Aqueous layer was extracted with CH2Cl2 (50.0 mL) and combined organic layers were dried (Na2SO4) and concentrated by rotary evaporation.

Chromatography of the residue on silica gel, using 50% EtOAc in hexane, gave compound 558 as a colourless oil (1.88, 93% overall yield).

25 -1 Rf (30% EtOAc in hexane) 0.1; []D -12.0 (0.35, CHCl3); FTIR (neat, cm ) 2949, 2871,

1695, 16621 1501, 1484, 1439, 1242, 1078, 747, 727; HRMS (ESI-TOF, m/z) calcd for

+ 1 C29H40N2O5Si [M+H] 525.2785, found 525.2726; H NMR (700 MHz, CDCl3) 

8.02—6.46 (8H, m, ArH), 4.94—4.66 (1H, m, N-HCH-PMP), 4.20—3.85 (3H, m, N-

HCH-PMP, ArC-CH and C(=O)N-HCH-C), 3.80 (3H, s, NCO2Me), 3.67—3.24 (6H,m,

Ar-OMe and C(=O)N-HCH-C and CH2OTES), 2.76 (1H, dd, J = 15.1, 6.1 Hz, NC(=O)-

HCH), 2.67 (1H, dd, J = 15.1, 4.5 Hz, NC(=O)-HCH), 2.46—2.14 (1H, m, HCH- S68

CH2OTES), 1.84 (1H, brs, HCH-CH2OTES), 0.78 (9H, t, J = 8.0 Hz, Si(CH2CH3)3),

13 0.42—0.32 (6H, m, Si(CH2CH3)3); C NMR (176 MHz, CDCl3) δ 170.8 (NC=O), 159.0

(ArPMBC), 153.2 (NCO2Me), 142.4 (ArC), 131.3 (ArC), 129.9 (ArPMBCH X 2), 129.3

(ArPMBC), 128.1 (ArCH), 124.2 (ArCH), 123.5 (ArCH), 115.4 (ArCH), 113.9 (ArPMBCH

X 2), 69.3 (N-C-CH2-N), 58.7 (CH2OTES), 55.4 (OCH3), 52.3 (OCH3), 51.5 (N-CH2-C-

N), 48.9 (N-CH2-PMP), 44.9 (ArC-CH), 38.0 (CH2CH2OTES), 36.9 (NC(=O)-CH2), 6.8

(Si(CH2CH3)3), 4.2 (Si(CH2CH3)3).

Compound 530 (by lactam reduction)

To a stirred solution of 558 (280 mg, 0.53 mmol) and NiCl2(dme) (12.0 mg, 0.053 mmol) in toluene (2.00 mL) under nitrogen atmosphere was added PhSiH3 (0.13 mL, 1.06 mmol). The resulting mixture was heated to 110 oC and stirred overnight. After cooling to room temperature, the reaction mixture was filtered through a short pad of celite and concentrated by rotary evaporation. Chromatography of the residue on silica gel, using

17% EtOAc in hexane, gave titled compound 530 as a colourless oil (210 mg, 78%)

S69

Compound 526 (by Swern oxidation)

o To a cooled (-78 C) and stirred solution of DMSO (1.20 mL, 17.1 mmol) in CH2Cl2 (15.0 mL) under nitrogen atmosphere was added (COCl)2 (0.73 mL, 8.53 mmol) dropwise. The resulting solution was stirred at -78 oC for 30 min and then a solution of 531 (940 mg,

o 1.92 mmol) in CH2Cl2 (5.00 mL) was added dropwise and stirred at -78 C until full consumption of silyl ether (checked by TLC). Then Et3N (4.20 mL, 29.1 mmol) was added dropwise and the resulting mixture was stirred and slowly warmed to 0 oC.

Saturated NH4Cl aq. solution (20.0 mL) was added to quench the reaction and separated aqueous layer was extracted by another portion of CH2Cl2 (20.0 mL). Combined organic layers were dried (Na2SO4) and concentrated by rotary evaporation. Chromatography of the residue on silica gel, using 30% EtOAc in hexane, gave titled compound 526 as a colourless oil (610 mg, 86%). Note: After addition of Et3N, raising the temperature from

–78 oC to rt caused the formation of silyl enol ether 532 and treatment of 532 with 2 M

HCl aq. solution in THF reformed the aldehyde 526.

Compound 559

S70

To a cooled (-78 oC) and nitrogen filled round bottom flask charged with aldehyde 557

(1.755 g, 4.30 mmol) in CH2Cl2 (40.0 mL) was added (CH2OTMS)2 (1.10 mL, 4.5 mmol) and TMSOTf (0.08 mL, 0.43 mmol). The resulting solution was stirred at -78 oC for 10 min and then warmed to room temperature and stirred for 1 hour. Saturated NaHCO3 aq

(10 mL) was added to quench the reaction and the mixture was extracted with CH2Cl2 (10 mL x 2). The combined organic layers were dried (Na2SO4) and concentrated by rotary evaporation to give compound 559 as a colourless oil (1.95 g, quant.) which was pure to go to next step without further purification. A small amount of sample was purified by column chromatography (silica gel, 80% EtOAc in hexane) for data collection.

25 -1 Rf (80% EtOAc in hexane): 0.29; []D -22.5 (c 0.63, CHCl3); FTIR (neat, cm ) 2995,

2885, 1696, 1662, 1510, 1483, 1440, 1357, 1243, 751; HRMS (ESI-TOF, m/z) calcd. for

+ 1 C25H29N2O6 [M+H] 453.2026, found 453.2031; H NMR (700 MHz, CDCl3)  7.94—

7.27 (1H, brs x 2, ArH), 7.23 – 7.10 (2H, m, ArH), 6.99 (1H, t, J = 7.3 Hz, ArH), 6.95

(2H, d, J = 8.8 Hz, ArPMBH), 6.77 (2H, brs, ArPMBH), 4.78 (2H, brs, HC(O,O) and HCH-

PMP), 4.14 (1H, brs, (O)N-HCH-C), 4.04 – 3.88 (1H, m, HCH-PMP), 3.98 (1H, t, J =

5.4 Hz, Ar-CH), 3.86 – 3.69 (5H, m, OCH3 and OCH2CH2O), 3.69 – 3.56 (5H, OCH3 and

OCH2CH2O), 3.56 – 3.34 (1H, m, (O)N-HCH-C), 2.76 (1H, dd, J = 15.0, 6.3 Hz,

HCHC(O)N), 2.66 (1H, dd, J = 15.1, 4.9, HCHC(O)N), 2.43 (1H, brs, HCH-CH(O,O)),

13 1.95 (1H, dd, J = 14.5, 4.5 Hz, HCH-CH(O,O)); C NMR (176 MHz, CDCl3)  170.7

(C(O)N), 159.0 (ArPMBC), 153.3 (NCO2Me), 142.2 (ArC), 131.0 (ArC), 129.9 (ArPMBCH

X 2), 129.3 (ArC), 128.3 (ArCH), 124.3 (ArCH), 123.5 (ArCH), 115.4 (ArCH), 113.9

(ArPMBCH X 2), 101.6 (HC(O,O)), 68.2 ((O)N-CH2-C), 64.8 (OCH2CH2O), 64.6

(OCH2CH2O), 55.4 (OCH3), 52.4 (NCO2CH3), 51.5 (C(O)NCH2), 49.0 (CH2-PMP), 44.4

(ArC-CH), 39.2 (CH2CH(O,O)), 37.0 (CH2C(O)N);

S71

Compound 560

To a solution of lactam 559 (364 mg, 0.80 mmol) in THF (3.00 mL) under nitrogen was added Mo(CO)6 (21.0 mg, 0.08 mmol) and phenylsilane (0.26 mL, 2.08 mmol) and the resulting mixture was heated to 75 oC and stirred for 6 to 8 h. The reaction solution was cooled to 0 oC and quenched with 1 M NaOH aq. slowly until gas evolution ceased. Water

(10.0 mL) and EtOAc (10.0 mL) were added and the aqueous layer was separated and extracted with EtOAc (10.0 mL). Combined organic layers were dried (Na2SO4) and concentrated. Chromatography of the residue on silica gel, using 30% EtOAc in hexane, gave the compound 560 as a colourless oil (267 mg, 76%).

25 -1 Rf (80% EtOAc in hxane): 0.63; []D -42.8 (c 0.72, CHCl3); FTIR (neat, cm ) 2948,

2881, 2798, 1699, 1509, 1477, 1439, 1476, 1359, 1242, 1125; HRMS (ESI-TOF, m/z)

+ 1 cald. for C25H31N2O5 [M+H] 439.2233, found 439.2221; H NMR (700 MHz, CDCl3) 

7.75 (1H, brs, ArH), 7.18 (1H, t, J = 7.8 Hz, ArH), 7.15 (2H, d, J = 8.5 Hz, ArH), 7.09

(1H, J = 7.3 Hz, ArH), 6.99 (1H, J = 7.1 Hz, ArH), 6.82 (2H, d, J = 8.6 Hz, ArH), 4.98

(1H, dd, J = 6.0, 3.1 Hz, HC(O,O)), 3.96 – 3.88 (2H, m, OCH2CH2O), 3.80 (3H, s, OCH3),

3.77 (3H, brs, OCH3), 3.76 – 3.73 (2H, m, OCH2CH2O), 3.68 (1H, t, J = 4.0 Hz, Ar-CH),

3.37 (1H, d, J = 13.1 Hz, N-HCH-C), 3.30 (1H, d, J = 13.2 Hz, N-HCH-C), 2.98 (1H, d,

J = 10.0 Hz, N-HCH-C), 2.74 (1H, brs, C-HCH-CH(O,O)), 2.63 – 2.59 (1H, m, N-HCH-

CH2), 2.52 (1H, dd, J = 15.2, 6.0 Hz, C-HCH-CH(O,O)), 2.24 (1H, d, J = 11.4, N-HCH-

13 C), 2.11 – 2.00 (3H, m, N-HCH-CH2 and N-CH-CH2); C NMR (176 MHz, CDCl3) 

158.7 (ArPMBC), 154.4 (NCO2Me), 142.7 (ArC), 132.4 (ArC), 130.9 (ArPMBC), 129.9

S72

(ArPMBCH x 2), 127.7 (ArCH), 122.9 (ArCH), 122.8 (ArCH), 115.6 (ArCH), 113.7

(ArPMBCH x 2), 102.5 (CH(O,O)), 68.0 (N-CH2-C), 65.0 (OCH2CH2O), 64.4

(OCH2CH2O), 62.1 (N-CH2-PMP), 58.8 (N-CH2-C), 55.4 (OCH3), 52.4 (OCH3), 49.3 (N-

CH2CH2), 41.8 (ArC-CH), 38.4 (C-CH2CH(O,O)), 24.2 (N-CH2CH2).

Compound 526 (by acid hydrolysis)

Procedure: To a stirred solution of 560 (4.78 g, 10.9 mmol) in ClCH2CH2Cl (55.0 mL) under N2 atmosphere was added allyl chloroformate (5.60 mL, 54.6 mmol), and NaHCO3

(4.6 g, 54.6 mmol) and the resulting mixture was heated to 80 oC for 1 h. After cooling to room temperature, the NaHCO3 was filtered and filtrate was concentrated by rotary evaporation and dried on high vacuum for 6 h. This crude dioxolane product 561 was then dissolved in THF (7.00 mL), followed by addition of acetic acid AcOH (14.0 mL)

o and H2O (7.00 mL). The resulting mixture was the stirred at 90 C until full consumption of dioxolane (24 h). The reaction mixture was cooled to room temperature, diluted with

EtOAc (100 mL) and washed successfully with H2O (100 mL x 2) and sat. NaHCO3 aq. solution (100 mL x 3). The combined organic layers were dried (Na2SO4) and concentrated by rotary evaporation. Chromatography of the residue on silica gel, using

30% EtOAc in hexane, gave aldehyde 526 as a colourless oil (2.82 g, 72% over two steps). S73

Compound 569

To a cooled (0 oC) and stirred solution of 535 (216 mg, 0.59 mmol) in THF (10.0 mL) was added LiAlH4 (1.0 M in THF, 2.40 mmol, 2.40 mmol) under N2. Cooling bath was then removed, and the solution was stirred at room temperature (22 oC) for 15 min. The

o reaction solution was cooled back to 0 C and quenched by dropwise addition of H2O

(0.30 mL) and 15% NaOH aq. (0.30 mL). EtOAc(15.0 mL) was added and the resulting mixture was stirred at room temperature for 15 min. Anhydrous Na2SO4 was then added and stirring was continued for another 30 min before the precipitates were filtered through a pad of celite. The resulting filtrate was concentrated, and the residue was simply purified by a short silica column (5 cm), eluted by 80% EtOAc in hexane, to remove any non-polar side products. The diol 546 obtained (178 mg, 61% yield), with acceptable purity, was put into next step directly without further purification.

o To a cooled (0 C) and stirred solution of diol (160 mg, 0.43 mmol) in CH2Cl2 (4.00 mL) was added Et3N (0.13 mL, 0.86 mmol), DMAP (16.0 mg) and TBDPSCl (0.13 mL, 0.47 mmol). The resulting mixture was stirred at 0 oC for 10 min and then warmed to room temperature and stirred until all consumption of starting material. The reaction solution was diluted with CH2Cl2 (15.0 mL) and washed with water (10.0 mL x 2). The organic S74

layer was dried (Na2SO4) and concentrated by rotary evaporation. Chromatography of the residue on silica gel, using 20% EtOAc in hexane, gave compound 569 (220 mg, 84%,

51% over 2 steps) as a white foam.

25 -1 Rf (20% EtOAc in hexane): 0.29; []D +45.2 (c 0.87, CHCl3); FTIR (neat, cm ) 3485,

2928, 2853, 1698, 1480, 1439, 1360, 1108, 1063, 751, 702; HRMS (ESI-TOF, m/z) calcd.

+ 1 for C37H47N2O4Si [M+H] 611.3300, found 611.3310; H NMR (700 MHz, CDCl3) 

7.74—7.71 (4H, m, SiPh2-t-Bu), 7.66 (1H, brs, ArH), 7.47—7.39 (6H, m, SiPh2-t-Bu),

7.15 (1H, td, J = 7.4, 1.3 Hz, ArH), 7.04 (1H, d, J = 7.3 Hz, ArH), 6.96 (1H, td, J = 7.4,

0.8 Hz, ArH), 5.69 (1H, ddt, J = 16.9, 10.2, 6.7 Hz, NCH2-CH=CH2), 5.05 (1H, dd, J =

10.0, 1.3 Hz, NCH2-CH=HCH), 4.99 (1H, dd, J = 17.1, 1.4 Hz, NCH2-CH=HCH), 4.39

(1H, s, N-CH-C-N), 4.08 (1H, d, J = 10.1 Hz, HCH-OTBDPS), 3.97 (1H, d, J = 10.3 Hz,

HCH-OH), 3.86 (3H, s, NCO2CH3), 3.80 (1H, d, J = 10.2 Hz, HCH-OH), 3.61 (1H, d, J

= 10.0 Hz, HCH-OTBDPS), 3.21 (1H, ddt, J = 13.9, 6.5, 1.3 Hz, N-HCH-CH=CH2), 3.10

(1H, ddt, J = 14.0, 6.7, 1.3 Hz, N-HCH-CH=CH2), 2.67 (1H, t, J = 6.9 Hz, Ar-CH), 2.42—

2.31 (2H, m, N-CH2-CH2), 2.19 (1H, dt, J = 14.5, 7.9 Hz, C-HCH-CH2-C), 1.75 (1H, dt,

J = 13.4, 7.0 Hz, C-CH2-HCH-C), 1.63 (1H, dt, J = 13.0, 7.6 Hz, C-CH2-HCH-C), 1.46

13 (1H, dt, J = 13.3, 6.7 Hz, C-HCH-CH2-C), 1.42—1.34 (2H, m, N-CH2-CH2); C NMR

(176 MHz, CDCl3) δ 154.3 (NCO2Me), 141.8 (ArC, weak signal, confirmed by HMBC),

136.1 (NCH2-CH=CH2), 136.0 (ArCH, TBDPS), 135.9 (ArCH, TBDPS), 134.2 (ArC),

133.5 (ArC, TBDPS), 133.4 (ArC, TBDPS), 123.0 (ArCH, TBDPS), 129.9 (ArCH,

TBDPS), 127.9 (ArCH, TBDPS), 127.9 (ArCH, TBDPS), 127.7 (ArCH), 123.8 (ArCH),

123.1 (ArCH), 117.7 (NCH2-CH=CH2), 115.6 (ArCH), 72.8 (N-C-CH-N), 70.3

(CH2OH), 66.8 (CH2OTBDPS), 64.1 (N-C-CH-N), 60.4 (N-CH2-CH=CH2), 52.4

(NCO2CH3), 50.9 (C(CH2O)), 47.6 (ArC-CH), 44.6 (N-CH2CH2), 36.4 (C-CH2CH2-C),

27.8 (N-CH2CH2), 27.6 (C-CH2CH2-C), 27.1 (C(CH3)3, TBDPS), 19.4 (C(CH3)3,

TBDPS). S75

Compound 572

To a stirred solution of alcohol 569 (327 mg, 0.54 mmol) in EtOAc (10.0 mL) was added

IBX (450 mg, 1.61 mmol) and the resulting mixture was stirred at 80 oC for 2 h. After cooling to room temperature, solids were filtered, and filtrate was concentrated by rotary evaporation. The residue was dissolved in CH2Cl2 (10.0 mL), followed by addition of silica gel (3.00 g) and the mixture was stirred at room temperature for 2 h. Silica gel was filtered, and the filtrate was concentrated by rotary evaporation to give 570 (296 mg, 92%,

93:7 dr by 1H NMR) as a colourless oil which was used directly without purification.

1 Rf (20% EtOAc in hexane): 0.63; H NMR (700 MHz, CDCl3) major  9.69 (1H, s,

CHO), 7.70—7.61 (5H, m, SiPh2-t-Bu), 7.48—7.35 (6H, m, SiPh2-t-Bu and ArH), 7.21—

7.05 (2H, m, ArH), 6.98 (1H, td, J = 7.3, 1.0 Hz, ArH), 5.68—5.51 (1H, m, NCH2-

CH=CH2), 5.04—4.87 (2H, m, NCH2-CH=CH2), 4.34 (1H, brs, N-CH-C-N), 4.23 (1H, d, J = 10.2 Hz, HCH-OTBDPS), 3.73 (1H, d, J = 10.6 Hz, HCH-OTBDPS), 3.63 (3H, s,

NCO2CH3), 3.13 (1H, dd, J = 13.9, 6.7 Hz, N-HCH-CH=CH2), 3.09—3.02 (2H, m, N-

HCH-CH=CH2 and ArC-CH), 2.53 (1H, ddd, J = 12.2, 8.9, 2.8 Hz, N-HCH-CH2), 2.32—

2.24 (2H, m, N-HCH-CH2 and C-HCH-CH2-C), 2.12—1.98 (1H, m, C-HCH-CH2-C), S76

1.75—1.60 (2H, m, C-CH2-HCH-C and N-CH2-HCH), 1.58—1.46 (2H, m, C-CH2-HCH-

C and N-CH2-HCH), 1.04 (9H, s, SiPh2-t-Bu);

To a solution of aldehyde 570 (180 mg, 0.296 mmol) in PhMe (4.50 mL) under N2 atmosphere was added Petasis reagent (Cp2TiMe2) (15% w/w, 2.00 mLm 1.48 mmol) and the solution was stirred at 80 oC in dark until full consumption of starting material

(checked by TLC, 2 h). After cooling to room temperature, the reaction mixture was diluted with PhMe (5.00 mL), followed by addition of Hoveyda-Grubbs 2nd generation catalyst (17.0 mg, 0.03 mmol) and the resulting mixture was stirred at 80 oC for 2 h. A

NaHCO3 aqueous solution and MeOH (0.50 mL each) was added he mixture was stirred at 80 oC for another 2 h to decompose metal reagents. After cooling to room temperature, precipitates were filtered and washed with EtOAc (10 mL x 2). Combined filtrates were concentrated by rotary evaporation and chromatography of residue on silica gel, using

30% EtOAc in hexane, gave compound 572 (110 mg, 71% over 2 steps) as a yellow oil.

25 -1 Rf (100% EtOAc) 0.7; []D +33.7 (c 0.66, CHCl3); FTIR (neat, cm ) 2927, 2853, 1704,

1479, 1438, 1349, 1320, 1229, 1107, 1082, 750, 701; HRMS (ESI-TOF, m/z) calcd. for

+ 1 C36H43N2O3Si [M+H] 579.3037, found 579.3015; H NMR (700 MHz, CDCl3)  7.67—

7.60 (4H, m, ArH), 7.45—7.31 (7H, m, ArH), 7.17 (1H, t, J = 7.7 Hz, ArH), 7.13 (1H, d,

J = 7.4 Hz, ArH), 7.06 (1H, t, J = 7.6 Hz), 5.84—5.67 (2H, m, NCH2-CH=CH-C), 3.71

(1H, d, J = 9.9 Hz, C-HCH-OTBDPS), 3.58 (3H, s, NCO2Me), 3.51 (1H, d, J = 9.8 Hz,

C-HCH-OTBDPS), 3.26 (1H, brs, ArC-CH), 3.01 (1H, dd, J = 15.7, 5.5 Hz, N-HCH-

CH=CH-C), 3.06 (1H, brs, N-CH-C-N); 2.79 (1H, brs, N-HCH-CH=CH-C), 2.67—2.62

(1H, m, N-HCH-CH2), 2.50 (1H, brs, N-HCH-CH2), 2.32—1.62 (6H, N-CH2CH2 and C-

13 CH2-CH2-C); C NMR (176 MHz, CDCl3)  153.7 (NCO2Me), 143.1 (C), 135.9 (CH),

135.9 (CH), 135.2 (C), 135.0 (CH), 134.2 (C), 134.0 (C), 129.6 (CH), 129.6 (CH), 127.7

(CH), 127.6 (CH), 127.1 (CH), 123.9 (CH), 123.2 (CH), 122.7 (CH), 119.1 (CH), 80.0

S77

(C), 69.2 (CH2), 67.1 (CH), 52.9 (CH2), 52.3 (CH), 49.8 (C), 48.5 (CH2), 45.2 (CH),32.6

(CH2), 30.3 (CH2), 27.0 (CH3), 19.5 (C). Some carbon signals are missing from spectra and the core structure were furtherly confirmed by subsequent reaction.

Compound 574

To a solution of 572 (97.0 mg, 0.168 mmol) in THF (2.30 mL) was added TBAF (1.0 M

o in THF, 1.68 mL) and the resulting solution was stirred at 70 C under N2 atmosphere 3 h. After cooling to room temperature, THF was removed by rotary evaporation and residue was added sat. NH4Cl aq. (15 mL) and extracted with EtOAc (15 mL x 2).

Combined organic layers were dried (Na2SO4) and concentrated by rotary evaporation.

Chromatography of residue on silica gel, using 50% EtOAc in hexane, gave compound

574 as a colourless oil (20 mg, 40%, low yield may due to insufficient extraction).

25 -1 Rf (30% EtOAc in hexane): 0.3; []D -24.5 (c 0.22, CHCl3); FTIR (neat, cm ) 3024,

2931, 1701, 1478, 1459, 1396, 1358, 1221, 1061, 1025, 752; HRMS (ESI-TOF, m/z)

+ 1 clacd. For C19H21N2O2 [M+H] 309.15976, found 309.15975; H NMR (700 MHz,

CDCl3)  7.77 (1H, d, J = 8.1 Hz, ArH), 7.25—7.22 (1H, m, ArH), 7.15 (1H, dt, J = 7.3,

1.5 Hz, ArH), 7.07 (1H, td, J = 7.4, 1.0 Hz, ArH), 5.74 (1H, ddd, J = 10.0, 4.9, 1.8 Hz,

NCH2-CH=CH-C), 5.43 (1H, ddd, J = 10.0, 2.9, 1.6 Hz, NCH2-CH=CH-C), 4.02 (1H, d,

J = 12.1 Hz, NCO2-HCH-C), 3.88 (1H, dd, J = 12.1, 1.5 Hz, NCO2-HCH-C), 3.46 (1H, d, J = 6.0 Hz, ArC-CH), 3.24 (1H, ddd, J = 16.7, 4.9, 1.6 Hz, N-HCH-CH=CH-C), 2.70—

2.63 (2H, m, N-HCH-CH=CH-C and N-HCH-CH2), 2.55 (1H, ddd, J = 14.8, 9.8, 7.0 Hz,

S78

C-HCH-CH2-C), 2.40 (1H, ddd, J = 15.1, 12.0, 3.3 Hz, C-HCH-CH2-C), 2.32—2.28 (1H, m, NCH2-HCH), 2.25—2.15 (2H, m, C-CH2-HCH and NCH2-HCH), 2.10—1.99 (3H, m,

13 N-CH-C-N, C-CH2-HCH-C and N-HCH-CH2); C NMR (176 MHz, CDCl3) δ 154.7

(NCO2), 142.3 (ArC), 131.5 (ArC), 128.3 (ArCH), 127.1 (NCH2-CH=CH-C), 127.0

(NCH2-CH=CH-C), 123.8 (ArCH), 122.4 (ArCH), 115.4 (ArCH), 77.8 (NCO2CH2), 75.6

(N-CH-C-N), 73.7 (N-CH-C-N), 53.6 (N-CH2-CH=CH-C), 50.3 (NCO2-CH2-C), 48.7

(N-CH2CH2), 43.4 (ArC-CH), 36.5 (C-CH2CH2-C), 34.4 (C-CH2CH2-C), 23.1

(NCH2CH2).

Compound 581

To a solution of 574 (4.00 mg, 0.013 mmol) in nEtOH (1.00 mL) was added KOH powder

(14.6 mg, 0.26 mmol) and the resulting mixture was heated to 90 oC and stirred for 1 h.

After cooling to room temperature, the mixture was diluted with EtOAc (5.0 mL), washed with water (5.0 mL), dried (Na2SO4) and passed through a short pad of silica gel, the filtrate was concentrated to give 581 as a colourless oil (3.00 mg, 87%).

25 -1 Rf (100% EtOAc): 0.28; []D -24.6 (c 0.13, CHCl3); FTIR (neat, cm ) 3264, 2923,

2851, 1604, 1461, 1258, 1132, 1051, 1020, 750; HRMS (ESI-TOF, m/z) calcd. for

+ 1 C18H23N2O [M+H] 283.1810, found 283.1812; H NMR (700 MHz, CDCl3)  7.09 (1H, tt, J = 7.6, 1.2 Hz, ArH), 7.06 (1H, dt, J = 7.3, 1.4 Hz, ArH), 6.91 (1H, td, J = 7.4, 1.0 Hz

ArH), 6.80 (1H, d, J = 7.7 Hz, ArH), 5.94 (1H, ddd, J = 9.8, 6.2, 1.7 Hz, NCH2-CH=CH),

5.51 (1H, ddd, J = 9.8, 2.7, 0.9 Hz, NCH2-CH=CH), 4.45 (2H, brs, NH and OH), 3.43

S79

(1H, d, J = 10.8 Hz, HCH-OH), 3.40 (1H, brs, Ar-CH), 3.36 (1H, d, J = 10.8 Hz, HCH-

OH), 2.98 (1H, dd, J = 15.6, 6.2 Hz, N-HCH-CH=CH), 2.67—2.61 (1H, m, N-HCH-

CH2), 2.57 (1H, dt, J = 15.8, 2.1 Hz, N-HCH-CH=CH), 2.51 (1H, ddd, J = 13.1, 11.4, 9.6

Hz, C-HCH-CH2-C), 2.29 (1H, s, N-CH-C-N), 2.25—2.16 (3H, m, N-HCH-CH2 and N-

CH2CH2), 2.13 (1H, ddd, J = 13.3, 11.4, 8.0 Hz, C-CH2-HCH-C), 1.82 (1H, ddt, J = 13.1,

13 8.0, 1.5 Hz, C-HCH-CH2-C), 1.75 (1H, ddd, J = 13.4, 9.6, 1.7 Hz, C-CH2-HCH-C); C

NMR (176 MHz, CDCl3)  148.9 (ArC), 135.0 (ArC), 132.5 (CH2-CH=CH-C), 127.7

(ArCH), 126.8 (CH2-CH=CH-C), 122.8 (ArCH), 121.5 (ArCH), 113.6 (ArCH), 74.2 (N-

C-CH-N), 73.0 (N-C-CH-N), 69.8 (CH2-OH), 51.9 (N-CH2-CH=CH-C)), 50.3 (C-

CH2OH), 50.2 (N-CH2CH2), 44.5 (ArC-CH), 34.8 (C-CH2CH2-C), 31.2 (C-CH2-CH2-C),

24.3 (N-CH2CH2).

Compound 582

To a round bottom flask charged with aldehyde 570 (915 mg, 1.50 mmol) under N2 atmosphere was added Petasis reagent (13.5% w/w, 12.0 mL) and the resulting solution was stirred at 80 oC in dark until full consumption of starting material (checked by TLC).

o After cooling to 60 C, aq. MeOH (MeOH:H2O 9:1, 5.00 mL) was added the resulting mixture stirred at 60 oC for 2 h to decompose excess Petasis reagent. The brown

S80

precipitates were filtered through a pad of silica gel and then washed with EtOAc (30.0 mL x 2). The filtrate was concentrated by rotary evaporation and the crude 571 was put into next step directly.

A solution of TBAF (1.0 M, 15.0 mL) was added to above crude alkene 571 and the resulting solution was stirred at 70 oC for 3 h. After cooling to room temperature, THF was removed by rotary evaporation and the residue was added NaClO4 aq. and stirred for

5 min before extraction with EtOAc. The combined organic layers were washed with brine, dried (Na2SO4) and concentrated to ~10 mL. Then 20% EtOAc in hexane (100 mL) was added and the resulting precipitates (n-Bu4N+X-) was filtered and the filtrate was concentrated. Chromatography of the residue on silica gel, using 20% EtOAc in hexane, gave alcohol 582 as a yellow oil (361 mg, 65% over 2 steps). 1H NMR indicated a mixture of epimers (dr 96:4).

-1 Rf (20% EtOAc in hexane) 0.22; FTIR (neat, cm ) 3465, 2947, 2867, 1699, 1481, 1440,

+ 1361, 1060, 915; HRMS (ESI-TOF, m/z) calcd. for C22H29N2O3 [M+H] 369.2173, found

1 369.2172; H NMR (700 MHz, CDCl3): mixture of diastereomers, dr 94:6; major isomer:  7.70 (1H, brs, ArH), 7.18 (1H, td, J = 7.8, 1.4 Hz, ArH), 7.12 (1H, dt, J = 7.4,

0.7 Hz, ArH), 6.99 (1H, td, J = 7.3, 1.0 Hz, ArH), 6.33 (1H, dd, J = 17.9, 11.0 Hz, C-

CH=CH2), 5.76 (1H, dddd, J = 17.0, 10.1, 6.8, 6.1 Hz, NCH2-CH=CH2), 5.28 (1H, dd, J

= 10.9, 1.6 Hz, C-CH=HCH), 5.17 (1H, dd, J = 17.9, 1.7 Hz, C-CH=HCH), 5.13—4.98

(2H, m, NCH2-CH=CH2), 4.34 (1H, s, N-CH-C-N), 3.87 (3H, s, OCH3), 3.61 (1H, d, J =

10.4 Hz, C-HCH-OH), 3.45 (1H, d, J = 10.4 Hz, C-HCH-OH), 3.32 (1H, ddt, J = 13.8,

6.8, 1.3 Hz, N-HCH-CH=CH2), 3.19 (1H, ddt, J = 13.8, 6.1, 1.5 Hz, N-HCH-CH=CH2),

3.01 (1H, dd, J = 8.6, 6.2 Hz, Ar-CH), 2.88 (1H, ddd, J = 13.0, 10.0, 3.0 Hz, N-HCH-

CH2), 2.57 (1H, brs, OH), 2.46 (1H, ddd, J = 12.9, 6.3, 3.5 Hz, N-HCH-CH2), 2.35—2.25

(1H, m, C-HCH-CH2-C), 1.93 (1H, ddd, J = 12.0, 7.6, 3.3 Hz, C-CH2-HCH-C), 1.74—

S81

1.61 (3H, m, C-HCH-CH2-C, C-CH2-HCH-C and N-CH2-HCH), 1.51 (1H, dddd, J =

13 13.8, 9.9, 8.6, 3.5 Hz, N-CH2-HCH); C NMR (176 MHz, CDCl3)  154.3 (NCO2Me),

142.8 (C-CH=CH2), 142.0 (ArC), 137.0 (NCH2-CH=CH2), 134.3 (ArC), 127.7 (ArCH),

123.9 (ArCH), 123.1 (ArCH), 117.2 (NCH2-CH=CH2), 115.7 (ArCH), 114.5 (C-

CH=CH2), 72.4 (N-C-CH-N), 69.3 (C-CH2OH), 64.2 (N-CH-C-N), 59.7 (N-

CH2CH=CH2), 52.5 (C-CH2OH), 52.1 (NCO2CH3), 46.4 (ArC-CH), 44.4 (N-CH2CH2),

34.1 (C-CH2CH2-C), 27.6 (C-CH2CH2-C), 26.4 (NCH2CH2).

Compound 587

o To a cooled (0 C) and stirred solution of alcohol 582 (100 mg, 0.27 mmol) in CH2Cl2

(10.0 mL) under N2 atmosphere was added Et3N (41 L, 0.40 mmol) and MsCl (37 L,

0.40 mmol). Cooling bath was left but not recharged and the stirring was continued until full consumption of starting material (1.5 h, checked by TLC). Saturated NH4Cl aq.

Solution (10 mL) was added and organic layer was separated, dried (Na2SO4) and concentrated by rotary evaporation to give 587 (122 mg, quant.) which was pure enough to go to the next step.

25 -1 Rf (100% CH2Cl2) 0.4; []D +52.0 (c 0.2, CHCl3); FTIR (neat, cm ) 2952, 2932, 1699,

1481, 1442, 1357, 1329, 1174, 954, 757; HRMS (ESI-TOF, m/z) calcd. for C23H31N2O5S

+ 1 [M+H] 447.1954, found 447.1947; H NMR (700 MHz, CDCl3)  7.69 (1H, brs, ArH),

7.18 (1H, t, J = 7.8 Hz, ArH), 7.11 (1H, d, J = 6.7 Hz, ArH), 6.99 (1H, td, J = 7.4, 1.0 Hz,

ArH), 6.26 (1H, dd, J = 17.8, 11.0 Hz, C-CH=CH2), 5.73 (1H, ddt, J = 16.8, 10.6, 6.4 Hz, S82

NCH2-CH=CH2), 5.27 (1H, dd, J = 10.9, 1.2 Hz, C-CH=HCH), 5.16 (1H, dd, J = 17.9,

1.3 Hz, C-CH=HCH), 5.09—5.00 (2H, m, NCH2-CH=CH2), 4.39 (1H, s, N-CH-C-N),

4.18 (1H, d, J = 9.2 Hz, C-HCH-OMs), 4.14 (1H, d, J = 9.1 Hz, C-HCH-OMs), 3.87 (3H, s, NCO2Me), 3.28 (1H, dd, J = 14.1, 6.6 Hz, N-HCH-CH=CH2), 3.19 (1H, dd, J = 13.9,

6.4 Hz, N-HCH-CH=CH2) 3.03 (3H, s, CH2OSO2Me), 3.00 (1H, t, J = 6.8 Hz, ArC-CH),

2.80 (1H, ddd, J = 12.2, 8.8, 3.2 Hz, N-HCH-CH2), 2.45—2.33 (2H, m, N-HCH-CH2 and

C-HCH-CH2-C), 1.90 (1H, dd, J = 9.1, 5.2 Hz, C-CH2-HCH-C), 1.81—1.49 (4H, m, N-

13 CH2-CH2, C-HCH-CH2-C and C-CH2-HCH-C); C NMR (176 MHz, CDCl3)  154.1

(NCO2Me), 141.9 (C-CH=CH2), 141.1 (ArC), 136.7 (NCH2-CH=CH2), 133.9 (ArC),

127.8 (ArCH), 123.9 (ArCH), 123.1 (ArCH), 117.1 (NCH2-CH=CH2), 115.6 (ArCH),

115.0 (C-CH=CH2), 75.2 (N-C-CH-N), 72.9 (C-CH2OMs), 63.2 (N-CH-C-N), 60.0 (N-

CH2CH=CH2), 52.7 (C-CH2OMs), 52.5 (NCO2CH3), 50.4 (C-CH2OSO2Me), 46.5 (ArC-

CH), 44.2 (N-CH2CH2), 37.5 (C-CH2CH2-C), 27.7 (C-CH2CH2-C), 27.0 (NCH2CH2).

Compound 588

To a solution of 587 (122 mg, 0.24 mmol) in PhMe (12.0 mL) under N2 atmosphere was added Hoveyda-Grubbs 2nd generation catalyst (7.7 mg, 0.024 mmol). The mixture was then stirred at 60 oC for 4 h. After cooling to room temperature, solvent was removed by rotary evaporation and chromatography of residue on silica gel, using pure Et2O gave 588 as a light brown oil (80 mg, 80%).

S83

25 -1 Rf (100% Et2O) 0.16; []D +50.5 (c 0.19, CHCl3); FTIR (neat, cm ) 2929, 2851, 1701,

1481, 1462, 1441, 1354, 1328, 1175, 955; HRMS (ESI-TOF, m/z) calcd. for C21H27N2O5S

1 419.1641, found 419.1649; H NMR (700 MHz, CDCl3)  7.44 (1H, d, J = 8.1 Hz, ArH),

7.20 (1H, t, J = 7.4 Hz, ArH), 7.13 (1H, d, J = 7.5 Hz, ArH, 7.08 (td, J = 7.4, 1.0 Hz,

ArH), 5.83 (1H, ddd, J = 10.1, 5.7, 1.8 Hz, NCH2-CH=CH-C), 5.74 (1H, dd, J = 10.1, 2.5

Hz, NCH2-CH=CH-C), 4.32 (1H, d, J = 9.4 Hz, HCH-OMs), 4.17 (1H, d, J = 9.5 Hz,

HCH-OMs), 3.80 (3H, s, NCO2Me), 3.28 (1H, brs, ArC-CH), 3.03 (1H, dd, J = 16.1, 5.8

Hz, N-HCH-CH=CH-C), 3.00 (3H, s, CH2OSO2Me), 2.93 (1H, brs, N-CH-C-N), 2.76

(1H, d, J = 16.2 Hz, N-HCH-CH=CH-C), 2.66—2.58 (2H, m, N-HCH-CH2 and C-HCH-

CH2-C), 2.37—2.14 (4H, m, N-HCH-CH2, C-HCH-CH2-C, C-CH2-HCH-C, N-CH2-

13 HCH), 2.02—1.95 (1H, m, NCH2-HCH), 1.85—1.76 (1H, m, C-CH2-HCH-C); C NMR

(176 MHz, CDCl3)  154.1 (NCO2Me), 142.6 (ArC), 134.8 (ArC), 131.7 (NCH2-

CH=CH-C), 127.4 (ArCH), 124.8 (NCH2-CH=CH-C), 124.2 (ArCH), 122.8 (ArCH),

119.4 (ArCH), 79.9 (N-CH-C-N) 74.9 (CH2OMs), 67.1 (N-CH-C-N), 52.7 (NCO2Me),

52.5 (N-CH2CH=CH-C), 48.3 (N-CH2CH2), 47.6 (C-CH2OMs) 45.3 (ArC-CH), 37.4

(CH2OSO2Me), 32.7 (C-CH2-CH2-C), 30.4 (C-CH2-CH2-C), 25.7 (N-CH2CH2).

Compound 590

To a stirred solution of 588 (50.0 mg, 0.12 mmol) in DMSO (2.50 mL) was added KCN

(65.0 mg) and the resulting mixture was stirred at 100 oC for 24 h. After cooling to room temperature, the reaction mixture was added EtOAc (15.0 mL) and water (15.0 mL).

S84

Aqueous layer was extracted with EtOAc (15.0 mL x 2) and combined organic layers were washed with brine (20.0 mL x 2), dried (Na2SO4) and concentrated by rotary evaporation. 1H NMR indicated a 1:1 mixture of two compounds. Chromatography of crude material on silica gel, using 10% acetone in hexane, gave 574 (12.0 mg, 32%) and

590 (10.0 mg, 31%) as colourless oil.

25 -1 Rf (30% EtOAc in hexane): 0.4; []D -9.7 (c 0.10, CHCl3); FTIR (neat, cm ) 2925,

2853, 1698, 1669, 1602, 1449, 1285, 1239, 1099, 756, 746; HRMS (ESI-TOF, m/z) calcd.

+ 1 for C18H21N2 [M+H] 265.1699, found 265.1701; H NMR (700 MHz, CDCl3)  7.13—

7.07 (2H, m, ArH), 6.82 (1H, td, J = 7.4, 1.0 Hz, ArH), 6.59 (1H, dt, J = 7.4, 1.0 Hz,

ArH), 5.96 (1H, ddd, J = 9.9, 3.1, 1.9 Hz, NCH2-CH=CH-C), 5.47 (1H, ddd, J = 9.9, 4.0,

2.1 Hz, NCH2-CH=CH-C), 3.55 (1H, d, J = 9.1 Hz, N-HCH-C), 3.48 (1H, ddd, J = 16.9,

4.1, 1.9, N-HCH-CH=CH-C), 3.20 (1H, dd, J = 5.2, 1.9 Hz, ArC-CH), 2.94 (1H, ddd, J

= 11.3, 5.1, 2.2 Hz, N-HCH-CH2), 2.90 (1H, dd, J = 9.1, 3.0 Hz, N-HCH-C), 2.43 (1H, ddd, J = 16.8, 3.1, 2.2, N-HCH-CH=CH-C), 2.36 (1H, td, J = 12.2, 4.5 Hz, C-HCH-CH2-

C), 2.29—2.15 (2H, m, NCH2CH2), 2.04—1.97 (1H, m, C-CH2-HCH-C), 1.94—1.79

13 (3H, m, N-HCH-CH2, C-HCH-CH2-C and C-CH2-HCH-C), 1.60 (1H, s, N-CH-C-N); C

NMR (176 MHz, CDCl3)  157.8 (ArC), 134.5 (ArC), 128.1 (NCH2-CH=CH-C), 127.9

(ArCH), 126.1, (NCH2-CH=CH-C) 124.3 (ArCH), 120.1 (ArCH), 112.4 (ArCH), 76.3

(N-C-CH-N), 70.3 (N-C-CH-C), 64.7 (N-CH2-C), 54.2 (N-CH2-CH=CH-C), 53.5 (N-

CH2CH2), 49.0 (N-CH2-C), 37.8 (ArC-CH), 36.3 (C-CH2-CH2-C), 29.0 (C-CH2-CH2-C),

25.1 (NCH2CH2).

S85

Compound 594

o To a cooled (-78 C) and stirred solution of DMSO (0.25 mL, 3.58 mmol) in CH2Cl2 (5.00 mL) was added (COCl)2 (0.15 mL, 1.79 mmol) dropwise and the resulting solution was

o stirred -78 C for 30 min. Then a solution of 582 (150 mg, 0.407 mmol) in CH2Cl2 (3.00

o mL) was added dropwise and the mixture was stirred at -78 C for 30 min before Et3N

(0.60 mL, 14.07 mmol) was added dropwise and the mixture was slowly warmed to 0 oC during which time the TLC indicated full consumption of starting material. The reaction mixture was quenched with sat. NH4Cl aq. (20.0 mL) and aqueous layer was separated and extracted with CH2Cl2 (15.0 mL). Combined organic layers were dried (Na2SO4) and concentrated to give aldehyde 593 as a mixture of epimers (2.5:1, checked by 1H NMR) which was put into next step directly.

o To a cooled (-78 C) and stirred suspension of Ph3PCH3Br (2.00 g, 5.60 mmol) in THF

(7.50 mL) under N2 atmosphere was added NaHMDS (2.0 M in THF, 2.50 mL) and the resulting yellow mixture was stirred at -78 oC for 1.5 h and the solution of crude 593 in

THF (2.50 mL) was then added. The mixture was slowly warmed to room temperature and stirred for another 1 h. Saturated NH4Cl aq. solution (20.0 mL) was added to quench the reaction at 0 oC and the mixture was extracted with EtOAc (15.0 mL x 3). Combined

S86

organic layers were dried (Na2SO4) and concentrated by rotary evaporation.

Chromatography of residue on silica gel, using 10% EtOAc in hexane, gave 594 (112 mg,

76% over 2 steps) as a pale-yellow oil.

25 -1 Rf (10% EtOAc in hexane): 0.43; []D +112 (c 0.10, CHCl3); FTIR (neat, cm ) 3068,

2947, 2926, 1699, 1480, 1438, 1359, 1108, 749, 702; HRMS (ESI-TOF, m/z) calcd. for

+ 1 C23H29N2O2 [M+H] 365.2229, found 365.2231; H NMR (700 MHz, CDCl3)  7.70 (1H, brs, ArH), 7.17 (1H, td, J = 7.7, 1.4 Hz, ArH), 7.10 (1H, d, J = 7.9 Hz, ArH), 6.98 (1H, td, J = 7.4, 1.0 Hz, ArH), 6.32 (1H, dd, J = 17.9, 10.9 Hz, C-CH=CH2), 6.08 (1H, dd, J =

17.0, 10.9 Hz, C-CH=CH2), 5.72 (1H, ddt, J = 16.8, 10.5, 6.4 Hz, NCH2-CH=CH2), 5.20

(1H, dd, J = 10.9, 1.5 Hz, C-CH=HCH), 5.09 (1H, dd, J = 17.9, 1.5 Hz, C-CH=HCH),

5.05—4.97 (4H, m, C-CH=CH2 and NCH2-CH=CH2), 4.47 (1H, s, N-CH-C-N), 3.86 (3H, s, NCO2CH3), 3.29 (1H,dd, J = 14.2, 6.3 Hz, N-HCH-CH=CH2), 3.20 (1H, dd, J = 14.0,

6.5 Hz, N-HCH-CH=CH2), 3.05 (1H, t, J = 6.5 Hz, ArC-CH), 2.83 (1H, ddd, J = 11.8,

8.1, 3.3 Hz, N-HCH-CH2), 2.40 (1H, ddd, J = 11.9, 8.1, 3.3 Hz, N-HCH-CH2), 2.36—

2.27 (1H, m, C-HCH-CH2-C), 2.06—1.98 (1H, m, C-HCH-CH2-C), 1.96—1.85 (2H, m,

C-CH2-HCH-C and NCH2-HCH), 1.78 (1H, dt, J = 13.1, 7.2 Hz, C-CH2-HCH-C), 1.53

13 (1H, dddd, J = 13.9, 8.1, 6.8, 3.4 Hz, NCH2-HCH); C NMR (176 MHz, CDCl3) δ 154.3

(NCO2Me), 146.8 (C-CH=CH2), 142.5 (C-CH=CH2), 142.1 (ArC), 136.9 (NCH2-

CH=CH2), 134.4 (ArC), 127.6 (ArCH), 123.80 (ArCH), 123.0 (ArCH), 116.7 (C-

CH=CH2), 115.5 (ArCH), 112.5 (C-CH=CH2), 111.6 (NCH2-CH=CH2), 73.4 (N-C-CH-

N)), 68.7 (N-C-CH-N), 60.7 (NCH2-CH=CH2), 54.5 (C(CH=CH2)2), 52.3 (NCO2CH3),

47.7 (ArC-CH), 44.1 (NCH2CH2), 36.9 (C-CH2CH2-C), 30.6 (C-CH2CH2-C), 28.1

(NCH2CH2).

S87

Compound 595

To a stirred solution of 594 (100 mg, 0.27 mmol) in PhMe (5.00 mL) under N2 was added

Hoveyda-Grubbs catalyst (17.0 mg, 0.027 mmol). The resulting solution was degassed

o o and backed filled with N2 and then stirred at 80 C for 1 h. After cooling to 35~40 C,

DMSO (50 eq. to catalyst, 0.10 mL) was added and the mixture was stirred at this temperature for 4 h. The solvent was removed by rotary evaporation and chromatography of the residue on silica gel, using 30% EtOAc in hexane, gave 595 as a colourless oil

(70.0 mg, 76%).

25 -1 Rf (20% EtOAc in hexane) 0.20; []D +174 (c 0.07, CHCl3); FTIR (neat, cm ) 2932,

2847, 1705, 1477, 1459, 1437, 1367, 1356, 1321, 1229, 1116, 754; HRMS (ESI-TOF,

+ 1 m/z) calcd. for C21H25N2O2 [M+H] 337.1916, found 337.1913; H NMR (700 MHz,

CDCl3)  7.56 (H, d, J = 8.0 Hz, ArH), 7.19 (1H, tt, J = 8.2, 1.1 Hz, ArH), 7.13—7.10

(1H, m, ArH), 7.06 (1H, td, J = 7.4, 1.0 Hz, ArH), 5.95 (1H, dd, J = 17.6, 10.7 Hz, C-

CH=CH2), 5.58 (1H, ddd, J = 9.8, 5.5, 1.6 Hz, NCH2-CH=CH-C), 5.54—5.50 (1H, m,

NCH2-CH=CH-C), 4.97 (1H, dd, J = 17.9, 0.9 Hz, C-CH=HCH), 4.93 (1H, dd, J = 10.7,

0.9 Hz, C-CH=HCH), 3.73 (3H, s, NCO2CH3), 3.34 (1H, t, J = 4.4 Hz, ArC-CH), 3.02

(1H, dd, J = 16.0, 5.6 Hz, N-HCH-CH=CH), 2.79 (1H, ddd, J = 13.3, 8.3, 3.4 Hz, C-

HCH-CH2-C), 2.70 —2.56 (3H, m, N-CH-C-N, N-HCH-CH=CH and N-HCH-CH2), 2.50

(1H, dt, J = 12.8, 7.9 Hz, C-CH2-HCH-C), 2.42 (1H, ddd, J = 13.5, 10.6, 6.8 Hz, C-HCH-

CH2-C), 2.28—2.21 (1H, m, NCH2-HCH), 2.19—2.14 (1H, m, N-HCH-CH2), 2.12—

13 2.08 (1H, m, NCH2-HCH), 1.98 (1H, ddd, J = 12.7, 10.7, 3.8 Hz, C-CH2-HCH-C); C

S88

NMR (176 MHz, CDCl3)  153.5 (NCO2Me), 146.4 (C-CH=CH2), 143.5 (ArC), 134.6

(ArC), 134.5 (NCH2-CH=CH-C), 127.3 (ArCH), 123.7 (ArCH), 122.5 (ArCH), 120.1

(NCH2-CH=CH-C), 118.7 (ArCH), 111.4 (C-CH=CH2), 80.5 (N-CH-C-N), 72.4 (N-CH-

C-N), 52.9 (N-CH2-CH=CH-C), 52.1 (NCO2CH3), 50.4 (C-CH=CH2), 49.0 (N-CH2CH2),

44.2 (ArC-CH), 36.5 (C-CH2CH2-C), 29.9 (C-CH2CH2-C), 24.0 (NCH2CH2).

Compound 604

To a stirred solution of MeCN (1.0 mL) and H2O (0.13 mL) charged with Pd(OAC)2 (5.60 mg, 0.025 mmol), 14,-benzoquinone (16.1 mg, 0.15 mmol) and HClO4 (1.0 M in H2O,

0.056 mL, 0.056 mmol) was added a solution of 595 (6.00 mg, 0.018 mmol) in PhMe

(0.20 mL) and the mixture was stirred at room temperature for 5 days. The mixture was quenched with sat. NaHCO3 aqueous solution (5.0 mL) and extracted with EtOAc (5.0 mL x 3). Combined organic layers were dried (Na2SO4) and concentrated. Preparative

TLC of the residue, eluted with 50% EtOAc in hexane, gave compound 604 (1.00 mg,

16% yield) as a colorless oil.

-1 Rf (30% EtOAc in hexane) 0.13; FTIR (neat, cm ) 2924, 2849, 1701, 1700, 1480, 1438,

+ 1363, 1231, 1117, 756; HRMS (ESI-TOF, m/z) calcd. for C21H25N2O3 [M+H] 353.1865,

1 found 353.1862; H NMR (700 MHz, CDCl3)  7.51 (1H, d, J = 8.0 Hz, ArH), 7.21 (1H, tt, J = 7.7, 1.2 Hz, ArH), 7.12 (1H, dt, J = 7.3, 1.5 Hz, ArH), 7.06 (1H, td, J = 7.4, 1.0 Hz,

ArH), 5.82 (1H, ddd, J = 9.9, 5.9, 1.7 Hz, NCH2-CH=CH-C), 5.63 (1H, dd, J = 9.8, 2.8

Hz, NCH2CH=CH-C), 3.72 (3H, s, NCO2Me), 3.37—3.34 (1H, m, ArC-CH), 3.21 (1H, S89

s, N-CH-C-N), 3.13—3.01 (2H, m, N-HCH-CH=CH-C and C-HCH-CH2-C), 2.72—2.58

(3H, m, N-HCH-CH=CH-C, N-HCH-CH2 and C-CH2-HCH-C), 2.39 (1H, ddd, J = 14.0,

10.9, 6.6 Hz, C-CH2-HCH-C), 2.27—2.11 (6H, m, C(=O)Me, NCH2CH2 and N-HCH-

13 CH2), 1.98 (1H, ddd, J = 13.1, 10.9, 3.7, C-HCH-CH2-C); C NMR (176 MHz, CDCl3)

 208.4 (C(=O)Me), 153.7 (NCO2Me), 142.9 (ArC), 134.0 (ArC), 128.8 (NCH2-CH=CH-

C), 127.6 (ArCH), 126.1 (NCH2-CH=CH-C), 123.7 (ArCH), 122.4 (ArCH), 118.5

(ArCH), 80.0 (N-CH-C-N), 67.7 (N-CH-C-N), 64.4 (C-COMe), 52.8 (N-CH2-CH=CH-

C), 52.4 (NCO2CH3), 49.1 (NCH2CH2), 44.2 (ArC-CH), 34.3 (C-CH2-CH2-C), 30.6 (C-

CH2-CH2-C), 25.9 (C(=O)CH3), 23.8 (NCH2CH2).

Compound 387

Procedure: To a cooled (0 oC) and stirred solution of 595 (15.0 mg, 0.045 mmol) in THF

(1.00 mL) was added LiALH4 (1.0 M in THF, 0.13 mL) and the mixture was heated to 65 oC for 30 min. After cooling back to 0 oC, the reaction mixture was quenched with water

(3 drops) and 15% NaOH aq. (3 drops). Stirring was continued for 10 min and EtOAc

(5.00 mL) and Na2SO4 were added with another 10 min stirring. Precipitates were filtered and the filtrate was concentrated by rotary evaporation and chromatography (pipet) of the residue on silica gel, using 20% EtOAc in hexane, gave compound 387 (10.2 mg, 80%) as a colourless oil.

S90

Compound 543

Preparation of c-Hex2BH: To a round bottom flask charged with anhydrous THF

. o (5.50 mL) under N2 was added BH3 Me2S (5.0 M in Et2O, 0.38 mL) and cooled to 0 C.

Cyclohexene (0.42 mL) was added slowly and the resulting mixture was stirred at 0 oC for 15 min and then 1 h at room temperature to give c-Hex2BH as a white suspension which was ready to use (ca. 0.3 M).

Hydroboration/oxidation: To a cooled (0 oC) and stirred solution of 595 (10 mg, 0.03 mmol) in anhydrous THF (0.50 mL) was added freshly prepared c-Hex2BH solution (0.3

M, 0.50 mL, 0.15 mmol) dropwise. The mixture was slowly warmed to room temperature

. and stirred for 30 min. Solid NaBO3 H2O (60.0 mg, 0.60 mmol) and water (0.50 mL) was added and the resulting mixture was stirred at room temperature for 2.5 h. Water (10.0 mL) was added and the mixture was extracted with CH2Cl2 (5.00 mL x 3). The combined organic layers were dried (Na2SO4) and concentrated by rotary evaporation.

Chromatography of the residue on 20 x 20 TLC plate, gave compound 609 (8 mg, 75%) as a colourless oil with acceptable purity (due to the high polarity, the obtained compound

609 contained some impurities which was difficult to purify. Further purification was not performed, and this intermediate was put into next step directly).

S91

Compound 609: FTIR (neat, cm-1) 3362, 2934, 1697, 1478, 1459, 1438, 1354, 1321,

+ 1228, 1042, 727; HRMS (ESI-TOF, m/z) calcd for. C21H27N2O3 [M+H] 355.2022, found

1 355.2010; H NMR (700 MHz, CDCl3)  7.55 (1H, d, J = 8.2 Hz, ArH), 7.19—7.15 (1H, m, ArH), 7.15—7.11 (1H, m, ArH), 7.01 (1H, td, J = 7.4, 1.0 Hz, ArH), 5.82 (1H, dt, J =

10.0, 3.6 Hz, NCH2-CH=CH-C), 5.54 (1H, d, J = 10.1 Hz, NCH2-CH=CH-C), 3.84 (3H, s, NCO2Me), 3.72 (1H, ddd, J = 11.3, 9.2, 4.9 Hz, CH2-HCH-OH), 3.65 (1H, ddd, J =

11.3, 5.8, 4.4 Hz, CH2-HCH-OH), 3.16 (1H, d, J = 16.5 Hz, N-HCH-CH=CH-C), 3.07—

2.98 (2H, m, N-HCH-CH=CH-C and ArC-CH), 2.90—2.80 (1H, m, N-HCH-CH2), 2.61

(1H, s, N-CH-C-N), 2.55—2.48 (1H, m, N-HCH-CH2), 2.30—1.50 (8H, m, CH2-CH2OH,

NCH2CH2, C-CH2CH2-C).

Hydrolysis: To a solution of 609 (8 mg, 22.6 mol) in MeOH was added 3M KOH aq. solution and the resulting mixture was stirred at 100 oC for 16 h. After cooling to room temperature, the mixture was added water (5.00 mL) and extracted with CH2Cl2 (10.0 mL x 3). Combined organic layers were dried (Na2SO4) and concentrated to give crude 610

(5.3 mg, 79% crude yield) which was put into next step directly.

Oxidative cyclisation: To a solution to 610 (3 mg, 10 mol) in CH2Cl2 (0.50 mL) was added NMO (0.2M in CH2Cl2, 0.1 mL) and TPAP (0.01M in CH2Cl2, 0.1 mL) and the resulting solution was stirred at room temperature for 1 h. The reaction mixture was passed through a short pad of silica gel and the filtrated was concentrated and preparative

TLC on the residue, using 40% EtOAc in hxane, gave 543 (1.87 mg, 60% yield, 36% overall yield from 595) as a colourless oil.

25 -1 Rf (40% EtOAc in hexane) 0.2; []D +19.3 (c 0.09, CHCl3); FTIR (neat, cm ) 2922,

2851, 1656, 1597, 1476, 1459, 1387, 1370, 1260, 754; HRMS (ESI-TOF, m/z) calcd. for

+ 1 C19H21N2O [M+H] 293.1654, found 293.1649; H NMR (700 MHz, CDCl3)  8.04 (1H, dt, J = 8.1, 0.5 Hz, ArH), 7.24—7.21 (1H, m, ArH), 7.18 (1H, dt, 7.5, 1.3 Hz, ArH), 7.07 S92

(1H, td, J = 7.4, 1.1 Hz, ArH), 5.74 (1H, ddd, J = 10.0, 2.9, 1.9 Hz, NCH2-CH=CH-C),

5.58 (1H, ddd, J = 10.0, 4.5, 1.9 Hz, NCH2-CH=CH-C), 3.40 (1H, ddd, J = 16.8, 4.4, 1.9

Hz, N-HCH-CH=CH), 3.31 (1H, t, J = 6.6 Hz, ArC-CH), 3.06 (1H, ddd, J = 11.5, 7.2, 5.2

Hz, N-HCH-CH2), 2.80 (1H, ddd, J = 16.7, 2.9, 2.0 Hz, N-HCH-CH=CH), 2.70 (1H, d, J

= 18.0 Hz, NC(O)-HCH-C), 2.54 (1H, dd, J = 18.0, 2.8 Hz, NC(O)-HCH-C), 2.37—2.34

(2H, m, C-CH2-CH2-C), 2.29 (1H, ddd, J = 12.0, 7.1, 4.8 Hz, N-HCH-CH2), 2.27 (1H, s,

N-CH-C-N), 2.16—2.06 (3H, m, N-CH2-CH2 and C-CH2-HCH-C), 1.90 (1H, ddd, J =

13 13.0, 8.5, 5.6 Hz, C-CH2-HCH-C); C NMR (176 MHz, CDCl3)  169.8 (N-C=O), 142.7

(ArC), 132.6 (ArC), 130.2 (N-CH2-CH=CH-C), 128.3 (ArCH), 124.0 (ArCH), 123.8

(NCH2-CH=CH-C), 123.6 (ArCH), 115.9 (ArCH), 71.9 (N-CH-C-N), 68.5 (N-CH-C-N),

53.5 (N-CH2-CH=CH-C), 50.3 (N-CH2CH2), 47.2 (NC(O)-CH2-C), 44.7 (N-CH2-

CH=CH-C), 42.1 (ArC-CH), 38.5 (C-CH2-CH2-C), 37.7 (C-CH2-CH2-C), 25.5 (NCH2-

CH2).

S93

Appendix 2: Chiral HPLC analysis and X-ray diffraction data

a. Chiral HPLC analysis on compound 394

Figure 1 Chiral HPCL of racemic 394

S94

Figure 2 Chiral HPCL of enantioenriched 394

S95

b. X-ray diffraction data of single crystal on compound 499

a. Figure 1 Asymmetric unit of the crystal structure of compound 499. The thermal ellipsoids are shown at the 50% probability level. Hydrogen atoms and atom labels are omitted for clarity. Colour scheme: carbon – grey, nitrogen – blue, oxygen – red.

All diffraction data for compound 499 were collected on a four-circle Agilent SuperNova (Dual Source) single crystal X-ray diffractometer using a micro-focus CuKα X-ray beam (λ = 1.54184 Å) and an Atlas CCD detector. The crystal temperature was controlled with an Oxford Instruments cryojet. Unit cell determination, data reduction and analytical numeric absorption correction were carried out using the CrysAlisPro programme.1 The crystal structures were solved with the ShelXT programme2 and refined by least squares on the basis of F2 with the ShelXL programme.3 All non-hydrogen atoms were refined anisotropically by the full-matrix least-squares method. Hydrogen atoms associated with carbon and

oxygen atoms were refined isotropically [Uiso(H) = 1.2 Ueq (C)] in geometrically constrained positions. The crystallographic and refinement parameters for compound 499 are shown in Table 1. b. Table 1. Crystallographic and refinement parameters of 499.

Compound 499

Empirical formula C24H26N2O4

S96

Formula weight 406.47 Temperature/K 151(2) Crystal system orthorhombic

Space group P212121 a / Å 8.0488(1) b / Å 9.8870(1) c / Å 25.9184(3) α / ° 90 β / ° 90 γ / ° 90 V / Å3 2062.55(4) Z 4 −3 ρcalc / g cm 1.309 μ / mm−1 0.725 F(000) 864.0 Crystal size / mm3 0.272 × 0.134 × 0.104 Radiation CuKα (λ = 1.54184) 2Θ range for data collection / ° 6.82 to 133.184

Index ranges − 9 ≤ h ≤ 9 −11 ≤ k ≤ 11 −30 ≤ l ≤ 30 Number of collected reflections 68536

Number of unique reflections 3650 [I > 2σ(I)]

Data/Restraints/¨arameters 3650/0/284

Goodness-of-fit on F2 1.057

Rint 0.0446 R (F), F > 2σ(F) 0.0280 wR (F2), F > 2σ(F) 0.0690 R (F), all data 0.0289 wR (F2), all data 0.0698 −3 Δr (max., min.) e Å 0.15/−0.18 Flack parameter −0.08(4) CCDC number 1904992 c. References: 1) CrysAllisPro. Agilent Technologies, Inc. (2014). 2) Sheldrick, G. M. A short history of SHELX. Acta Crystallogr., 2008, A64, 112–122. DOI: 10.1107/S0108767307043930. 3) Sheldrick, G. M. Crystal structure refinement with SHELXL. Acta Crystallogr., 2015, C71, 3–8. DOI:1107/S2053229614024218

S97

[1] Rundell. C.; PhD thesis, 2016, University College London. [2] Ortega, V.; Del Castillo, E.; Csákÿ, A. G. Org. Lett. 2017, 19 (22), 6236–6239. [3] Leung, P. S. W.; Teng, Y.; Toy, P. H. Org. Lett. 2010, 12 (21), 4996–4999. [4] Lin, C.-H.; Hong B.-C.; Lee, C.-H. RSC Adv. 2016, 6, 8243-8247; [5] Chang, C.-F.; Huang, C.-Y.; Huang, Y.-C.; Lin, K.-Y.; Lee, Y.-J.; Wang, C.-J. Syn. Commun. 2010, 40, 3452-3466. [6] Evans, D. A.; Mito, S.; Seidel, D. J. Am. Chem. Soc. 2007, 129, 11583–11592. [7] X. Wang, D. Xia, W. Qin, R. Zhou, X. Zhou, Q. Zhou, W. Liu, X. Dai, H. Wang, S. Wang, L. Tan, D. Zhang, H, Song, X. Liu, Y. Qin, Chem 2017, 2, 803–816. [8] H. Tanino, K. Fukuishi, M. Ushiyama, K. Okada, Tetrahedron 2004, 60, 3273– 3282.

S98