UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Investigations towards the synthesis of vinblastine and vinblastine analogs Pedersen, Ove 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1992_spring_pedersen_ove.pdf [ 2.73MB ]
Metadata
JSON: 831-1.0061752.json
JSON-LD: 831-1.0061752-ld.json
RDF/XML (Pretty): 831-1.0061752-rdf.xml
RDF/JSON: 831-1.0061752-rdf.json
Turtle: 831-1.0061752-turtle.txt
N-Triples: 831-1.0061752-rdf-ntriples.txt
Original Record: 831-1.0061752-source.json
Full Text
831-1.0061752-fulltext.txt
Citation
831-1.0061752.ris

Full Text

INVESTIGATIONS TOWARDS THE SYNTHESIS OF VINBLASTINE ANDVINBLASTINE ANALOGSbyOVE PEDERSENM.Sc., University of Copenhagen, 1984.A THESIS SUBMITTED IN THE PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Department of Chemistry)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAJanuary 1992©Ove Pedersen, 1992In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of ChemistryThe University of British ColumbiaVancouver, CanadaDate March 5. 1992DE-6 (2/88)iiABSTRACTThis dissertation describes research towards a feasiblesynthesis of vinbiastine 1, by coupling of a suitablecatharanthine analog with vindoline 4. In this connection areinvestigation of the modified Polonovski reaction wasundertaken in order to gain a better understanding of thefactors affecting the yield of the reaction.The Diels-Alder reaction of the methyl a-chloroacrylate31 with the dihydropyridine 30 afforded the isomeric DielsAlder adducts 32 and 33. The two adducts were separated andtransformed into the indole-amides 40 and 51 as well as thecorresponding thioamides. Photochemical cyclization of theamides and the thioamides gave lactam 55 and thiolactam 61.However, with a best overall yield of 7% of lactam 55 theinvestigated strategy did not lead to a practical synthesisof vinbiastine.The low yields obtained in the coupling ofcatharanthine derivatives with vindoline 4 necessitated areinvestigation of the modified Polonovski reaction. Inusing fractional factorial design as experimental strategyand exocatharanthine 89 as a suitable catharanthine analog,a new intermediate was discovered and the yield of dimericalkaloid was improved from 35% to 70%.The better understanding obtained of the factorsinfluencing the yield of the modified Polonovski reactioniiishould now make it possible to couple catharanthine analogswith vindoline 4 in good yields.CH3h14.C H3H0CH3HCH3CH3 CH3C30 31iv0PhONc’CO2H3PhOIlCH0132 33Cl40H3Cc02510HH3H3C H3H3H5589VTABLE OF CONTENTSTitle Page iAbstract iiTable of Content VList of Figures ViiiList of Schemes ixList of Tables XList of Abriviations XiiiAcknowledgements XV11. Introduction 11.1. Background 11.2. Biosynthetic Considerations 31.3. Synthetic Considerations 51.4. Synthetic Strategy Chosen for a FeasibleSynthesis of Vinbiastine 1 9and Discussion 14Synthesis of the Isoquinuclidine Skeleton 14Transformation of the Endo Isomer 32 18Transformation of the Exo Isomer 33 27Photochemical Cyclization of the Amides40 and 51, and of the Thioamides 41 and 52 372.5. Evaluation of the Chosen Strategy 572.6. The Modified Polonovski Reaction 582.7. Synthesis and Characterization ofExocatharanthine 89 652.8. Investigation of the Formation of CatharanthineN-oxide 10 and Exocatharanthine N-oxide 90 . . .842.9. Reinvestigation of the Modified PolonovskiReaction 912.10. Identification and Characterization of theSecond Intermediate 99 in the ModifiedPolonovski Reaction 1112.11. Preliminary Investigation of the Reactivityof the 19’, 20’-Double Bond in 19’, 20’-Anhydrovinblastine 91 1263. Experimental 1353.1. General Experimental Conditions 1353.2. Catharanthine N-oxide 10 1373.3. N-Benzyloxycarbonyl-l, 2-dihydropyridine 30. . . .1383.4. a-Chloro methyl acrylate 31 1402. Results2.1.2.2.2.3.2.4.vi3.5. N-Benzyloxycarbonyl-endo-7-methoxycarbonyl-7-chloro-2-azabicyclo [2,2,2] octan-5-ene 32and N-Benzyloxycarbonyl-exo-7 —methoxycarbonyl7-chloro-2-azabicyclo [2,2,2] octan-5-ene 33. .1423. 6. N-Benzyloxycarbonyl-encZo-7-methoxycarbonyl-7--chloro-2-azabicyclo [2,2,2] octan-endo-6-ol 34and N-Benzyloxycarbonyl-endo- 7-methoxy-carbonyl-7-chloro-2-azabicyclo [2,2,2] octanendo-5-ol 35 1463.7. N-Benzyloxycarbonyl-endo-7--methoxycarbonyl-7--chloro-2-azabicyclo [2,2,2] octan-6-one 36and N-Benzyloxycarbonyl-endo-7-methoxy-carbonyl-7-chloro-2-azabicyclo [2,2,2] octan5-one 37 1503.8. N-Benzyloxycarbonyl-endo-7-methoxycarbonyl-7-chloro-2-azabicyclo [2,2,2] octan-6-one2’, 2’ -dimethyl-1’ , 3’ -propanediyl acetal 38.. . .1543.9. N-(3’ ‘-Indolylmethylenecarbonyl)-endo-7-methoxy-carbonyl -7 -chloro-2 -azabicyclo[2,2,2] octan-6-one 2’,2’-dimethyl-1’,3’-propanediyl acetal 40 1563.10. N-(3’ ‘-Indolylmethylenethiocarbonyl )-endo-7-methoxy-carbonyl-7-chloro-2-azabicyclo[2,2,2] octan-6-one 2’, 2’ -dimethyl-1’ .3,-propanediyl acetal 41 1583.11. N-Benzyloxycarbonyl-exo-7-methoxycarbonyl-7-chloro-2-azabicyclo [2, 2,2] octan-endo-6-ol 44and N-Benzyloxycarbonyl-exo-7-methoxycarbonyl-7-chloro--2-azabicyclo [2,2,2] octanendo-5-ol 45 and N-Benzyloxycarbonyl7-methoxycarbonyl-2-azatricyclo [2,2,2,06,7]octane 46 1603.12. N-Benzyloxycarbonyl-exo--7-methoxycarbonyl-7-chloro-2--azabicyclo [2,2,2] octan-6-one 47and N-Benzyloxycarbonyl-exo-7-methoxy-carbonyl—7-chloro-2-azabicyclo [2,2,2] octan5-one 48 and N-Benzy1oxycarbony--methoxy-carbonyl-2-azatricyclo [2,2,2,0‘ ]octane 46 1653.13. N-Benzyloxycarbonyl-exo-7-methoxycarbonyl-7-chloro—2--azabicyclo [2,2,2] octan-6-one2’, 2’ -diinethyl-1’ , 3’ -propanediyl acetal 49.. . .1683.14. N-(3’ ‘-Indolylmethylenecarbonyl)-exo-7-methoxycarbonyl -7 -chioro- 2- azabicyclo[2,2,2] octan-6-one 2’ ,2’-dimethyl-l’ 13’-propanediyl acetal 51 1713.15. N-(3’ ‘-Indolylmethylenethiocarbonyl)-exo-7-methoxycarbonyl-7-chloro-2-azabicyclo[2,2,2] octan—6-one 2’, 2’ -dimethyl-1’,3’-propanediyl acetal 52 1743.16. 20-Desethyl-15,20-dihydro-5-oxo-catharanthin-20-one 2’ ,2’ -dimethyl-1’ ,3’- propanediylacetal 55 and the Isomeric By-product 56 177vii3.17. 20-Desethyl-15, 20-dihydro-5-thioxo-catharanthin-20-one 2’ ,2’-dimethyl-l’ 13’- propanediylacetal 61 and the Isomeric By-product 62 1803.18. Exocatharanthine 89 1833.19. Exocatharanthine N-oxide 90 1853.20. 19’,20’-Anhydrovinblastine 91 and Epi19’ , 20’ -anhydrovinbiastine 94 1873.21. 7-Hydroxy cathraranthine N-oxide 92 1933.22. 7-Hydroxy exocatharanthine N-oxide 93 1953.23. 4’-Benzyl-19’ ,20’-anhydrovinblastine 108 1973.24. 19’-Hydroxy vinbiastine 109 and 19’-Hydroxyleurosidine 110 1983.25. Bis-indolic mesylate 112 from the Mesylationof 19’-Hydroxy vinbiastine 109 2033.26. General Experimental Conditions for theOptimization of the Photochemical Cyclizationof the Amides 40 and 41 and the Thioamides51 and 52 2053.27. General Experimental Conditions Used in theOptimization of the Yield of Exocatharanthine89 2063.28. General Experimental Conditions Used in theInvestigation of the Formation of CatharanthineN-oxide 10 and Exocatharanthine N-oxide 90 .2073.29. General Experimental Procedure Used in theSystematic Investigation of the ModifiedPolonovski Reaction 2083.30. Preparation of the Second Intermediate 99 2124. References 2145. Appendix 2205.1. Factorial Design 2205.2. Fractional Factorial Design 224viiiList of FiguresFigure 2-1. ‘H NMR spectra of the Diels-Alder adduct33 in DMSO-d6 17Figure 2-2. X-ray diffraction analysis of ketone 36 20Figure 2-3. X-ray diffraction analysis of thioamide 41.. .26Figure 2-4. X-ray diffraction analysis of alcohol 45 29Figure 2-5. X-ray diffraction analysis the tricyclic byproduct 46 30Figure 2-6. The conformers 52A and 52B of thioamide 52.. .37Figure 2-7. A graphic representation of the resultsof the fractional factorial designs intable 2-6 and 2-7 45Figure 2-8. A graphic representation of the resultsof the fractional factorial design intable 2-9 52Figure 2-9. The overall yields of cyclized products 55 and61 from the Diels-Alder adducts 32 and 33. . .58Figure 2-10. The structure of the isolated carbomethoxydihydrocleavamines 84Figure 2-11. The CD spectrum of 19’,20’-anhydro-vinblastine 91 and epi 19’,20’-anhydro-vinbiastine 94 93Figure 2-12. Dehydration of vinblastine 1 94Figure 2-13. Monitoring of the Polonovski reactionby HPLC 101Figure 2-14. Possible structures of the secondintermediate 99 112Figure 2-15. -H NMR spectrum of the second intermediate99 114Figure 2-16. The structure of the mesylate decompositionproduct 112 determined by X-ray diffractionanalysis 133Figure 3-1. Apparatus used in the photolysis 205ixList of SchemesScheme 1-1. Possible pathways for the biosynthesis ofvinbiastine 1 4Scheme 1-2. Synthesis of 15’,20’-anhydrovinblastine 5 6Scheme 1-3. Hypothetical formation of vinblastine 1 byintroduction of an ethyl group into asuitable ketone 9Scheme 1-4. Strategies used in the synthesis ofcatharanthine 3 and desethylcatharanthine 17 11Scheme 1-5. Synthetic strategy chosen for a feasiblesynthesis of vinbiastine 1 13Scheme 2-1. Synthesis of the diene 30 and thedienophile 31 14Scheme 2-2. Synthesis of the Diels-Alder adducts 32 and33 15Scheme 2-3. Transformation of the Diels-Alder adduct 32.. 19Scheme 2-4. Synthesis of the amide 40 and thethioamide 41 22Scheme 2-5. Transformation of the Diels-Alder adduct 33 .28Scheme 2-6. Synthesis of the amide 51 and thethioamide 52 32Scheme 2-7. Photochemical cyclization of the amides 40and 51, and the thioamides 41 and 52 38Scheme 2-8. Possible path ways in the photochemicalcyclization reaction of aromatic a-chloroacetates 56Scheme 2-9. The mechanism of the Polonovski reaction 59Scheme 2-10. The pathways in the modified Polonovskireaction 60Scheme 2-11. The coupling of catharanthine 3 withvindoline 4 61Scheme 2-12. Possible mechanisms for the coupling ofcatharanthine 3 with vindoline 4 63Scheme 2-13. Coupling of exocatharanthine 89 withvindoline 4 92Scheme 2-14. The reaction pathways in the modifiedPolonovski reaction 100Scheme 2-15. Transformation of 15’20’-double bond in15’,20’-anhydrovinblastine 5 126Scheme 2-16. Mesylation of the major diol 109 followed bydecomposition to 112 134Scheme 5-1. The factors and their levels used in thefactorial design in scheme 5-2 221Scheme 5-2. The factorial design 222Scheme 5-3. A three factor two level factorial design. .226Scheme 5-4. Fractional factorial design 227xList of TablesTable 2-1. Chemical shifts observed in the ‘3C APTNMR spectrum of thioamide 41 24Table 2-2. Chemical shift correlations obtained in theHETCOR spectrum of thioamide 41 in theregion 0-100 ppm 25Table 2-3. Chemical shifts observed in the C APTNMR spectrum of thioamides 52A and 52B 34Table 2-4. Chemical shift correlations obtained in theHETCOR spectrum of thioamide 52A in the region0-100 ppm 36Table 2-5. Investigation of the effect of the nature ofthe solvent, the wave length and thetemperature on the photochemical cyclizationof amide 40 40Table 2-6. Investigation of the photochemical cyclization of amide 40 using fractional factorialdesign with four factors on three levels 41Table 2-7. Investigation of the photochemical cyclization of amide 51 using fractional factorialdesign with four factors on three levels 43Table 2-8. Investigation of the photochemical cycliza—tion of thioamide 41 using fractional factorialdesign with seven factors on two levels 48Table 2-9. Investigation of the photochemical cyclization of thioamide 41 using fractional factorialdesign with four factors on three levels 50Table 2-10. The photochemical cyclization of thioamide52 compared to thioamide 41 53Table 2-li. Coupling of catharanthine derivativeswith vindoline 4 64Table 2-12. Chemical shifts observed in the H NMRspectrum of exocatharanthine 89.4 68Table 2-13. Chemical shifts observed in the C APT NMRspectrum of exocatharanthine 89 69Table 2-14. Chemical shift correlations obtained in theHETCOR spectrum of exocatharanthine 89 70Table 2-15. Chemical shift correlations obtained in theCOSY spectrum of exocatharanthine 89 71Table 2-16. Observed enhancements in SINEPT experimentson exocatharanthine 89 72Table 2-17. Observed enhancements in NOE experimentson exocatharanthin 89 72Table 2-18. Assignment of the -‘-3C chemical shifts ofcatharanthine 3 77Table 2-19. Assignment of the H and chemical shiftsof exocatharanthine 89 78Table 2-20. Conditions investigated for the metalcatalyzed isomerization of catharanthine 3to exocatharanthine 89 80xiTable 2-21. Optimization of the isomerization ofcatharanthine 3 to exocatharanthine 89 82Table 2-22. The factors and their levels used inthe fractional factorial design intable 2-23 88Table 2-23. Investigation of the formation of7-hydroxy catharanthine N-oxide 92 89Table 2-24. The influence of acid on the formationof 7-hydroxy catharanthine N-oxide 92 90Table 2-25. Investigation of the effect of thereaction temperature, the purity of the mchloroperbenzoic acid and the addition rate ofTFAA on the yield of 19’,20’-anhydro-vinblastine 91 96Table 2-26. The Polonovski reaction carried outat low temperature 97Table 2-27. Investigation of the effect of the amountof vindoline 4, the nature of solvent andthe amount of TFAA on the yield of l9’,20’-anhydrovinblastine 91 98Table 2-28. Addition of TFAA to exocatharanthine N-oxide90 prior to the addition of vindoline 4 103Table 2-29. Addition of vindoline 4 prior to the additionof TFAA. The importance of concentrating thereaction mixture after the reaction has gone to“completion” at -65°C 105Table 2-30. The effect of concentrating (versus notconcentrating) the reaction mixture after thereaction has gone to “completion” at -65°C onthe yield and product distribution 107Table 2-31. The type of carbon signals expected for eachof the five possible structures in figure2-14, together with th actual types ofcarbons found in the APT spectrum forthe second intermediate 99 113Table 2-32. Chemical shifts observed in the ‘H NMRspectrum of the second intermediate 99 115Table 2-33. Chemical shift correlations obtained in theCOSY spectrum of the second interediate 99. .116Table 2-34. Chemical shifts observed in the C APT NMRspectrum of the second intermediate 99 117Table 2-35. Chemical shift correlations obtained in theHETCOR spectrum of the second intermediate99 118Table 2-36. Observed enhancements in the SINEPT spectrumon the second intermediate 99. Iradiation at7.39.ppm and observation in the C chemicalshift region of 0 - 90 ppm 120Table 2-37. Observed enhancements in the SINEPT experimentof the second intermediate 99. Irradiation at457 ppm and 4.09 ppm and observation in thechemical shift region of 80 - 175 ppm.. . .122xiiTable 2-38. Observed enhancements in the SINEPT experimentof the second intermediate 99. Irradiationat 7.73 ppm (H-9) and 746 ppm (H-12)and observation in the C chemical shiftregion of 83 - 175 ppm. . 123Table 2-3g. Assignment of the ‘H and ‘C chemical shiftsto the second interediate 99 125Table 2-40. Assignment of the C chemical shifts invinblstine 1 and leurosidine 113 130Table 2-41. The chemical shifts of the aliphaticmethylenes and quaternary carbons groupsin vinblastine 1, leurosidine 113and the two diols 109 and 110 132Table 5-1. Confounding of main effect and two-factorinteractions in a fractional factorialdesign of three factors each on two levels. .227xiiiList of Abbreviations:Ac AcetateAddn AdditionAPT Attached proton testAr ArylAr—H Aromatic protont-BuOH tert—ButanolCaic CalculatedCD Circular DichroismCOSY Correlated spectroscopyd Doublet2-D Two-dimensionalDCC 1, 3-Dicyclohexylcarbodiimidedd Doublet of doubletsDMSO Dimethyl sulfoxidedt Doublet of tripletsEt EthylEther Diethyl etherEt3N Triethyl amineeq Equivalentevac Evacuationexo Exocatharan-thineexp ExperimentG. Don Genus DonHETCOR Hetereonuclear correlationHPLC High pressure liquid chromatographyxivhrs Hoursi Isoipr2NH Iso-propylamineIR Infaredm Multipletm/z Mass to charge ratioMe-morph N-MethylmorpholineMeOH MethanolMHz Mega hertzzn-Cl PBA zneta-Chloroperbenzoic acidmp Melting pointMS Mass spectroscopynm NanometerNMR Nuclear magnetic resonanceNOE Nuclear Overhauser effectNu NucleophilePCC Pyridinium chlorochromatePh PhenylPLC Preparative thin layer chromatographyppm Parts per millionp-TsOH para-Toluene sulfonic acidRT Room temperature5 Singletsec SecondsSINEPT Selective insensitive nuclei enhancementby polarization transfersolv Solventxvt Terttemp TemperatureTFAA Trifluoroacetic acidTHF TetrahydrofuraneTLC Thin layer chromatographyUV Ultravioletvac VacuumVCR VincristineVLB VinblastinexviACKNOWLEDGEMENTSI wish to express my appreciation to Professor James P.Kutney for the opportunity to pursue this project and forhis advice both during the progress of this research and inthe preparation of this thesis. I wish to express myappreciation to the members of Dr. Kutney’s research groupfor their advice, enlightening discussions and friendshipand in this connection I would especially like to thank Dr.Katalin Honty for her invaluable advice. I would also liketo express my appreciation to my colleagues at NielsClauson-Kaas Chemical Research Laboratory for their supportthroughout the preparation of this thesis. Finally, I wishto thank Dr. Mike McHugh for proof reading and criticism ofthis thesis and to thank Dr. J. Trotter, Department ofChemistry, University of British Columbia, for performingthe X-ray diffraction analyses.11. INTRODUCTION1.1. BACKGROUNDThe bisindole Catharanthus alkaloids vinbiastine(vincaleukoblastine or VLB) 1 and vincristine (leurocristineor VCR) 2 are important anticancer agents which are usedroutinely in the treatment of a number of human cancers.1 R = CH32 R = CHOC H3Vinbiastine 1 was first isolated in 19581,2 andvincristine 2 three years later in l961 by extraction fromthe leaves of the periwinkle Catharanthus roseus G. Don.Catharanthus roseus although native to Madagascar has nowspread throughout the tropics4. Attention had been drawn tothis plant on the basis of its reported use in folklore asan oral hypoglycemic agent5. The reported hypoglycemicCF-I3 Ac02CH32properties could not, however, be verified clinically butthe investigations led to the discovery of significantoncologic activity of some of the plant extracts’,culminating in the isolation of vinbiastine 1 andvincristine 21,2. Some 70 different alkaloids have beenisolated from Catharanthus roseus5 with vindoline 4 andcatharanthine 3 being the major constituents. Vinbiastine 1and vincristine 2, being minor constituents, are found inonly 100- 700 mg/kg dried leaves for the former and 6 -30 mg/kg dried leaves for the latter6.The major clinical use of vinbiastine 1 is in thetreatment of Hodgkin’s disease and vincristine 2 is mainlyused against childhood leukemia, Wilm’s tumor and non-Hodgkin’s lymphomas79. The mechanism of action of 1 and 2is not known but the activity is probably connected to thefact that they both are mitotic inhibitors causing metaphasearrest in dividing cells1-0. A major problem with bothalkaloids is their relatively high toxicity7°. Vinbiastine1 shows mainly hematologic toxicity and vincristine 2neurotoxicity.The structure of vincristine 2 was established in 1965by an X-ray diffraction study of a single crystal ofvincristine methiodide-1. The structure of vinbiastine 1followed from its known relationship to vinblastine 212.31.2. BIOSYNTHETIC CONSIDERATIONSSince catharanthine 3 and vindoline 4 are the majoralkaloidal components of Catharanthus roseus Atta-Ur-Rahmanproposed that they might be the biological precursors ofvinblastine 1 and vincristine 213. Feeding experiments byScott et al. with carbon-14 labelled vindoline 4 and tritiumlabelled catharanthine 3 using whole Catharanthus roseusplants gave incorporation of labelled material into 15’,20’-anhydrovinbiastine 5 as well as vinbiastine 14a• Similarfeeding experiments using apical cuttings of Catharanthusroseus plants gave incorporation of labelled catharanthine 3and vindoline 4 into vinbiastine 14b• Experiments usingcell free extracts have also shown the incorporation oflabelled catharanthine 3 and vindoline 4 into 15’,20’-anhydrovinbiastine 5151) as well as the transformation oflabelled 15’,20’-anhydrovinblastine 5 to vinbiastineKutney et al have shown that coupling of catharanthine 3with vindoline 4 using cell free extract initially resultsin the formation of the iminium intermediate 717a (scheme 1-1). The same intermediate is also formed when 15’,20’-anhydrovinbiastine 5 is incubated in cell free extractsl7b.Furthermore it has been shown that the iminium intermediate7 can be reduced to the enamine 12 which subsequently can beoxidized to vinblastine 17c,27,28 Whether the actualbiosynthesis of vinbiastine 1 in Catharanthus roseus plantsproceeds via 15’,20’-anhydrovinblastine 5 or via the enamine12 is still to be established.4Scheme 1-1. Possiblevinbiastine 1.H3+H3HH347CH3H2C H3INDOLINEH1INDOLINE5I ifH INDOLINE12pathways for the biosynthesis of51.3. SYNTHETIC CONSIDERATIONSThe first approach to these bisindole alkaloids, knownas the “chloroindolenine approach”, consisted of treatingl613-carbomethoxy-2O3-dihydrocleavamine 6a or 1613-carbo-methoxycleavamine 6b with tert-butyl hypochlorite to givethe corresponding 7-chioroindolenine 8. Treatment of 8Sb: 15,2O’9a9b: 15’,2O’H6aClt—BuOC 1CH3HH38a6b/HC1’MQH,/CH3H36with vindoline under acidic conditions gave the dimer13,18 X-ray diffraction analysis revealed, however, that 9possessed the unnatural stereochemistry (R) at C-16’19.1) vindoline2) (CF3CO)20CH1’—EO °CNaB H4CO H05Scheme 1-2. Synthesis of 15’,20’-anhydrovinblas-tine 5.•C H33 10HPcCH37H3CH3CO2H37The simultaneous discovery in 1975 by Kutney et al2°and by Potier et al.21, that treatment of catharanthine N-oxide 10 with trifluoroacetic anhydride in the presence ofvindoline 4 gave 15’,20’-anhydrovinblastine 5 with thenatural stereochemistry at C-16’, opened for the first timea viable route to the synthesis of vinbiastine 1 (scheme 1-2). However, despite considerable effort, transformation ofthe 15’,20’-double bond in 5 to either vinblastine 1 or it’s20’ epimer leurosidine were unsuccessful22’3.The first reported synthesis of vinblastine 1 consistedof coupling the N-oxide of the catharanthine derivative 11with vindoline 4 and subsequent removal of the acetyl groupat the 20’ position24. The structure of 11, formed in amodified Prévost reaction from catharanthine 325, has,however, been questioned after it was found, in an extensivestudy of the chemistry of the C-15 and C-20 positions incatharanthine 3, that nucleophilic approach from the a-sideof the quinuclidine system was impossible26, thus drawingthe correctness of structure 11 into question. Vinbiastine 1118has also been synthesized from the enamine 12, treating iteither with thallium(III) acetate27 or with iron(III)chloride and oxygen28 followed by reduction with sodiumborohydride. The later reaction has been reported to give anover-all yield of 40% of vinbiastine 1 based on vindoline428Ti (OPc >31) orFeC13’022) NaBH4HC H3CH312CH3HCO2H3CH PcCH3 02CH3191.4. SYNTHETIC STRATEGY CHOSEN FOR A FEASIBLE SYNTHESIS OFVINBLASTINE 1In the investigation of the reactivity of the 15’,20’-double bond in 15’,20’-anhydrovinblastine 5 it was observedthat electrophiles always attacked the double bond from the3-face of the piperidine ring system22’3.Assuming that theobserved approach of electrophiles to the 15’,20’-doublebond in 5 is due to steric hindrance of the piperidine ringsystem then addition of an ethyl group to the ketone 13 in130Scheme 1-3. Hypothetical formation of vinbiastine 1 byintroduction of an ethyl group into a suitable ketone.H3EH5EHINOOLINEHtNDOLINE110scheme 1-3 should lead to vinblastine 1. Experimentalsupport for this view is found in the synthesis ofvelbanamine 15 by Büchi etH3We therefore adopted the strategy of introducing theethyl group last as outlined in scheme 1-3. This in turnrequires the synthesis of the catharanthine analog 16 shownbelow.Several syntheses of catharanthine 3 and desethylcatharanthine 17 have been reported in the literature. Inscheme 1-4 are summarized the different strategies employedin the formation of the isoquinuclidine part ofcatharanthine 3 and the final ring closure leading to theformation of the catharanthine skeleton. Two differentstrategies have been applied in the formation of theLiEtOHH H14 1516113 R=Et17 R=H24a R1H R2025a R1Et R2S24b R1=H R2=PhCHOCO25b R1Et =CH3O OScheme 1-4. Strategies used in thecatharanthine 3 and desethylcatharanthine 17.27 R = PhCH228 R PhCHO Osynthesis ofOSi (CH3)2ButCO2H321H191CO2H302Ph19/fl\H3®CCO2H320;02CH3220H 0CH3R1 H2602Ph0C23P12isoquinuclidine part of the catharanthine skeleton. Instrategies i3O, 231, and 332 the isoquinuclidine system isformed by an intermolecular or intramolecular reaction of adienophile attached to indole with a suitable 1,2-dihydropyridine. The final cyclization to the catharanthineskeleton is achieved either through an intramolecularphotochemical cyclization as in strategy 1, anintramolecular nucleophilic substitution as in strategy 2 oran intramolecular addition reaction as in strategy 3. Instrategy 433,34 the isoquinuclidine unit was first built andthen attached to the indole system, followed by an intramolecular photochemical cyclization to the catharanthineskeleton. In strategy 529,35 the carbomethoxy group is not apart of the isoquinuclidine system but is introduced afterthe acid catalyzed cyclization to the ibogamine skeleton.Based on the background outlined above a retrosyntheticanalysis was carried out as outlined in scheme 1-5. Thesteric requirements observed for 15’,20’-anhydrovinblastine522 and velbanamine 1529 dictated our immediate sub-targetas the catharanthine analog 16, which was to be coupled tovindoline 4. The resulting bisindole 14 was subsequently tobe transformed into vinbiastine 1. Alternatively, the ethylgroup could be introduced into the catharanthine analog 16prior to the coupling with vindoline 4 as outlined in scheme1-5. The strategy chosen for the formation of 16 was similarto route 4 in scheme 1-4 where the isoquinuclidine systemwas formed first, transformed, coupled to indole and13finally cyclized to give the desired catharanthine analog16H3PhC 02C H332/33Scheme 1-5. Synthetic strategy chosen for a feasiblesynthesis of vinbiastine 1.0HCO2H3INDOLINE1OHCO2H329H NDOLINE1416142. RESULTS AND DISCUSSION2.1. SYNTHESIS OF THE ISOQUINUCLIDINE SKELETONThe necessary diene and dienophile for the formation ofthe desired isoquinuclidine skeleton were synthesized asoutlined in scheme 2-1. Addition of benzyl chloroformate inhNaBH4I + PhCH2OCOC1 ICH3OH/—?8°C -30(867.)+ HCHO + CH3OH2S04/i00°CH0CH3(357.)Scheme 2-1. Synthesis of the diene 30 and the dienophile 31.anhydrous ether to a suspension of pyridine and sodiumborohydride in anhydrous methanol at -78°C gave 30 in 86%yield36. The NMR spectrum of 30 revealed that it waspresent as a mixture of two conformers in a ratio of 4:6.The chemical shifts of the olefinic protons in 30 appearedat 6.71, 5.83, 5.53 and 5.10 ppm (for the major conformer)15unambiguously showing 30 to be a 1,2-dihydropyridine. The ‘HNMR of 30 also indicated the presence of the corresponding1,4—dihydropyridine isomer as a contaminant. Based on theintegral of the H—4 protons, an easily recognizablemultiplet at 2.79 ppm37, the content of the 1,4-dihydropyridine isomer was calculated to be about 7%.The dienophile 31 was prepared by treatment oftrichioroethylene with aqueous formaldehyde and methanol inconcentrated sulfuric acid at 1000C38,9. The structure ofthe dienophile 31 was established by a coupling constant of1.5 Hz for the olefinic protons.)CH3benzener e f’ 1 ux032 +(357.)+.c130 31‘02CH333(267.)Scheme 2-2. Synthesis of the Diels-Alder adducts 32 and 33.16The Diels-Alder reaction of 30 with 31 in refluxingbenzene gave only two products, 32 and 33, in a combinedyield of 70% (scheme 2-2). The yield of 32 and 33 afterseparation by repeated column chromatography was 35% and 26%respectively. The ratio of 1.3:1 does not reflect the realratio in which the two adducts were formed since compound 33showed some decomposition under the chromatographicconditions. The ‘H NMR spectra of the two Diels-Alderadducts are relatively complicated due to splitting of mostof the signals, as illustrated by the ‘H NMR spectrum ofcompound 33 in figure 2-1. Heating either compound to 120°C.in DMSO—d6 causes the split signals to collapse,establishing that this splitting is due to conformers andnot to the presence of isomers. For each Diels-Alder adductthe ratio of the two conformers is somewhat solventdependent. For example, in DMSO-d6 the ratio between the twoconformers of compound 32 is 62:38, but in CDC13 the ratiois 52:48 . For compound 33 the ratio is 71:29 in DMSO-d6,but 57:43 in CDC13. However, for both compounds, it is thesame conformer which predominates in both solvents. Themost noticeable difference in the 1H NMR spectra (in CDC13)of the two Diels-Alder adducts is the splitting of themethoxy signal, which in one case is 4 Hz and in the othercase is 64 Hz. The cause of this splitting can only arisefrom different conformations of the carbamate group, and thespectrum showing the larger splitting is therefore assignedto structure 32, in which the methoxy group is closest to17a) ‘H NMR spectrum of 33 in DMSO-d6 at 25°C.7 6 5 4. 3 2b) -H NMR spectrum of 33 in DMSO-d5 at 120°C.Figure 2-1. ‘H NMR spectra of the Diels-Alder adduct 33 inDMSO-d6.18the carbamate group. In the NMR spectrum of 33 in DMSO-d6at 120°C the doublet at 5.19 ppm (J = 4.0 Hz) is assigned toH-i. This assignment is verified by irradiation of theolefinic proton at 6.44 ppm which causes the doublet at5.19 ppm to collapse to a singlet. The -H NMR spectrum of 32in DMSO-d6 at 120°C revealed the H-i proton to be a doublet(J = 4.0 Hz) too. Since the H-i protons in bothcompoundsare doublets the structure of the Diels—Alder Alder adductsmust be as shown in scheme 2-2.2.2. TRANSFORMATION OF THE ENDO ISOMER 3240.Treatment of 32 with borane dimethyl sulfide complex at26- 28°C in THF and subsequent oxidation with 30% hydrogenperoxide gave 80 - 85% of 34 and 7 - 10% of 35 (scheme 2-3).Both alcohols gave the expected molecular ion at 353/355 m/zin the low resolution mass spectrum and the presence of thehydroxy group was established through IR absorption at 3445cm and at 3460 cm for 34 and 35 respectively. Theseparation of the alcohols 34 and 35 was difficult, and itwas found to be more convenient to oxidize the alcoholmixture to the corresponding ketones which were readilyseparated by fractional crystallization. The position of thehydroxy group in 34 and 35 was established by oxidation tothe corresponding ketones 36 and 37. The observedregioselectivity must be due to a directional effect of thecarbamate nitrogen through complexation to borane in thehydroboration step. Furthermore, complexation of the borane190PCH3O ci321) BH3S(CH2THF/26°C2) NaQH’30X HD2THF/MeOH/26°C(80—857.) (7—lOX)036(777.)+/ccPCH21/Ref1LLxCH3H H3HO OH, p—TsOHbenzene/reflUX0CH3C H3o’JJCH338(997.)034 351C 137Scheme 2-3. Transformation of the Diels-Alder adduct 32.20molecule to the carbamate nitrogen would cause the hydroxygroup in 34, and perhaps also in 35, to be endo.Oxidation of the alcohol 34 with pyridiniumchlorochromate in refluxing dichioromethane gave thecorresponding ketone 36 in 77% yield. Compared to 34 the IRspectrum of 36 showed lack of the absorption at 3445 cm’and an increased absorption at the position of the estercarbonyl at 1740 cm, due to the keto group. The ‘H NMRspectrum in CDC13 revealed that the ketone 36 exists as twoconformers in a ratio of 40:60. The H-i proton in 36 at4.85 ppm (major conformer) is a sharp singlet, establishingthe keto group to be in position 6. This was confirmed by Xray diffraction analysis (figure 2-2), which also confirmedthe position as well as the stereochemistry of the chlorineC14C13c102•03Figure 2-2. X-ray diffraction analysis of ketone 36.21atom. The mass spectrum of the isomeric ketone 37 showed theexpected molecular ion at 353/351 m/z. In solution theisomeric ketone 37 also exists as two conformers in a ratioof 39:61. The H-i proton in the 1-H NMR spectrum of 37 at5.01 ppm is a dd with J = 3.5 Hz and J’ = 2 Hz, establishingthe keto group to be in position 5.The ketal 38 was formed in 99% yield by heating theketone 36 with 2,2-dimethyl-i,3-propanol at ref lux inbenzene overnight using p-toluenesulfonic acid as catalyst(scheme 2-3). The mass spectrum of 38 did not show themolecular ion but only M+ - Cl at 402 m/z. The presence ofchlorine was, however, confirmed by elemental analysis. Themethyl groups of the ketal group appear in the 1H NMRspectrum at 1.22 and 0.81 ppm (major isomer).The benzyloxycarbonyl group was removed byhydrogenolysis in THF at room temperature using 10%palladium on activated carbon as catalyst. The crude amine39 was used without work-up in the subsequent couplingreaction with DCC and 3-indole acetic acid, giving the amide40 in 54% yield (scheme 2-4). In an attempt to improve theyield of 40 amine 39 was treated with 3-indole acetylchloride in dichloromethane in the presence oftriethylamine. However, the crude product obtained in thisway turned out to be very difficult to purify, resulting ina much lower yield than in the DCC promoted reaction. Themass spectrum of 40 gave a weak molecular peak at 460/462m/z followed by a peak at 424 m/z due to the loss of22hydrogen chloride. The presence of chlorine was established0Ph0 1) H2, Pd/CN THFI25°CCH3O ci 0CH32)COORCH 0 •CH338CH3CH21, 25°C CH3(547.)C H3O— PPQoCR3Toluene, 75°CCH341(537.)Scheme 2-4. Synthesis of the amide 40 and the thioamide 41.by elemental analysis. The IR spectrum showed acharacteristic N-H stretch at 3250 cm and the UV spectruma typical indole absorption pattern at 272, 279 and 289 nm.The most remarkable difference in the 1H NMR spectrum of 40compared to that of 38 (besides the replacement of thebenzyloxycarbonyl group with 3-indole acetyl group) is thatdifferent conformers are no longer observed and the positionof the H-i proton has shifted down field from 5.38 ppm in 38(major conformer) to 5.88 ppm in 40.H’4023Treatment of 40 with Lawesson’s reagent76 in toluene at75°C for 10 hours gave the corresponding thioamide 41 in 53%yield (scheme 2-4). Using phosphorous pentasulfide insteadof Lawesson’s reagent gave about the same yield butcomplicated the work-up. The mass spectrum of 41 gave theexpected molecular ion at 476/478 m/z. In ‘H NMR spectrumthe H-i proton was assigned to the singlet at 6.88 ppm. Inorder to verify this a 3-C APT (Attached Proton Test)spectrum (table 2-1) and a HETCOR (HeteronuclearCorrelation) spectrum (table 2-2) of 41 were obtained. Asexpected, the aliphatic region of the 13C APT spectrumcontained 5 carbons with an odd number of hydrogensattached, namely at 22.0, 23.1, 27.4, 53.3 and 53.4 ppm.The signals at 22.0 and 23.1 correspond to the two methylgroups of the ketal and the signal at 27.4 ppm is attributedto C-4. Judged by their relative intensities the smallersignal at 53.4 ppm was assigned to C-i and the larger signalat 53.3 ppm to the methoxy group. The HETCOR spectrumconfirmed this interpretation by correlating the signal at53.3 ppm with the three proton singlet at 3.73 ppm in theNMR spectrum and the signal at 53.4 ppm with the proton at6.88 ppm.The downfield shift observed for H-i in 41 comparedto the H-i proton in the corresponding amide 40, caused bythe introduction of sulfur, must either be due to a changein equilibrium conformation or to the larger size of sulfurwhich results in an increased deshielding of H-i.24Table 2-1. Chemical shifts observed in the ‘3C APT NMRspectrum of thioamide 41a5 (ppm) +/_b Interpretation S (ppm) 1_b Interpretation22.0 - CH3-C 109.5 + Indole-C23.1 - CH3-C 111.2 - Indole-CH27.4 - C-4 118.3 - Indole-CH29.1 + >C< 119.5 - Indole-CH37.1 + -CH2 122.1 - Indole-CH41.2 + CH2 122.3 - Indole-CH42.6 + CH2 126.8 + Indole-C49.9 + CH2 136.0 + Indole-C53.3 - CH3O- 170.3 + C=S53.4 - C-i 202.9 + C=062.2 + C-771.5 + -CH2O-73.2 + -CH2O-98.1 + c-6a Solvent: CDC13b The +/- sign indicate whether the signal s positive (>C<,-CH2) or negative ( >CH-, CH3-) in the C APT spectrum.25Table 2-2. Chemical shift correlations obtained in theHETCOR spectrum of thioamide 41 in the region 0-100shift (ppm) ‘H correlation22.0 0.8023.1 1.2027.4 2.1229.1 no correlation37.1 3.09 and 1.8841.2 4.24 and 4.1042.6 1.88 and 1.7149.9 3.4053.3 3.7353.4 6.8862.2 no correlation71.5 3.35 and 4.0173.2 3.35 and 4.4298.1 no correlationa Solvent: CDC13.Raucher et al.4’ have reported that reaction of thecompound 42 with Lawesson’s reagent76 in 1,2-dimethoxyethaneat 65°C caused the chlorine to epimerize:26H’ 1,2—dietoxyethariel65°CCH3O ci. CH342In order to establish if this had happened in our case,41 was submitted for X-ray crystallography. The X-raydiffraction analysis of 41 (figure 2-3) clearly showed thatthe chlorine had remained in the exo position. During thework-up of 41 no other isomer was observed, and it seemsthat the remaining material was lost due to decompositionunder the reaction conditions.‘SC43LBFigure 2-3. X-ray diffraction analysis of thioamide 41.272.3. TRANSFORMATION OF THE EXO ISOMER 3340Treatment of the Diels-Alder adduct 33 with boranedimethyl sulfide complex in THF at room temperature andsubsequent oxidation with basic hydrogen peroxide at 0 - 4°Cgave the alcohols 44 and 45 in a combined yield of 73% andthe by-product 46 in 6% yield (scheme 2-5). Upon monitoringthe reaction by TLC it was found that the by-product 46 wasformed during the oxidation stage. Initially the oxidationwas carried out at room temperature leading to a 30% yieldof the two alcohols (44 and 45) and a 47% yield of the byproduct 46. However, lowering of the temperature of theoxidation to 0 - 4°C reduced the amount of by-product to 6%and increased the combined yield of the alcohols to 73%. Attemperatures below 0°C the oxidation became too slow. Themass spectra of 44 and 45 gave the expected molecular ionsat 353/355 m/z. The IR spectra of 44 and 45 showed a strongabsorption at 3440 cm and 3400 cm respectively, due tothe hydroxy group. Both alcohols are seen in the 1H NMRspectrum to exist as two conformers in a 46:54 ratio. In theNMR spectrum of 44 the H-i proton is a doublet (J = 2 Hz)at 4.67 ppm (major conformer). This establishes the hydroxygroup to be on C-6. In the -H NMR spectrum of 45 the H-iproton is a dd (J = 4 Hz, J’ = 1 Hz) at 4.65 ppm (majorconformer), thereby establishing the hyciroxy group to be inposition 5. The directional effect of the carbamate nitrogenmakes it reasonable to assume that the hydroxy groups in 44and 45 are endo as drawn in scheme 2-5. This assumption was280FC i-L-4.oI HCO2H31) BHS(CH321T F”25°CP) NaUH’307. ,THF/MEQH/O—4°C0Pho_LjCO2H3 CH349(737.)Cl.330 0P ho+ Cl0H +44HPC02CH345 46(6Z)(737.)PCCC H2 C 12/R ef lux0P470P hoN1+ Cl_1OCO2H348(207.)p—T sO H(527.)H3C CH3HO OH,benzeneirefluxScheme 2-5. Transformation of the Diels-Alder adduct 33.29Figure 2-4. X-ray diffraction analysis of alcohol 45.shown to be correct for alcohol 45, an X-ray diffractionanalysis of which is shown in figure 2-4. The X-ray of 45also confirmed the regiochemistry as well as thestereochemistry of the chlorine atom to be as predicted forthe Diels-Adler adduct 33, based on its 1H NMR spectrum.The mass spectrum of the by-product 46 showed amolecular ion at 301 m/z and did not indicate the presenceof chlorine. The absence of chlorine was confirmed by acorrect elemental analysis based on a molecular weight of301. The IR spectrum of 46 did not show the presence of ahydroxy group but the broad absorption at 1680 cm1indicated an ester and/or a carbamate group. The 1H NMRspectrum confirmed the presence of both a benzyloxycarbonyland a methoxycarbonyl group. A tricyclic structure such asC14otCA033046 seemed to be in best agreement with the spectroscopicdata and was finally confirmed by X-ray diffractionanalysis (figure 2-5).Figure 2-5. X-ray diffraction analysis the tricyclic byproduct 46.The separation of the two alcohols 44 and 45 turnedout to be very difficult and could only be achieved byiterative chromatography on 0.5 mm silica plates. Thecorresponding ketones, on the other hand could be separatedby column chromatography, and the mixture of the twoalcohols was therefore oxidized to the corresponding ketonesC13C14 c1zCt3cliclzcis01cii02 0204 043147 and 48 in a yield of 52% and 20% respectively (scheme 2-5). Alternatively, the oxidation was carried out on thecrude reaction mixture obtained after the hydroxylationwithout removing by-product 46 first. In this way 34% of 47,14% of 48 and 11% of compound 46 were obtained. Compared tothe Diels-Alder adduct 32 the regioselectivity in thehydroboration step of 33 is reduced. This is probably due tointeraction between the nitrogen lone pair and theperiplanar chlorine bond (and/or interaction with chlorinelone pairs), thereby reducing the capability of the nitrogenlone pair to interact with borane in the hydroboration step,thus leading to lower regioselectivity. The mass spectra of47 and 48 gave the expected molecular ions at 351/353 m/z.In the -H NMR spectrum 47 was found to be a 43:57 mixture oftwo conformers. H-i is a sharp singlet at 4.72 ppm (majorconformer), establishing the keto group to be in position 6.The ketone 48 was found to be a 48:52 mixture of twoconformers. The H-i proton at 5.06 ppm is a dd (J = 4 Hz,J’ = 2 Hz) showing the keto group to be in position 5.Treatment of the ketone 47 with 2,2-dimethyl-1,3-propanediol and p-toluene sulfonic acid in refluxing benzenegave the desired ketal 49 in 73% yield (scheme 2-5). Themass spectrum of 49 did not show the molecular ion at 437m/z but an ion at 402 corresponding to loss of chlorine. Acorrect elemental analysis confirmed the expected molecularformula as well as the presence of chlorine. The NMRspectrum showed 49 to be a 32:68 mixture of two conformers.32The reactivity of the carbamate group in 49 turned outto be quite different from that of the isomeric ketal 38 inthat the carbamate group in 49 could not be removed underthe conditions used for 38 (hydrogenation using 10%palladium on carbon in THF at 1 atm). Attemptedhydrogenation of 49 in a Parr hydrogenator at 30 - 35°C and48 psi was also unsuccessful. Treatment of 49 withiodotrimethylsilane (formed in situ42) did cause thestarting material to disappear on TLC but isolation of theanticipated amine was unsuccessful. The removal of thecarbamate group was finally achieved by passing dry hydrogenbromide through an anhydrous solution of 49 in benzene.C C H3Toluene, 60°C1 )H8rbenzeneCod2)cJ—( ,Et3NHCH3N, 25°C(757.)PClo.5çCH3CO2H3 CH349/H51H3C(71Z)H’52C H3Scheme 2-6. Synthesis of the amide 51 and the thioamide 52.33The amine hydrogen bromide was treated, without furtherpurification, with triethylamine and 3-indole acetylchloride41 in acetonitrile to give the desired amide 51 in71% yield based on the ketal 49 (scheme 2-6).Treatment of the amide 51 with either Lawesson’sreagent 5376 or reagent 5443 in toluene at 60°c gave thedesired thioamide 52 in 75 - 76% yield. However, HPLCanalysis of this compound revealed it to be a mixture of twocomponents. Reinvestigation of 52 by TLC confirmed the HPLCobservation, and the two components 52A and 52B weresubsequently isolated by preparative TLC on alumina plates,eluting iteratively with dichioromethane. TLC comparison of52A, 52B and the thioamide 41 under the same conditionsshowed conclusively that neither 52A nor 52B was identicalto the thioamide 41. The low resolution mass spectra of 52Aand 52B are virtually identical, both showing the expectedmolecular ion at 476/478 m/z for the desired thioamide 52.The -H NMR spectra of 52A and 52B showed severaldifferences, the most noticeable being the chemical shift ofH-i. In 52A H-i appeared at 7.05 ppm and in 52B at 5.05 ppm.The results of the 13C APT NMR spectra of 52A and 52B arelisted in table 2-3. Comparison of the 13C APT NMR of 52Aand 52B strongly suggests that both the carbon skeletons andthe patterns of substitution are identical. In order toverify that the signal at 7.05 ppm in 52A is due to H-i aHETCOR experiment was carried out. The result is listed intable 2-4.3422.0 21.8 22.022.5 22.0 23.126.9 27.1 27.429.3 30.1 29.137.7 34.6 37.140.3 36.3 41.241.2 39.7 42.650.3 52.6 49.953.2 53.3 53.356.3 62.5 53.466.7 65.5 62.271.6 70.1 71.572.5 70.3 73.298.3 98.6 98.1a The table is continued on the following page.b Solvent: CDC13.C See note at the end of the table on the next page.Table 2_3a Chemical shifts observed in the 13C APT NMRspectrum of thioamides 52A and 52Bb.52A 52B 416 (ppm) 1_c 8(ppm) 6(ppm) +!_C+++++++++++++++++++++++++++35Table 2-3 continued.52A 52B 418 (ppm) 1_b ö(ppm) 6(ppm)109.2 + 109.8 + 109.5 +111.3— 111.2 — 111.2—118.0- 118.4- 118.3-119.3— 119.1— 119.5—122.1— 121.6— 122.1 —122.5- 123.8- 122.3 -126.8 + 127.1 + 126.8 +135.9 + 135.8 + 136.0 +168.8 + 168.9 + 170.3 +203.8 + 204.1 + 202.9 +b The +1- sign indicate whether the signal s positive (>C<,-CH2) or negative ( >CH-, CH3-) in the 1 c APT spectrum.This experiment clearly demonstrates that the singlet at7.05 ppm is due to H-i. The mass spectral data and theAPT spectra of 52A and 52B strongly suggests that the twocompounds are structurally identical. The only conclusionpossible on basis of this physical data is that 52A and 52Bare unusually stable conformers of the same molecule 52.36Table 2-4. Chemical shift correlations obtained in theHETCOR spectrum of thioamide 52A in the region 0-100 ppma.shift (ppm) correlation26.9 2.1529.3 no correlation37.7 1.91 and 3.0640.3 1.7641.2 4.4150.3 3.53 and 3.6453.2 3.8756.3 7.0566.7 no correlation71.6 3.29 and 3.7372.5 3.38 and 4.26a Solvent: CDC13.The only region of the molecule that might give rise torestricted rotation is around the C-N bond in the thioamidegroup. Conformer 52A containing the most deshielded H-iproton must then be the conformer with sulfur syn to the H-lproton as depicted in figure 2-6. Supporting evidence forthis interpretation is the fact that heating either 52A or52B to 100°C in toluene results in both cases in theformation of a mixture of the two “compounds”. Furthermore,comparison of the 1H NMR spectra of 52A and 52B in DMSO-d637at room temperature and 120°C show that while the ‘H NMRspectra of the two “compounds” are different at roomtemperature then they are identical at 12000..c1 ClH3C. 02CH3 CO2H352BFigure 2-6. The conformers 52A and 52B of thioamide 52.24. PHOTOCHEMICAL CYCLIZATION OF THE AMIDES 40 AND 51, ANDOF THE THIOAMIDES 41 AND 52Irradiation of the amide 40 in aqueous methanol using aPyrex filter gave 4% of the desired lactam 55 and 15% of theundesired isomer 56 (scheme 2-7). Both compounds gave theexpected molecular ion at 424 m/z. The two compounds aredifferentiated by the chemical shifts of the indole protons.In the ‘H NMR spectrum of 55 the indole N-H is a singlet at9.04 ppm. The protons at 0-9 and 0-12 appear as doublets at7.54 ppm (J = 8.0 Hz) and 7.28 ppm (J = 8.0 Hz)respectively. The protons at 0-10 and C-il appear astriplets at 7.07 and 7.12 ppm. This assignment was confirmedI—H338RCO2H3C H355 R = 061 R = SScheme 2-7. Photochemical cyclization of the amides 40and 51, and the thioamides 41 and 52./I-1C H3 H3C H340 R = 041 R = ShuCH3+C H31256 R = 062 R = ShuH’C1H351 R = 052 R = S39by irradiation of the triplets at 7.07 and 7.12 ppm, wherebythe doublets at 7.54 ppm and 7.28 ppm collapse to singlets.In the ‘H NMR spectrum of 56 the indole N-H proton appearsat 8.72 ppm as a singlet. The four indole C-H protons appearas a doublet at 7.25 ppm (J = 8.0 Hz), a triplet at 7.09 ppm(J = 8.0 Hz), a singlet at 6.94 ppm and a doublet at6.84 ppm (J = 8.0 Hz). Irradiation of either the doublet at7.25 or at 6.84 ppm causes the triplet at 7.09 ppm tocollapse to a doublet. The doublets at 7.25 and 6.84 ppm aretherefore assigned as C-b and C-12 and the triplet at7.09 ppm to the proton at C-li. Irradiation of the singletat 6.94 ppm results in a sharpening of the indole N-H signalat 8.95 ppm establishing the singlet at 6.94 ppm as theproton at C-2.In order to improve the photochemical cyclization theinfluence of various factors on the yield of 55 wasexamined. The preliminary investigation summed up in table2-5 indicated that polar protic solvents are necessary forthe photochemical cyclization to take place. Comparison ofentry 20 with 21 suggests that the cyclization is favored byan increase in reaction temperature. In methanol (Entries 18- 20) the cyclization is about three times more efficient atwavelengths above 300 nm compared to shorter wavelengths.However, this observation does not hold true in aqueousmethanol where a cut off at 200 nm is only slightly worsethan a cut off at 220 nm, whereas longer wavelengths show amarked decrease in the yield. On the basis of these40Table 2_5a• Investigation of the effect of the nature ofthe solvent, the wave length and the temperature on thephotochemical cyclization of amide 40b•exp solvent nm temp (°C) % 55 % 561 THF 200 ref lux 0.0 1.02 THF 220 reflux 0.0 1.53 THF 300 reflux 1.0 0.54 THF/H20 300 ref lux 0.0 3.55 EtOAc 200 ref lux 0.0 0.06 EtOAc 220 reflux 0.0 0.07 EtOAc 300 ref lux 0.0 0.08 EtOAc 300 23 0.0 0.09 EtOAc/Et3N 300 23 0.0 0.010 EtOAc/HOAc 300 23 0.5 0.011 Benzene 200 reflux 1.5 0.012 Benzene 220 ref lux 1.0 0.013 Benzene 300 ref lux 1.0 0.014 CH3N 200 ref lux 0.0 0.015 CH3N 220 reflux 0.0 1.516 CH3N 300 ref lux 0.0 0.517 CH3N/H20 300 reflux 0.0 2.5a Table 2-5 is continued on the following page.41Table 2-5 continued.exp solvent nm temp (°C) % 55 % 5618 CH3Q 200 ref lux 0.0 4.019 CH3O 220 reflux 0.5 4.520 CH3Q 300 reflux 1.0 15.021 CH3O 300 23 1.0 7.022 CH3OH/H20 200 ref lux 0.0 13.523 CH3OH/H20 220 reflux 0.5 15.024 CH3QH/H20 300 ref lux 0.0 5.525 CHOH/H0/EtN 300 ref lux 0.0 2.026 CH3OH/H0/Na CO 300 23 0.0 4.5b The yields were determined by HPLC (conditions asdescribed in section 3.26) using pure 55 and 56 asexternal standards.Table 2_6a• Investigation of the photochemical cyclizationof amide 40 using fractional factorial design with fourfactors on three levels.Factor level 1 level 2 level 3temp: temperature (°C) 0 25 65base: none N-methyl triethylmorpholine aminenm: wavelength (nm) 220 300 260solv: solvent MeOH THF CH3Na The table is continued on the following page.42Table 2-6 continued.exp temp base nm solv %55b %56b1 0 none 260 THF 0 22 25 none 220 CH3O 0 123 65 none 300 CI{3N 0 04 0 Me morph 300 CH3O 0 15 25 Me morph 260 CH3N 4 146 65 Me morph 220 THF 0 07 0 Et3N 220 CH3N 0 58 25 Et3N 300 THF 0 09 65 Et3N 260 CI-130H 4 71/3 El (55) 0.0 0.0 0.0 1.31/3 E2 (55) 1.3 1.3 0.0 0.01/3 E3 (55) 1.3 1.3 2.7 1.31/3 El (56) 2.7 4.7 5.7 6.71/3 E2 (56) 8.7 5.0 0.3 0.71/3 E3 (56) 2.3 4.0 7.7 6.3b The yields were determined by HPLC (conditions asdescribed in section 3.26) using pure 55 and 56 asexternal standards.43Table 2-7. Investigation of the photochemical cyclization ofamide 51 using fractional factorial design with four factorson three levelsa.exp temp base nm solv %55b %56b1 0 none 260 THF 0 112 25 none 220 CH3O 0 53 65 none 300 CH3N 0 24 0 Me morph 300 CH3O 0 85 25 Me morph 260 CH3N 0 286 65 Me morph 220 THF 31 207 0 Et3N 220 CH3N 0 178 25 Et3N 300 THF 0 09 65 Et3N 260 CH3O 35 221/3 E 1(55) 0.0 0.0 10.3 10.71/3 E 2(55) 0.0 10.3 0.0 10.31/3 E 3(55) 22.0 11.7 11.7 0.01/3 E 1(56) 12.0 12.7 20.7 18.31/3 E 2(56) 17.7 18.7 3.3 10.31/3 E 3(56) 14.7 13.0 20.3 15.6a The levels of the factors are defined in table 2-6.b The yields were determined by HPLC (conditions asdescribed in section 3.26) using pure 55 and 56 asexternal standards.44preliminary findings a more systematic investigation wascarried out on the two isomeric amides 40 and 51. In orderto make the investigation as efficient as possiblefractional factorial design was used as the experimentalstrategy44. A design with four factors, each on threelevels, was chosen, as shown in table 2-6 for the amide 40and as shown in table 2-7 for the amide 51. The fractionalfactorial design dictates how the four factors are to bevaried in the 9 experiments while all other factors are keptconstant. In the last six rows in each table the averageyield of 55 and 56 are summed up for each level of the fourfactors. A graphic representation of the results is shown infigure 2-7. It should be noted that in the followinganalysis of the fractional factorial design it is assumedthat no interactions exist between the four factors. As inthe preliminary investigation of 40 all the factors examinedin table 2-6 had only a very small effect on the yield ofthe desired product 55. However, for the isomeric amide 51all the factors examined show some influence on theformation of 55. From the results in table 2-7 (and figure2-7) it is clear that the photochemical cyclization of 51 to55 is favored by high reaction temperature, presence of baseand wave lengths below 300 nm. The observed effect of thedifferent solvents is rather odd. That cyclization shouldtake place in a non-polar aprotic solvent and in a polarprotic solvent but not in an aprotic polar solvent ofcomparable transparency seems unlikely. A closer look at the45Cyclization of 40 (o) and 51 () to 55X YIELD Z YIELD X YIELD X YIELD30 TEMPERPTURE OflSE 30 WPVE LENGTH SOLVENT20 20 20 2010 / 10 // °_A2 Zr0 25 65 None Et3N 220 260 300 THF MeCN MeDHMe—morphCyclization of 40 (a) and 51 (es) to 56X YIELD X YIELD X YIELD X YIELD30TEMPERflTURE BSE30 IJPVE LENGTH 30 SOLVENT10 :: A0 25 65 None Et3N 220 260 300 THF MeCN MeOHrle—riiorphFigure 2-7. A graphic representation of the results of thefractional factorial designs in table 2-6 and 2-7.levels of the other factors in the three experiments inacetonitrile (table 2-7 entry 3, 5 and 7) reveals that ineach of these experiments at least one of the other factorsis at a level where the graphic analysis in figure 2-7 showsthat no reaction occurs. This is the most likely the reasonwhy no reaction is observed in acetonitrile. The fact thatno response is observed in a majority of the experiments inthe fractional factorial design makes the interpretation of46the design less reliable. In contrast to the desired product55, the undesired product 56 is readily formed from bothamides under most of the different reaction conditionsexamined. The effect of the different factors on thecyclization of the two amides to 56 is seen to follow asimilar trend.Raucher et al.34 reported that photochemical cyclizationof the amide 57 gave only trace amount of the desiredproduct 59, but that the corresponding thioamide 58 cyclizedin 41% yield to the thiolactam 60.hvC H3On the basis of this report the thioamides 41 and 52were synthesized from the corresponding amides. Irradiationof thioamide 41 in aqueous methanol using a Vycor filtergave 11% of thiolactam 61 and 10% of the undesired isomer 62(scheme 2-7). Both compounds gave the expected molecular ionat 440 mhz. The two thiolactams are differentiated by theindole protons in their ‘H NMR spectra. In the NMRspectrum of 61 the N-H proton is a singlet at 7.86 ppm. TheH-9 and H-12 protons appear as doublets at 7.58 ppm (J = 7.447Hz) and 7.28 ppm (J = 7.4 Hz). The H-b and H-il protons aretriplets appearing at 7.17 ppm (J = 7.4 Hz) and 7.13 ppm (J= 7.4 Hz). Irradiation of the proton at 7.58 ppm causes thetriplet at 7.13 ppm to collapse into a doublet, confirmingthe assignments. In the 1H NMR spectrum of 62 the N-H protonis a singlet at 8.50 ppm. The remaining four indole protonsresonate as a doublet at 7.29 ppm (J = 7.0 Hz), a singlet at7.22 ppm, a triplet at 7.11 ppm (J = 7.0 Hz) and a doubletat 6.81 ppm (J = 7.0 Hz). Irradiation of the triplet at7.11 ppm causes, as expected, the doublets at 7.29 ppm and6.81 ppm to collapse to singlets. For steric reasons C-9 onthe indole ring must be substituted, since a molecular modelshows that the quinuclidine ring system cannot stretchacross to C-12 in the indole ring.Six factors were screened for their effect on thephotochemical cyclization of 41 using a fractional factorialdesign with seven factors each on two levels. The seventhvariable which the chosen design allowed was not assigned toany factor but was instead used to give some idea of thesize needed for a response to be significant. Table 2-8shows the design and the results obtained. The effect of thedifferent factors on the yield of the desired thiolactam 61is given in column E(61). Only the temperature and thesolvent have any significant effect on the yield of 61 andcomparison of the two levels for these two factors showedthat the better yield of 61 was obtained at 65°C and inmethanol. For the undesired product 62 the presence or48absence of triethylamine and the light frequency have thelargest effect on the yield with the presence oftriethylamine and wavelength up to 220 nm being thepreferred levels of these two factors. The remaining fourfactors have only a minor or insignificant effect on theyield of 62. In order to further investigate the mostimportant factors found in the experiments above afractional factorial design with four factors each on threelevels was chosen. The general plan, as well as the resultsof the design, is shown in table 2-9. A graphicalrepresentation of the results is given in figure 2-8.Table 2_8a Investigation of photochemical cyclization ofthioamide 41 using fractional factorial design with sevenfactors on two levels.Factor level 1 level 2time: time (mm) 20 40temp: temperature (°C) 25-30 65-85base: base (Et3N) absence presencemt : light intensity low highnm : wavelength (nm) >220 >300solv: solvent MeOH MeOH\H20 (3:2)a The table is continued on the following page.CD-)C’01Ca)t’JHCDMMM0)(DQts)HH—S--SHctU)(DC’C’C’C’(D(DC)L%J•%)HHL’J)HHr1I—’(DHF’.)01)[“30L’3H[‘a)H[‘.)H[“3HI-’.c1,QC)Cl)Cl)01Ca)CD01CDCa)CD0rtHCDrFjCflM[‘a)[‘a)HCDQ)(D(DHC.a)01IS.)[‘.3HH[“3IS.)HHI-jo...01CDCa)01Ca)CDU)H-CDOrtCDHCa)HHHHC’aIS.)[‘.300)••••••IS.)HHIS.)Ht-’.)[“3HU)•H•01Ca)CD00101CD[‘S3IS.)[‘a)HHCa)[“301[‘.30IS.)HHHH[“3[‘3[“3[‘-.3(J1Ca)CD00101H-H[“a)HH‘.0CDH0[‘.3[-.3H[‘a)HIS.)[‘s.)H(QIj•.••010101CDCa)C’)[-.3[‘a)H0Ca)01IS.)01CDHH[‘3[‘.3[‘3IS.)HHH(DO•••••<100010101CDCA)IS.)[‘3HHQJc-I-IS.)01Ca)Ca.)0Ca)IS.)HH••Q’O000000C’U)[“30)o\O0)0)HHHHHC’(I)‘.‘.0—J01H[‘-.3i01HCa.)IS.)HCa)[‘.3IS.)(a)0001CDi01IS.)CDC’ [‘a)50Changing the solvent from tetrahydrofuran to acetonitrile ormethanol results in a significant increase in the yield ofboth products. The photochemical cyclization is clearlyfavored by polar solvents. The examined variations in thetemperature, base type and wavelength have only a smalleffect on the yield of the desired thiolactam 61. The trendobserved for the temperature is opposite to that indicatedin the two level design in table 2-8. The cause of thisdiscrepancy is not clear and would require furtherinvestigation to be unravelled. The effect of thetemperature on the formation of 62 is more pronounced, witha significant increase in the yield going from 25 to 10°C.Table 2_9a• Investigation of the photochemical cyclizationof thioamide 41 using fractional factorial design with fourfactors on three levels.Factor level 1 level 2 level 3temp: temperature (°C) 10 25 65base: N-methyl triethyl diisopropylmorpholine amine aminenm: wavelength (nm) 260 300 330solv: solvent MeOH THF CH3Na The table is continued on the following page.51Table 2-9 continued.exp temp base nm solv %61b %62b1 10 Me morph 330 THF 16 142 25 Me morph 260 CH3O 22 383 65 Me morph 300 CH3N 20 304 10 Et3N 300 CH3O 21 395 25 Et3N 330 CH3N 16 286 65 Et3N 260 THF 10 57 10 ipr2NH 260 CH3N 20 518 25 i-pr2NH 300 THF 11 109 65 i-pr2NH 330 CH3O 16 291/3 El (61) 19 19 17 201/3 E2 (61) 16 16 17 121/3 E3 (61) 15 16 16 191/3E1 (62) 35 27 31 351/3E2 (62) 25 24 26 101/3E3 (62) 21 30 24 36b The yields were determined by HPLC (conditions asdescribed in section 3.26) using pure 61 and 62 asexternal standards.52Cyclization of 41 to 61 () to 62 (o)7. ‘IIELO 7. IIELO X 13ELO 7. YIELOTEMPERATURE BASE WAVE LENGTH SOLVENT:: :: 70:: 0::10 10 10 10 0I I I I I10 25 65 Me—morph j—pr2NH 260 300 330 THF MeCN IleOHEt3NFigure 2-8. A graphic representation of the results of thefractional factorial design in table 2-9.As found in the design in table 2-8 the yield drops withincreasing wavelength. In order to determine if yield couldbe increased further by increasing the solvent polarityexperiment 3 in table 2-9 was repeated using methylformamide (E = 111) instead of acetonitrile (E = 36). Ayield of 42% of 62 and 8% of 61 was obtained. This change insolvent and solvent polarity gave the same total yield ofcyclization as in experiment 3 but profoundly changed theratio of the two products. In the fractional factorialdesign in table 2-8 it was found that the formation of theby-product 62 was strongly disfavored in the absence of abase, whereas the desired thiolactam 61 was unaffected. Inorder to check if the disfavoring of 62 could be translatedinto an increased yield of 61 experiment 2 in table 2-9 was53repeated but without any base. A 22% yield of 61 and 14%yield of 62 was obtained. Thus, the decrease in formation of62 did not result in an increase in the yield of 61.Lowering of the temperature to —10°C under these conditionsdid not have any effect on the yield of the two products.The photolysis of the isomeric thioamide 52 wasexamined under the conditions found to give the best yieldof compound 61 from the thioamide 41. The results are shownin table 2-10. Both the pure conformers 52A and 52B wereTable 2-10. The photochemical cyclization of thioamide 52compared to thioamidecompound time % 61 % 62 % 41 % 52A % 52B * 52(mm) (total)41 15 22 14 0-- -52A 15 1 1 — 40 29 6952B 15 2 1- 33 30 6352A 75 3 2 - 22 17 3952B 75 4 2- 17 14 31a Conditions: MeOH, >260 nm, 25°C, no base.b The yields were determined by HPLC (conditions asdescribed in section 3.26) using pure 61 and 62 asexternal standards.54used in these studies. After 15 mm photolysis all of thestarting material 41 had been consumed and 22% of thedesired product 61 had formed. In the same period of timeonly one third of the starting material of 52A and 52B hadbeen consumed, producing only 1 - 2% of the desired product61. The remaining and initially pure 52A and 52B had now, inboth cases, become a mixture of the two conformers. It isclear from table 2-10 that irradiation of the thioamide 52leads mainly to decomposition. It appears that the decay ofexited 52 occurs mainly through translation into rotationalenergy around the thioamide bond and through some chemicalprocess other than cyclization.The mechanism of the photochemical cyclization ofaromatic a-chloroacetamides has been studied in somedetail45 and even though less work has been carried out onthe corresponding ct-chloroacetates, the results obtainedindicate that they behave similar to the achloroacetamides45.Assuming that the results obtained forthe aromatic a-chloroacetamides holds true for thecorresponding a-chloroacetates, the mechanism ofphotochemical cylization of aromatic a-chloroacetates can berepresented by the simplified model outlined in scheme 2-8.The excited molecule 64 has several possible fates: 1) Itmay decay to the ground state. 2) It may undergo homolyticcleavage of the carbon-halogen bond. 3) It may undergoelectron transfer from the aromatic system to the acetategroup and subsequent heterolytic cleavage of the carbon-55halogen bond. Homolytic scission of the carbon-chlorine bondhas been found generally to be followed by protonabstraction from the solvent or from the molecule itselfrather than cyclization. The major route to cyclization hasbeen found to be electron transfer from the aromatic ringsystem to the acetate group followed by heterolytic cleavageof the halogen bond and subsequent ring closure of the thusformed diradical. As it is seen from figure 2-7 the amide 51gave a much higher yields of cyclized products than theamide 40. The higher reactivity of 51 can be explained by ahigher reactivity of the chlorine bond due to itscoplanarity with the lone pair of the carbamate nitrogen.That the overall yield of the cyclization for both amides 40and 51 (figure 2-7) and the thioamide 41 (figure 2-8)increased with increasing polarity of the solvent, pointstowards a mechanism involving charged intermediates. It istherefore reasonable to assume that the photochemicalcyclization of the investigated amides and thioamides followa mechanistic path involving electron transfer andsubsequent heterolytic cleavage of the carbon-halogen bondas outlined in scheme 2-8. This reaction mechanism couldalso explain why the thioamide 41 gave a much higher yieldof cyclized products than amide 40, in that the introductionof sulfur probably makes the electron transfer easier byinitially accepting the electron from the aromatic ringsystem and subsequently transferring it the a-chloroestergroup. Model studies have indicated that the diradical560II163jhv0*CH21J ®IcI._0CH2’D1[—cH2..c1]63 67C 1 LC H2C 1]65HCl.:][H=°1Scheme 2-8. Possible path ways in the photochemicalcysclization reaction of aromatic a-chloroacetates.formed upon the release of the chloride ion has so short alifetime that cyclization takes place from the dominantconformer of the starting material in its ground state48.The change in the ratio of the desired lactam 55 to the06657undesired lactam 56 as well as in the yields of the twocyclized products by increasing the reaction temperaturemight be a reflection of a shift in the equilibrium of thestarting conformers. Model studies on N-chloroacetyl-3-indolylamines have shown that cyclization to position 4 onthe indole ring is preferred in a 6:1 ratio to the 2position49. This correlates with the general preference forthe 4 position observed for both the amides and thethioamide 41 resulting in the undesired isomer.2.5. EVALUATION OF THE CHOSEN STRATEGY.Having investigated the photochemical cyclization ofthe amides 40 and 51 and thioamides 51 and 52 to theibogamine skeleton the time has come to evaluate the resultsof this strategy against the goal of the project, namely aviable synthetic route to vinbiastine 1. The overall yieldsfrom the Diels-Alder adducts to the desired lactam 55 andthiolactam 61 are shown in figure 2-9. Taking intoconsideration that at least another five steps remain in thesynthesis of viriblastine 1 the low overall yield of theformation of the ibogamine skeleton makes the chosenstrategy unfeasible as a synthetic route to vinbiastine 1and further work on this strategy was therefore stopped atthis point.58PCO2H30H33Figure 2-9. The overall yields of cyclized products 55 and61 from the Diels-Alder adducts 32 and 33.2.6. THE MODIFIED POLONOVSKI REACTIONThe discovery by Max and Michael Polonovski50, thattertiary amines can be demethylated by treatment of thecorresponding N-oxides with acetic anhydride, has provided aversatile tool in alkaloid synthesis51. The mechanism of thereaction is outlined in scheme 2_952. The intermediate02C55CH30s32H6102H359species 7O and have been isolated. It has beenfound that the nature of the acylating reagent has adramatic effect on the regiochemical course of thereaction55. In the modified Polonovski reaction56, when theamine N-oxide is treated with trifluoroacetic anhydrideinstead of acetic anhydride, the diminished nucleophilicityof the trifluoroacetate ion disfavors the formation ofintermediate The modified Polonovski reaction mainlyfollows the two pathways shown in scheme 2-10. In pathway Athe most acidic proton a to nitrogen is abstracted, givingt0flcI i-,. flc20 I—C—N——--I I- -HOPcH H670flc 0\+/ I / flc20C_N —C—N/ \ Oflc71 7211+ \ \—C--N—flc—b N—Pc + iiO + Pc20/ /CH3—O0Pc 7573Scheme 2-9. The mechanism of the Polonovski reaction.60the “normal Polonovski intermediates. In pathway B afragmentation reaction takes place in which the C-C bondadjacent to the C-N bond is cleaved. According to Grob’srules57 for hetereoatomic fragmentation reactions the C-Cbond to be cleaved must be antiperiplanar to the N-C bond.cF3C00-0 cF3001+ 1+—N—C---—C—C-—H I r —N—C—C——C—HI I I ICF3OO -CF3COO B\+ /N=C/ SC—C—H CF3OO CF3OQI I \+ / +1 IN=zC C—C—H(0 // HI”\ /Nu—C—C—H CCII /\81 82Scheme 2-10. The pathways in the modified Polonovskireaction.Kutney et al.2° and Potier et ai.21 were the first toutilize the modified Polonovski reaction in the coupling of61catharanthine 3 with vindoline 4 to give 15’,20’-anhydro-vinbiastine 5 with the natural stereochemistry (S) at C-16(scheme 2-11). The best yield of 5 was found to be 50% byboth groups. Under optimal conditions for the formation of15’,20’-anhydrovinblastine 5 Potier et al. also found 12% ofepi-15’,20’-anhydrovinblastine 21• It was shown by bothgroups that the amount of epi-15’,20’-anhydrovinblastine 9increases with increasing reaction temperature201.1) rn—Cl PBP2) VINDOLINE3) TFPfl, —50°C-C H3H3 CVINDOLINECO2H39Scheme 2-11. The coupling of catharanthine 3 with vindoline4.As a reasonable rationale for the observed Polonovskifragmentation process two routes were suggested as outlined302CH3INDOLINE562in scheme 2-12: 1) A concerted mechanism in which the attackby vindoline on C-16’ and the fragmentation of the C-16’,C-21’ bond takes place simultaneously. 2) A stepwise mechanismin which fragmentation of the C-16’,C-21’ bond generates anintermediate which subsequently reacts with vindoline togive the dimer. Molecular models suggested that, for stericreasons, vindoline must approach from the a-face ofcatharanthine 3, leading to the dimer with the naturalconfiguration. For this reason it is difficult to imaginethe formation of the dimer with unnatural configurationthrough a concerted process. However, a stepwise mechanismwould be able to explain the formation of both dimers. Ifthe initially formed intermediate 85a remains in the “iboga”conformation, attack by vindoline 4 would occur from the aface leading to the dimer with the natural conformation. Byconformatiorial change of 85a to the conformer 85b the 3-faceof the indole unit becomes the more accessible, leading tothe unnatural configuration at C-16’ upon attack byvindoline 4. It is reasonable to suppose that an increasedreaction temperature might lead to a conformational changefrom 85a to 85b giving an increase in dimer with theunnatural configuration. It is also possible that aconcerted and a stepwise mechanism are operatingsimultaneously and that the latter is favored by increasingtemperature.6385bCF3OOScheme 2-12. Possible mechanisms for the coupling ofcatharanthine 3 with vindoline 4.Concert.ed processVINDOLINE‘Co CF3CF3OO-C H3-C H35Stepuise process02CH3INDOLINECo CF383•CH3844:02CH3VINDOLINEI85a-C H3-C H3 IH-CH3I NDOL INE9INDOLINE3564The coupling of various catharanthine derivatives ofcatharanthine 3 with vindoline 4 has been examined (table 2-11). It was found that the yield of dimers is stronglydependent on the reaction conditions20’1 and that none ofthe examined catharanthine derivatives gave as good a yieldas catharanthine 3 itself.Table 2-11. Coupling of catharanthine derivatives withvindoline 4.VINDOLINE88a —R3•1Compound R1 R2 R3 R4 R5 0 Reference87a—87fH INDOLINEa = Et H H 50 20,21b H H Et H H 10 21bb H H Et H H 20a 20bC = H Et H 20 21bd H H H Et H 11 21be -0- Et H H 20 58e -0- Et H H 6 59f H H Et H OAc 6 60R2=H, R3=H, R4=H, R5=H.a +20% where R1=Et,65In the strategy chosen by us for the synthesis ofvinbiastine 1 it is necessary to couple the catharanthinederivative 16 or 29 with vindoline 4 to form the appropriatedimeric intermediate. Based on the literature it was clearthat a moderate to significant drop in the yield of dimer byusing either 16 or 29 had to be anticipated. This, combinedwith the provisional status of the Polonovski reaction, withrespect to the mechanism, as well as the factors affectingthe yield of the desired dimer, prompted a reinvestigationof the Polonovski reaction primarily in order to determinethe factors having the most significant effect on the yield.C H32.7. SYNTHESIS AND CHARACTERIZATION OF EXOCATHARANTHINE 89Utilization of catharanthine 3 itself in thereinvestigation of the Polonovski reaction is complicated bytwo factors. Firstly, catharanthine N-oxide 10 is unstabledue to a facile [2,3] sigmatropic rearrangementreaction2b,Th making the isolation of the pure N-oxidevery difficult. Secondly, the formed 15’,20’-anhydro-1666vinbiastine 5 reacts very readily with oxygen under theformation of the corresponding epoxide16.Thus we decided to look at exocatharanthine 89, astructurally closely related derivative of catharanthine 3.6 538 4N141513 2 16 21 \\cH3C02H3 19 1822 23H H10510HvindolineH5 88eindoline8967A preliminary investigation showed that exocatharanthine N-oxide 90 did not undergo the rearrangement reaction observedfor catharanthine N-oxide 10 and could easily be isolated inits pure form. 19’,20’-Anhydrovinblastine 91 did not formthe corresponding epoxide when exposed to the air, andappeared more stable than 15’,20’-anhydrovinblastine 5.Furthermore 19’,20’-anhydrovinblastine 91 opened up anopportunity to compare the reactivity of the 19’,20’-doublebond in 91 with the 15’,20’-double bond in 5, as well as apotential new route to vinblastine 1.Exocatharanthine 89, a synthetic isomer of catharanthine 3, was first reported by Szantay eL a146. The compoundwas formed as a side product during the catalytichydrogenation of catharanthine 3 in methanol, usingpalladium on carbon as catalyst, in a yield of 20 - 30%. Noeffort was done at the time to optimize the yield ofexocatharanthine 89. Exocathararithine 89, formed in thisway, exhibited a typical indole UV spectrum and the lowresolution mass spectrum showed a molecular ion at 336 m/z.A molecular weight of 336 was confirmed by a correctelemental analysis. This molecular weight is the same asthat of catharanthine 3. However, TLC clearly showed thatexocatharanthine 89 is not identical to catharanthine 3, andthis fact pointed towards that exocatharanthine 89 might bean isomer of catharanthine 3. The 13C APT spectrum ofexocatharanthine 89 (table 2-13) showed the samedistribution of the 21 carbons on the different “bond types”68found for catharanthine 3, confirming thatexocatharanthine 89 is an isomer of catharanthine 3. TheTable 2-12. Chemical shifts observed in the ‘H NMR spectrumof exocatharanthine 89aShift (ppm) # H Correlation1.56 Cd)1.81 (d)2.16 (m)2.32 (s)2.78 (dt)2.90-3.053.08-3.163.22-3.413.43-3.573.70 (s)4.00 (s)5.36 (q)7.07 (t)7.16 Ct)7.25 (d)7.50 (d)7.65 (s)H-18H-17H-14H-15H-17H-6, H-3H-3H-5, H-6H-5H-23H-21H-19H-10H-ilH-12H-9N-Has that(m)Cm)Cm)(m)31121212131111111a Solvent: CDC13.69Table 2-13. Chemical shifts observed in the ‘3C APT NMRspectrum of exocatharanthine 89 (solvent CDC13).Shift (ppm) 1_a Correlation12.85 C-lB21.44 + C-627.31 C-1429.72 + C-1537.32 + C-1750.28 + C-352.72 C-2353.02 + C-555.53 + C—1663.59 C-21110.34 C-il110.38 + C-7118.22 C—9118.36 C—19119.36 C—10121.83 C—12128.72 + C-8135.12 + C-13136.98 + C-2137.09 + C-20174.47 + C-22a indicate whether the signal is positive or negative.70Table 2-14. Chemical shift correlations obtained in theHETCOR spectrum of exocatharanthine 89a•13C correlation 13C ‘H correlation(ppm) (ppm) (ppm) (ppm)12.85 (C—18) 1.56 110.34 (C—il) 7.1621.44 (C-6) 2.95 and 3.28 110.38 (C-7) -27.31 (C-14) 2.16 118.22 (C-9) 7.5029.72 (C—15) 2.32 118.36 (C-19) 5.3637.32 (C-17) 1.81 and 2.78 119.36 (C-iC) 7.0750.28 (C-3) 2.99 and 3.10 121.83 (C-12) 7.2552.72 (C-23) 3.70 128.72 (C-8)-53.02 (C-5) 3.30 and 3.50 135.12 (C-13) -55.53 (C—16) none 136.98 (C—2) —63.59 (C-21) 4.00 137.09 (C—20)-174.47 (C-22) —a Solvent: CDC13.71Table 2-15. Chemical shift correlations obtained in theCOSY spectrum of exocatharanthine 89aShift (ppm) # H Correlation1.56 (d)1.81 (d)2.16 (m)2.32 (s)2.78 (dt)2.90-3.053.10 (m)3.22-3.413.50 (m)3.70 (s)4.00 (s)5.36 (q)7.07 (t)7.16 (t)7.25 (d)7.50 (d)7.65 (s)5.36,2.78,3.10,5.36,3.10,3.50,2.78,3.50,3.30,2.32.2.32, 2.16.2.78, 2.32, 1.81.2.95, 2.16, 1.81.2.16, 1.81.3.30.2.16.2.952.95.(m)(m)311212121311111112.32, 1.56.7.50.7.07.a Solvent: CDC13.72Table 2-16. Observed enhancements in SINEPT experiments onexocatharanthine 89a•Irradiation (ppm) Observed enhancement (ppm)5.36 (H-19) 63.59 (100%), 29.72 (29%).4.00 (H—21) 136.98/137.09 (100%), 50.28 (17%).1.56 (H—18) 137.09 (100%).a Solvent: CDC13.Table 2-17. Observed enhancements in NOE experiments onexocatharanthine 89aIrradiation (ppm) Observed enhancement (ppm)5.36 (H-19) 4.00 and 1.56.4.00 (H—21) 5.36 and 3.50.3.70 (H—23) 7.65, 7.25 and 5.36.1.56 (H-18) 5.36 and 2.32.a Solvent: CDC13.73highest field signal in ‘H NMR spectrum of exocatharanthine89 (table 2-12) is a doublet at 1.56 ppm (J = 6.0 Hz)integrating to three protons. This signal was assigned tothe methyl group 0-18. The COSY spectrum (table 2-15) showedthat the methyl group was coupled to the olefinic proton at5.36 ppm, thereby establishing exocatharanthine 89 to be adouble bond isomer of catharanthine 3, with the double bondshifted from position C-15/C-20 in catharanthine 3 toposition C—15/C-19 in exocatharanthine 89.Position 15. 18. 19, and 21:As mentioned above the ‘H NMR spectrum (table 2-12) andthe COSY spectrum (table 2-15) established H-18 and H-19 at1.56 ppm and 5.36 ppm respectively. The HETCOR spectrum(table 2-14) showed C-18 to be at 12.85 ppm and C-19 to beat 118.36 ppm. The lowest field aliphatic signal at 63.59ppm, with a negative amplitude in the 13C APT spectrum, wasassigned to 0-21. In the HETCOR spectrum this carbon signalwas correlated to the proton at 4.00 ppm. The SINEPTexperiments listed in table 2-16 are optimized for thedetection of three bond couplings with a coupling constantof 7 Hz between C and H by irradiation of the appropriatehydrogen. Irradiation of the olefinic proton at 5.36 ppm (H19) resulted in a maximum enhancement at 63.59 ppmcorresponding to 0-21 and a smaller enhancement at 29.72 ppmindicating 0-15. The 13C APT spectrum confirmed that thesignal at 29.72 ppm was a methylene group and the HETCOR74spectrum established the corresponding protons to be at 2.32ppm (table 2-14). The NOE experiments listed in table 2-17shows that irradiation at 5.36 ppm (H-19) resulted in an NOEenhancement at H-21 (4.00 ppm) as well as at the methylgroup at 1.56 ppm. Irradiation of the methyl group at 1.56ppm gave a NOE enhancement at H-19 (5.36 ppm), but not at H-21. Instead the irradiation of the methyl group resulted inan enhancement at 2.32 ppm, confirming this signal to belongto the C-15 methylene group, and firmly establishing thestereochemistry of the double bond to be E.Next to the H-19 proton in the NMR spectrum isobserved a small quartet (J = 6 Hz) at 5.26 ppm integratingto 1/6 of a proton. The COSY spectrum established that thisproton was coupled to a doublet at 1.65 ppm integrating to1/6 of a methyl group. These to signals are assigned to H-19and H-18 in the corresponding Z-isomer of exocatharanthine,and from the HETCOR spectrum C-18 and C-19 (in the z-isomer)were found to resonate at 13.10 and 118.36 ppm respectively.Position 14 and 23:The remaining aliphatic carbons with an odd number ofhydrogens in the 13C APT spectrum resonate at 27.31 and52.72 ppm and are assigned to 0-14 and C-23 respectively.The HETCOR spectrum correlated H-14 at 2.16 ppm and H-23 at3.70 ppm.75Position 3, 5, 6, 16 and 17:Of the remaining aliphatic carbons 3, 5, 6, 16 and 17,C-16 was assigned to the signal at 55.53 ppm, the only oneof the five carbon signals that did not show correlation toany hydrogens atoms in the HETCOR experiment. The remainingfour aliphatic carbon, all methylene groups, were found inthe 13A APT spectrum to resonate at 21.44, 37.32, 50.28 and53.02 ppm. In the COSY experiment listed in table 2-15 theH—14 proton at 2.16 ppm was found to couple to protons at3.10, 2.78, 2.32, and 1.81 ppm. Decoupling of H-14 at 2.16ppm in the ‘H NMR revealed, besides confirming the observedcross peaks, a coupling to a proton in the multiplet 2.90 -3.05 ppm. Consequently coupling from H-14 to all the vicinalprotons were accounted for. The proton signal at 2.32 ppmwas correlated above to C-15 at 29.72 ppm. The HETCORexperiment listed in table 2-14 correlated the protonsignals at 3.10 and 2.99 ppm and to the carbon at 50.28 ppmand the proton signals at 2.78 and 1.81 ppm to the carbonsignal at 37.32 ppm. Taking the size of the chemical shiftsinto consideration the carbon signals at 50.28 and 37.31 ppmare assigned to C-3 and C-17 respectively. Based on theirchemical shifts the remaining carbon signals at 53.02 and21.44 ppm are then assigned to C-5 and C-6 respectively. TheHETCOR experiment established the H-5 protons to be at 3.30and 3.50 ppm and the H-6 protons to be at 2.95 and 3.28 ppm.76Position 1, 9, 10, 11 and 12:The singlet observed in the NMR spectrum at 7.65 ppmwas readily assigned to the indolic N-H proton. Irradiationof the protons in methoxy group at 3.70 ppm (table 2-17)resulted in a small NOE enhancement of the N-H proton aswell as the doublet at 7.25 ppm establishing this proton tobe the indolic proton H-12 (table 2-17). Thus the otherdoublet at 7.50 ppm must be H-9. From the COSY spectrumlisted in table 2-15 it was found that the doublet at 7.50ppm was coupled to the triplet at 7.07 ppm therebyestablishing this signal to be H-b. By way of exclusion itfollowed that the triplet at 7.16 ppm must be H-li. From theHETCOR experiment (table 2-14) C-9, C-b, C-li and C-12 werefound to resonate at 118.22, 119.36, 110.34 and 121.83 ppmrespectively.Position 2, 7, 8, 13, 20, 22:The carbonyl C-22 is readily assigned to the chemicalshift at 174.47 ppm. Irradiation of the methyl group at 1.56ppm resulted in enhancement at the carbon at 137.09 ppm(table 2-16), which therefore was assigned to C-20.Irradiation of H-21 at 4.00 ppm resulted in a enhancement ofthe carbon signals at 137.09, 136.98 and 50.28 ppm. Thesignal at 136.98 was assigned to C-2. By comparison with thecarbon spectrum of catharanthine (table 2-18), the carbonsignals at 110.38, 128.72 and 135.12 ppm were assigned to C7, C-8 and C-13 respectively.77Table 2-18. Assignment of the chemical shifts ofcatharanthine 3a•Position ‘3C (ppm) Position 13C (ppm)2 136.0 13 134.73 52.9 14 30.45 49.3 15 123.46 21.0 16 55.07 110.2 17 38.08 128.4 18 10.59 117.7 19 25.910 118.9 20 148.511 121.3 21 61.512 110.2 22 173.623 52.0a Solvent: CDC13.The numbering of the carbon atoms in exocatharanthine89 is shown below and the assignment of the chemical shiftsfor 89 are summarized in table 2-19.6 5“013N2 21 20HCO2H3 19 1822 2378Table 2-19. Assignment of the ‘H and 13C chemical shifts ofexocatharanthine 89 (solvent CDC13).Position12356789101112‘314151617181920212223H ã(ppm) C ö(ppm)136.9850.2853 .0221.44110.38128. 72118.22119.36110.34121.83135.1227.3129.7255.5337.3212.85118.36137.0963. 59174.4752. 727.652. 99/3. 103.30/3.502.95/3.287.507.077.167 .252.162.321.81/2.781.565.364.003.7079In an attempt to optimize the isomerization ofcatharanthine 3 to exocatharanthine 89 various conditionsfor metal catalyzed isomerization of the double bond underhomogeneous conditions were looked at47. In table 2-20 arelisted the different conditions examined. However, it turnedout that isomerization did not occur under any of theseconditions. The attention was therefore turned back to thehydrogenation conditions under which the isomerization wasfirst discovered46. A preliminary investigation establishedthat hydrogen has to be present in order for theisomerization to take place. The fact that hydrogen has tobe present for the isomerization to take place indicatesthat the mechanism of the reaction is a metal hydride 1,2-addition-elimination reaction. Even though isomerizationseemed to be slightly favored over hydrogenation when aprehydrogenated catalyst was used (table 2-21, entry 1 and4) it was for practical reasons decided to continue theoptimization using a non-prehydrogenated catalyst in ahydrogen atmosphere. Change of the concentration ofcatharanthine 3 in the range of 1 - 3 mg/mL does not seem toinfluence the 87b/89 ratio (table 2-21, entry 1 and 3).However, too much catalyst resulted in an increase in15, 20-dihydrocatharanthine 87b (table 2-21, entry 1 and 2).Changing the solvent from methanol to benzene gave asignificant improvement of the 87b/89 ratio (table 2-21,entry 4 and 6).80Table 2-2O. Conditions investigated for the metalcatalyzed isomerization of catharanthine 3 to exocatharanthine 89aEXP METAL CATALYST SOLVENT °C/hr WORKUPb1 Pd Pd(CF3CO2) CH3OCH RT/4 Athen and60/1.5 B2 Na2PdC14 AcOH 50/24 C3 PdC14-AgBF CH3N RT/24 Athen and70/4 B4 PdC12(Ph N) CHC13 61/7 C5 PdC12(Ph N) CH3N 82/3 A6 PdC12(Ph N) benzene 79/11 B7 PdC12(Ph3) CHC13 RT/24 Athen and61/5 Ba,b The table is continued on the following page, see note aand b at the end of the table.81Table 2_0a continued.EXP METAL CATALYST SOLVENT °C/hr WORK_UPb8 Rh RhC13.H20 EtOH 78/4 C9 RhC13.H20 t—BuOH 108/4 C10 RhC13.H20 i—PrOH 82/4 C11 RhC1(Ph3P) benzene RT/20 Bthen and79/3 C12 Ru RuHC1(Ph3P) benzene 79/5 C13 Fe Fe(CO)5 n-octane 127/4 B/D14 Fe(CO)5 n-octane 70/96 Cthen127/12a The reactions were carried out using 10 mg ofexocatharanthine 89.b Method of work-up: A) Reductive (NaBH4).B) Ligand exchange (aqueous KCN).C) Extractive.D) Decomposition by FeC13/EtOH.82Table 2-21. Optimization of the isomerization ofcatharanthine 3 to exocatharanthine 89aexp solvent TPb [3] Pd-C:3 Temp MethodC 3:87b:89(%) (mg/mL) (°C)1 MeOH 1.00 1.2:1 rt A 0:7:32 MeOH 1.67 2.4:1 rt A 0:9:13 MeOH 3.00 1.0:1 rt A 0:7:34 MeOH 0.75 1.0:1 rt B 0:8:25 benzene 2.20 1.1:1 rt B 0:6:46 benzene 1.10 1.1:1 rt B 0:6:47 benzene 1.00 1.0:1 40 B 0:5:58 benzene 1.79 1.0:1 45 B 0:5:59 benzene 0.015 1.00 1.2:1 rt B 2:3:510 benzene 0.030 1.00 1.2:1 rt B 1:5:411 benzene 0.100 1.10 1.1:1 rt B 1:4:512 benzene 1.000 5.10 1.0:1 rt B 10:0:013 benzene 0.030 1.00 1.2:1 60 B 0:3:714 toluene 0.030 0.89 1.2:1 80 B 0:2:8a The reaction were carried out using 10- 50 mg ofcatharanthine 3.b (vol/vol) of thiophene (TP) added to the solvent.C A: prehydrogenated catalyst. B: non-prehydrogenatedcatalyst. For a description of method A and B see chapter3.27 in the experimental section.83Also in benzene it holds true that an increase in theconcentration of catharanthine 3 does not influence theproduct ratio (table 2-21, entry 5 and 6). Increasing thereaction temperature from room temperature to 40°C gavefurther increase in the yield from 40 to 50% ofexocatharanthine 89 (table 2-21, entry 6,7). Poisoning ofthe catalyst with thiophene gave a significant increase inthe isomerization at room temperature but also prevented thereaction from going to completion (table 2-21, entry 6 and 9- 11). Too large amount of catalyst poison totally inhibitedany reaction (table 2-21, entry 12). However increasing thereaction temperature to 60°C in the presence of 0.03%thiophene resulted in an increased amount of exocatharanthine 89 and consumption of all the catharanthine 3(table 2-21, entry 10 and 13). Changing to a less toxicsolvent as well as increasing the temperature another 10°Cit was possible to obtain a further improvement of the yieldof exocatharanthine 89 (table 2-21, 13 and 14). Increasingthe temperature above 80°C resulted in a decrease in theyield of exocatharanthine 89 and the formation of severalside products. Isolation of these side products and asubsequent investigation by mass spectroscopy revealed thatthey were carbomethoxydihydrocleavamines. Comparison oftheir ‘H NMR spectra as well as their fragmentation patternwith the data reported in the literature63b, it was possibleto determine the structure of the isolated carbomethoxydihydrocleavamines as shown in figure 2-10.8486b R1= H, R = Et86c R1= Et, R2 = HFigure 2-10. The structure of the isolated carbomethoxydihydrocleavamines.In large scale preparations of exocatharanthine 89 itturned out to be necessary to lower the reaction temperatureto 70°C to prevent the formation of carbomethoxydihydrocleavamines. In the optimized large scale procedureexocatharanthine 89 could be obtained 83% yield. Theobtained exocatharanthine 89 was shown by proton MNRspectroscopy to be a mixture of the E and the Z isomer in atypical ratio of 5:1. It was not possible to separate thismixture.2.8. INVESTIGATION OF THE FORMATION OF CATHARANTHINE N-OXIDE10 AND EXOCATHARANTHINE N-OXIDE 90The examination of the formation of catharanthine N-oxide 10 and exocatharanthine N-oxide 90 was prompted by areinvestigation of the modified Polonovski reaction usingexocatharanthine 89 instead of catharanthine 3. In theHH86aH85preliminary experiments, following the experience of Dr.Kutney’s research group, about 50% excess ofexocatharanthine 8928, as well as an excess ofchloroperbenzoic acid for the generation of thecorresponding N_oxide2(Th11t), was used in the coupling withvindoline 4. In this way the coupling reaction afforded anaverage of 30% of 19’,20’-anhydrovinblastine 91. The applied-c H31) peracid2) vindoline3) TFRR4) NaBH4excess of exocatharanthine 89 could not be accounted forafter the coupling reaction and the loss of valuablestarting material necessitated a closer look at the wholeprocedure, starting with the N-oxide formation. Treatment ofcatharanthine 3 with an equimolar amount of mchloroperbenzoic acid in dichloromethane at -70°C gave 50%of catharanthine N-oxide 10 and 16% of 7-hydroxycatharanthine N-oxide 92. In order to keep therearrangement of the formed Noxidem),62 to a minimum thetemperature of the dissolved catharanthine N-oxide 10 waskept below 10°C at all times. By using two equivalents of mchloroperbenzoic instead of one the yield of 7—hydroxyH89INDOLINE9186catharanthine N-oxide 92 could be increased to 70%.Similarly, exocatharanthine N-oxide 90 and 7-hydroxyexocatharanthine N-oxide 93 were prepared in 71% and 68%yield respectively.PERPCID•C HI3 A15’2°8 19,2O+U HI13Q10 t L19,2O19,OCatharanthine N-oxide 10 has been characterized byKutney et a1.2° and Potier et al.2Th. Exocatharanthine N-oxide 90 exhibited a typical indole UV spectrum absorbing at221, 274 (shoulder), 280 and 289 nm. The expected molecularion at 352 m/z was observed as well as a peak of significantintensity at 336 corresponding to loss of oxygen from themolecular ion. The ‘H NMR spectrum showed that the 19’,20’-double bond was still intact and the carbons alpha to N-4showed, as would be expected, a significant downfield shiftcompared to exocatharanthine 89. The mass spectrum of 7-HHO87hydroxy exocatharanthine N-oxide 93 showed a small molecularion at 368 mhz with daughter peaks at 352 and 336 m/zcorresponding to loss of one and two oxygens respectively.The UV spectrum of 93 did not contain the characteristics ofan indole nucleus. Compared to 90 the 13C NMR spectrumshowed the absence of a quaternary olefinic carbon and theappearance of a quaternary aliphatic carbon at 85.0 ppm. Thepresence of the 19,20-double bond was demonstrated in theNMR spectrum.A fractional factorial design44 with seven variables attwo levels was carried out in order to screen variousfactors which might influence the reaction of catharanthine3 with m-chloroperbenzoic acid. The variables investigatedin the design are listed in table 2-22 and in table 2-23 isgiven the layout of the design as well as the results.Row E(92) in table 2-23 shows the effect of the differentfactors on the formation of 7-hydroxy catharanthine N-oxide92. It is seen that the excess of peracid and the mode ofaddition have the largest effect on the formation of 92whereas the effect of the remaining factors areinsignificant. In fact, addition of another equivalent ofperacid to the experiment leads to conversion of all theformed catharanthine N-oxide 10 into 7-hydroxy catharanthineN-oxide 92. Row E(3) in table 2-23 shows the effect of thedifferent factors on the amount of remaining catharanthine3. With an absolute uncertainty in the order of 1% on theyields none of the factors can be said to have a significant88influence on the amount of remaining catharanthine. Since anexcess of peracid was used in all the experiments in thefractional factorial design all the catharanthine should, intheory, have been used up. A prolonged reaction time did nothave any influence on either the ratio or the amount of thethree compounds, showing that the reaction has indeed goneto completion. That catharanthine remains despite the excessof peracid must be due to protonation of the catharanthine 3Table 2-22. The factors and their levels used in thefractional factorial design in table 2-23.Factorsolv: solventconc: concentrationtemp: temperaturemode of addn:peracid purity:peracid excess:dichioromethane100 mg/mL-10°Cperacida added as asolid to the solution of 3.reaction vessel notcovered with Alfoil.acetonitrile20 mg/mL-30°C3 added as a solidto the peracidasolution.reaction vessel iscovered with Alfoil.oDoo.J 05%level 1 level 2light:n0,u 0DrOLi 0a zn-chloroperbenzoic acid.89Table 2_238. Investigation of the formation of 7-hydroxycatharanthine N-oxide 92.expb solv conc temp mode of light peracid peracid %92C %3Caddn purity excess1 1 1 1 2 2 2 1 24 52 2 1 1 1 1 2 2 10 13 1 2 1 1 2 1 2 10 14 2 2 1 2 1 1 1 22 25 1 1 2 2 1 1 2 16 56 2 1 2 1 2 1 1 20 07 1 2 2 1 1 2 1 20 18 2 2 2 2 2 2 2 12 1¼ El 18 18 17 15 17 17 22¼ E2 16 16 17 19 17 17 12E(92) 2 2 0 4 0 0 10¼E1 3 3 2 1 2 2 2¼E2 1 1 2 3 2 2 2E(3) 2 2 0 2 0 0 0a The levels of th factors are defined in table 2-22.b The reactions were performed as described in the generalprocedure in chapter 3.28 on a 50 mg scale.C The yield is determined by HPLC using 92 and 3 as externalstandards. The absolute uncertainty on the yield is ± 1%.90and/or that the peracid reacts faster with catharanthine N-oxide 10 than it does with catharanthine 3. The latter canbe ruled out since it is possible to minimize the amount of7-hydroxy catharanthine N-oxide 92 by lowering the excess ofperacid used. In order to determine if protonation ofcatharanthine slows down the formation of N-oxide andtherefore indirectly promotes the formation of 7-hydroxycatharanthine N-oxide 92 2.3 equivalence of acetic acid wereadded to the catharanthine 3 solution 7 mm prior to theaddition of the peracid, but otherwise using the sameconditions as in experiment 3 in table 2-23. The result isshown in table 2-24. This experiment confirmed that protoTable 2-24. The influence of acid on the formation of 7-hydroxy catharanthine N-oxide 92.exp solvent conc temp mode of light peracid peracid % 92 % 3addn purity excess9 1 2 1 1 2 1 2 27 13nation of catharanthine 3 slows down the N-oxide formationand thereby indirectly promotes the formation of 7-hydroxycatharanthine N-oxide 92. In large scale preparation (200 -500 mg) it was found that equimolar amounts of catharanthine913 and peracid overall gave the best yield of the desiredcatharanthine N-oxide 10, namely 85 - 90%.For exocatharanthine 89 the use of equimolar amounts ofperacid was also found to be optimal, giving a yield of 90 -95% of exocatharanthine N-oxide 90 with equal amounts ofunreacted exocatharanthine 89 and 7-hydroxy exocatharanthineN-oxide 93 accounting for the remaining 5 - 10% of startingmaterial.Acetonitrile and acetone were also examined as possiblesolvents. The yield of N-oxide in these solvents was foundto the same as in dichloromethane.2.9. REINVESTIGATION OF THE MODIFIED POLONOVSKI REACTIONCoupling of exocatharapthine 89 using the standardconditions2Ob gave a yield of 30 - 35% of 19’,20’-anhydro-vinbiastine 91 (scheme 2-13). Compound 91 gave the expectedmolecular ion of 792 m/z and the low resolution massspectrum contained all the characteristic fragments of theexpected dimeric structure64. In the NMR spectrum of 91the protons at C-lB’ appear as a doublet at 1.68 ppm (J =6.5 Hz) and H-l9’ as quartet at 5.49 ppm. The CD spectrum infigure 2-ha indicates that 91 has the correctstereochemistry at C-16’6. Another compound was isolatedfrom the reaction mixture with the desired molecular ion at792 m/z and the expected fragmentation pattern. The ‘H NMRspectrum showed that this compound also contained a C-19’,C-20’-double bond. The CD spectrum in figure 2-lib indicates92that this compound must be the C-16’ epimer 94. Analogs ofCO3HiJL1•C H3901) VINDOLINE/ 2)/1/ 3) NaBH4/_CH3 H3+H / ‘VINOOLINECO2H391Scheme 2-13. Coupling of exocatharanthine 89 with vindoline4.91 have been synthesized by Miller et al., by dehydration ofvinblastine with concentrated sulfuric acid to give thedesacetylated and dehydrated products 95 - 9765 (figure 2-12). Comparison of the chemical shifts found for H-19’ inthe dimers 96 and 97 with the chemical shift of 5.49 ppmfound for H-19’ in 19’,20’-anhydrovinblastine 91 stronglyindicate that no isomerization of the 19’,20’-double bondCD2H3 CO2H3H INDOLINECH393has taken place during the coupling of exocatharanthine 89with vindoline 4.a: CD of 91b: CD of 94—3350 nmFigure 2-11. a) The CD spectrum of 19’,20’-anhydro-vinbiastine 91. b) The CD spectrum of epi 19’,20’-anhydro-vinbiastine 94 (solvent CH3N).Le1—1—2-4’200 250 300LE10200 250 300 350 nm94H2S04VINBLflSTINER: 17_(desacety1)V1fld0l1te96—5.46 ppmHppm-H—--5.28 ppmFigure 2-12. Dehydration of vinblastine 1.Using fractional factorial design44 as experimentalstrategy various factors were investigated to study theirinfluence in the modified Polonovski reaction with respectH95+H+CH39795to the yield of 19’,20’-anhydrovinblastine gi. Investigationof the effect of the reaction temperature, the purity of them-chloroperbenzoic acid and the addition rate of TFAA on theyield of 91 (table 2-25) disclosed that only the reactiontemperature had a significant influence on the yield of 91.Comparison of the two temperature levels shows that areaction temperature of -60°C gives a better yield than at -40°C. Experiment 5 in table 2-26 confirmed this conclusion.Lowering the temperature below -73 to -78°C did not increasethe yield but resulted in a substantial increase in thereaction time. Comparison of experiments 1 and 3 withexperiments 2 and 4 in table 2-25 indicates that, at -40°C,either the addition rate of TFAA or the peracid puritystrongly affects the yield whereas this is not the case at-60°C. Fast addition of TFAA results in a temperatureincrease of 10 to 15°C in less than 15 seconds but thetemperature drops back down to the actual reactiontemperature within two minutes. Further experimentsestablished that this brief increase in temperature resultsin a significant drop in the yield when the reaction iscarried out at -40°C.The fractional factorial design in table 2-27 showsthat with respect to the amount of vindoline 4, the natureof the solvent and the amount of TFAA both the nature of thesolvent and the amount of TFAA have a significant effect onthe yield of 19’,20’-anhydrovinblastine 91. Changing thesolvent to acetone, but otherwise using the same conditions96Table 2-25. Investigation of the effect of the reactiontemperature, the purity of the zn-chloroperbenzoic acid andthe addition rate of TFAA on the yield of 19’,20’-anhydrovinbiastine 91a•Factor level 1 level 2Reaction temp -60°C -40°CPeracid purity 96% 69%Addn rate of TFAA < 1 sec 20 mmexp Reaction PeracidC addn rate HPLCb Isolatedtemp purity of TFAA * 91 % 94 % 911 -60°C 96% 20 mm 34 2-2-40°C 96% < 1 sec 17 3-3 -60°C 69% < 1 sec 33 2-4 -40°C 69% 20 mm 27 1-½E(9J-) 33.5 25.5 25.0½E(91) 22.0 30.0 30.5E(91) 11.5 4.5 5.5a 200 mg exocatharanthine 89 in CH21 used. The reactionswere performed as described in chapter 3.29 using steps laand 7a. The reactions were not followed by HPLC.b HPLC conditions as described for exocatharanthine 89 inchapter 3.28. The yield of 94 is estimated from the heightratio of 91 and 94 assuming that 94 has the same responsefactor as 91, since no actual calibration curve has beenestablished for 94.c zn—chloroperbenzoic acid.97Table 2_6a The Polonovski reaction carried out at lowtemperature.exp Reaction PeracidC Addn rate HPLCb Isolatedtemp purity of TFAA % 91 % 94 % 915 -73°C 96% < 1 sec 41 2-a See note a under table 2-25.b See note b under table 2-25.c see note c under table 2-25.as in experiment 7 (table 2-27), gave an increase in theyield (isolated) of 91 from 44 to 51%. After an extensiveinvestigation of acetonitrile, acetone and acetonitrile/dichloromethane (1:1) as solvents for the reaction it becameclear that the increased yield obtained by switching fromdichioromethane to acetonitrile/di-chioromethane or acetonewas not due to the solvent as such, but to an additionalcoupling which took place during the evaporation step. Theevaporation step is the point in the procedure where thereaction is “finished” and the cooling bath is removed whilethe solvent is evaporated off. The temperature of thereaction is raised to about 20°C during the evaporation.After the solvent has been evaporated off methanol is added,the reaction is cooled down to -20°C and reduced with sodiumborohydride. A closer examination of the evaporation steprevealed that in acetone, no coupling took place at -60°Cbut first when the reaction was heated above _2000.98Table 2_278. Investigation of the effect of the amount ofvindoline 4, the nature of the solvent and the amount ofTFAA on the yield of 19’,20’-anhydrovinblastine 918.Factor level 1 level 2Vindoline 4 1.5 eq 1.0 eqSolvent CH21/C3N CH21. (1:1)TFAA 4.8 eq 1.2 eqexp Vindoline Solvent TFAA HPLCb Isolated(4) %91 %94 %916 1.5 eq CH21/C3N 1.2 eq 19 3 -7 1.0 eq CH1/CN 4.8 eq 48 7 448 1.5 eq CH21 4.8 eq 30 5 -9 1.0 eq CH21 1.2 eq 8 1 -½1 1 24.5 33.5 39.02 28.0 19.0 13.5E(9l) 3.5 14.5 25.5a See note a under table 2-25.b See note b under table 2-25.99Even in dichioromethane, where an increased yield of 91 isobtained by lowering the temperature from -60 to —73°C,additional product is formed upon heating the reaction above-30°C. This apparent contradiction of the dependence of theyield on the temperature can only be explained by theexistence of two different intermediates of exocatharanthine89, each capable of coupling with vindoline 4 to give thedesired dimer (scheme 2-14) . Addition of TFAA toexocatharanthine N-oxide 90 must lead to acylation of 90 togive the acylated intermediate 98 which then reacts withvindoline 4 at -60°C to give 19’,20’-anhydrovinblastine 91.To account for the observed formation of dimer on heating ofthe reaction mixture a second intermediate must be present,structurally different from 98, which is capable of reactingwith vindoline 4 but only at a much higher temperature thanthe temperature needed for the reaction of 98 with vindoline4. That the yield of 91 actually increases on lowering thereaction temperature from -60 to -73°C points towards acompetitive reaction between vindoline 4 and a reactiontransforming 98 into a second intermediate 99 also capableof coupling with vindoline 4 but only at a highertemperature.In order to substantiate the idea of twoexocatharanthine intermediates capable of coupling withvindoline 4 the coupling reaction was followed by HPLC usingthe same conditions as those described for exocatharanthine89 in chapter 3.28. In figure 2-13 are shown the I-IPLCTFflRCO3HLC10100Scheme 2-14.Polonovski reaction.02CH389CF300-CO2H390OCF3H98SECOND INTERFIEDIPTE99H3VINDOLINE_73o>—30°CCF3OOH3100PcH3NaBH491The reaction pathways in the modified101A .3B161: 1.2 mm. 6: 11.2 mm.2: 2.4 mm. 7: 9.7 mm.3: 3.1 mm. 8 :11.4 mm.4: 3.9 mm.5: 9.3 mm.6: 11.2 mm1: m-Chloroperbenzoic acid. 2: 7-Hydroxy-exocatharanthine N-oxide 93. 3: Exocatharanthine N-oxide 90. 4: Vindoline 4.5: Second intermediate 99. 6: Iminium dimer 100. 7: Epi19’ ,20’ -anhydrovinbiastine 94. 8: 19’, 20’ -Anhydrovinbiastine91.A) Exocatharanthine N-oxide 90 and vindoline. 4 B) Acylatedexocatharanthine N-oxide. C) The reaction mixture at -73°Cafter the reaction has stopped. D) The reaction mixtureafter concentration and heating. E) The reaction mixtureafter reduction with NaBH4.Figure 2-13. Monitoring of the Polonovski reaction by HPLC(same conditions as for exocatharanthine 89 in chapter3.28).156E 807102chromatograms at different stages of the reaction. It shouldbe noted that the peaks attributed to the variousintermediates might not represent the actual intermediateitself but rather an artifact created from that particularintermediate by reaction with methanol.In experiment 10 (table 2-28) TFAA was added toexocatharanthine N—oxide 90 prior to the addition ofvindoline 4 (step 1). After completion of the reaction ofTFAA with 90 the excess of TFAA was evaporated of f prior tothe addition of vindoline 4 (step 2). This removal of excessTFAA is absolutely necessary in order to prevent theacylation of vindoline 4 and any dimer(s) present when thereaction is heated above —30°C. It is clear from the HPLCchromatogram that no coupling takes place unless thereaction is heated above —30°C. If, however, the vindoline 4is added prior to the addition of TFAA, as was done inexperiment 11 (table 2-29), the reaction is 78% “complete”in only 16 mm at -65°C. These two experiments clearlydemonstrate that exocatharanthine N-oxide 90 in the presenceof TFAA only is transformed into an intermediate (99)capable of coupling with vindoline to the desired dimer, butonly at temperatures above -30°C. This observation by HPLCwas subsequently confirmed by NMR (figure 2-15). As tothe nature of this second intermediate 99, it must either bea fragmentation product of 98 or a reaction product of 98with the trifluoroacetate ion released upon reaction oftrifluoroacetic anhydride with the N-oxide 90.103Table 2_8a Addition of TFAA to exocatharanthine N-oxide 90prior to the addition of vindoline 4.EXP CONDITIONSb. HPLCC10 1) 90 at _6000, add TFAA, Remaining N-oxide 90:stirring for 2 hr. 20% after 16 mm7% after 1 hr 3 mm3% after 1 hr 56 mm2) -60- -30°C over 30 mmwhile evac. on vac. line,stop evacuation and cool to3) -40°C, yellow foam dissolved HPLC profile looks thein 2 ml, of CH21, cool to same after as beforethe evaporation4) -55°C, add 4, stirring for1 hr, then heating to5) -30°C, stirring for 2.5 hrs, time temp %4 %99 %100dthen heating to 0 -52 100 100 040 -33 98 96 057 -30 94 94 0a The table is continued on the following page.b,c,d See the end of the table on the next page.104Table 2-28 continued.9% 9111% 94200 mg of exocatharanthine 89 in CH21, 1.00 equivalentof m-chloroperbenzoic acid and 1.5 equivalent of viridoline 4 used. The experiment was performed as described inchapter 3.29 (using steps la and 7b) except for thechanges in addition, temperatures and reaction timesdescribed in table 2-28.c HPLC conditions as described for exocatharanthine 89 inchapter 3.28.d The time is in minutes and time 0 mm is the point whenvindoline 4 was added. The amount of 4 and secondintermediate 99 is the height of the peak at time x mmin percent of the peak height at time 0 mm. The amount ofintermediate 100 is the peak height at time x mm inpercent of the peak height at time 335 mm.EXP CONDITIONSb. HPLCC6) 0°C, stirring for 4 hrs,then workuptime temp %4 %99 %100d94 —36 97 86 0138 —42 88 77 0165 -30 92 75 0193 0 96 38 46335 0 55 8100Yield:b105Table 2_9a• Addition of vinc7oline 4 prior to the additionof TFAA. The importance of concentrating the reactionmixture after the reaction has gone to “completion” at—65°C.EXP CONDITIONSb HpLCc,d11 1) 90 and 4 at -65°C, addition time %90 %4 %99 %100of TFAA, stirring for 3 hr., 0 100 100 0 016 15 65 91 7839 5 55 93 9096 0 55 103 103159 0 51 100 100end 0 3 0 1852) —65 -, -40°C over 20 mmwhile evac. on vac. line togive a red foam. Vacuumreleased and the foam dissolved in 1 mL CH21 at3) -45°C over 15 mm4) -45°C evac. on vac. line for25 mm. Vacuum released andheating froma The table is continued on the following page.b,c,d See the end of the table on the next page.106Table 2-29 continuedCONDITIONSb5) -45°C to RT over 1 hr.6) Yield: HPLC: 72% 9112% 94Isolated: 70% 91b 100 mg of exocatharanthine 89 in CH21, 1.00 equivalentof m-chloroperbenzoic acid and 1.0 equivalent of vindoline 4 used. The experiment was performed as described inchapter 3.29 (using steps la and 7b) except for thechanges in temperatures and reaction times described intable 2-22.c See c under table 2-28.d The time is in minutes and time 0 mm is the point whenTFAA was added. The amount of exocatharanthine N-oxide90, vindoline 4 and the second intermediate 99 is theheight of the peak at time x mm in percent of the peakheight at time 0 mm. The amount of intermediate 100 isthe peak height at time x mm in percent of the peakheight at time 159 mm. The amounts at time “end” refersto the composition of the reaction after evaporation andheating (entry 5).EXP HPLCC107Table 2_30a• The effect of concentrating (versus not concentrating) the reaction mixture after the reaction has goneto “completion” at -65°C on the yield and product distribution.EXP CONDITIONSb I-IPLC°12 1) 90 at -60°C, addition of Remaining 90:TFAA, stirring for 1.5 hr., <1% after 22 mmheating to2) -45°C, evac. on vac. linefor 20 mm, then heating to-30°C for 5 mm whereby ayellow foam is formed,cooling to3) -50°C, the yellow foam is HPLC profile looks thedissolved in 2 mL CH21 the same as before evap12A 1) 99 at -65°C, addition of 4, time temp %4 %99 %100dstirring for 1 hr.2) -65°C, add 4.8 eq. TFAA, 0 -65 100 100 0stirring for 1 hr. 45 -55 92 89 03) -50°C for 30 mm 238 -50 85 85 04) -50°C + another 4.8 eq. 545 0 0e < 100TFAA, stirring for 4.5 hr.a The table is continued on the following page.b,c,d,e See the end of the table on the next page.108Table 2-30 continued.EXP CONDITIONSb HPLCC12A5) -30°C, evac. on vac. linefor 20 mm, thick oil,then heating to6) 0°C for 2.5 hr7) Yield 55% of 9112% of 94123 1) 99 at -65°C, addition of 4, time temp %4 %99stirring for 2 hr., then2) -50°C, stirring for 5 hr., 0 -60 100 100 0heating to 90 -50 99 92 03) 0°C and stirring for 2 hr. 203 -50 97 95 0480 0 81 <17 1004) Yield 11% of 917% of 94b 100 mg of exocatharanthine 89 in CH21 and 1.00 equivalent m-chloroperbenzoic acid used. After step 3 inexp. 12 the reaction mixture was divided equally into exp.12A and 12B. 1.5 equivalent vindoline 4 (used in exp.12A and 12B. The experiments were performed as describedin chapter 3.29 (using steps la and 7c) except for thechanges in addition, temperatures and reaction timesdescribed in table 2-30.c See c under table 2-28.d See d under table 2-28.c Substantial amounts of acylated vindoline 4.109From the HPLC chromatogram of experiment 11 in table 2-29 it can be seen that the reaction stops after 1 hr 30 mmat —65°C because all the N-oxide 90 has been converted toeither the exocatharanthine intermediate 99 or to thedimeric intermediate 100. Prolonged reaction time at thistemperature does not give any increase in eitherintermediates. After evaporation of the solvent at -45°C andsubsequent heating the peak height of the dimericintermediate 100 has increased 85% compared to its heightprior to evaporation and heating (table 2-29) giving anisolated yield of 70% of 91. It is clear that a considerableamount of the second intermediate 99 is formed at -65°C incompetition with 100. The importance of the concentrationstep (step 2, table 2-29) on the yield, as well as on theratio of dimers with the natural stereochemistry to the onewith unnatural stereochemistry at C-16’, can be seen fromexperiment 12 in table 2-30. Concentration prior to theheating not only gives a higher yield of 91 (table 2-30,exp. 12A compared to exp 12B) but also suppresses theformation of the epimer 94.In conclusion, the traditional way of performing themodified Polonovski reaction consists of three steps; 1)formation of N-oxide, 2) coupling with vindoline at -50°C,3) reduction with sodium borohydride and work-up. As aresult of this investigation into the factors affecting theyield of the modified Polonovski reaction a side reactionleading to a second intermediate 99 was discovered. This110intermediate 99 has also been shown to be capable ofcoupling with vindoline 4, under appropriate conditions.Taking advantage of this discovery an increased yield of thedesired dimeric alkaloid is obtained compared to thetraditional procedure. This improved procedure of thePolonovski reaction consists of five steps; 1) formation ofN-oxide, 2) coupling with vindoline at - 73 to - 78°C, 3)evaporation of the solvent below -30°C, 4) heating of thealmost dry residue to 0°C, 5) reduction with sodiumborohydride and work-up. By monitoring the reaction by HPLCit has been found that incorporation of these two new stepsafter the point where the reaction was “complete” at -70°Cresulted in an increase of the peak height relating to theintermediate 100 of up to 75% in the case of catharanthine3, 85% in the case of exocatharanthine 89 and 45% in thecase of dihydrocatharan-thine 87b. It is, however, notadvisable to first transform all the N-oxide 90 into thesecond intermediate 99 since 99 reacts with vindoline 4 toform both 16’S and 16’R dimers, whereas the HPLC profile ofthe coupling at -73°C shows that only the 16’S dimer isformed. Concentration plays an important role at bothtemperature stages of the coupling reaction. When themodified Polonovski reaction was performed in thetraditional way it has been found that dilution lead to adecrease of the yield of dimer2Ob, . In the improvedprocedure the concentration has an extreme influence on thecoupling reaction in terms of the yield and the product111ratio between the 16’S and 16’R isomers. If the reactionmixture is not concentrated as much as possible the secondintermediate 99 not only shows a higher degree ofdecomposition and thereby lower yield upon heating but alsoa substantial increase in formation of the undesired 16’Rdimer.2.10. IDENTIFICATION AND CHARACTERIZATION OF THE SECONDINTERMEDIATE 99 IN THE MODIFIED POLONOVSKI REACTIONAs demonstrated in the previous chapter a secondexoca-tharan-thirie intermediate 99 is formed in the modifiedPolonovski reaction which is also capable of coupling withvindoline 4 to give the desired dimer. The exocatharanthineN-oxide 90 was transformed exclusively into thisintermediate in the absence of vindoline 4. In figure 2-14are listed all the possible structures for this secondintermediate 99. In table 2-31 are summarized the differenttypes of carbon-13 signals that are to be expected for eachof the five possible structures shown in figure 2-14together with the actual types of carbons found for thesecond intermediate 99 excluding the signals from thetrifluoroacetate group. It can be seen that the onlypossible structures for the second intermediate are 101, 103and 99. These possibilities all contain one quaternaryaliphatic carbon and one tertiary aliphatic carbon. From thel3 APT spectrum listed in table 2-34 the chemical shift forthe quaternary aliphatic carbon was found to be at11296 5OCOCF3121 5CO2H31922 2398 10185.89 ppm. This excludes structure 101 since the chemicalshift for a positively charged carbon would be expected tobe above 200 ppm ( e.g. (CH3) ; 328 ppm and (Ph)3C ;211 ppm).1o11‘3-CH3 H3399Figure 2-14. Possible structures of the second intermediate99.C 02C H310 103CF3O 02CH3113Table 2-31. The type of carbon signals expected foreach of the five possible structures in figure 2-14,together with the actual types of carbons found inthe APT spectrum for the second intermediate 99(excluding the signals from the trifluoroacetate group).Structure =0< =CH- >C< >CH- >CH2 —CH398 6 5 1 2 5 2101 6 6 1 1 5 2102 7 6 0 1 5 2103 6 6 1 1 5 299 6 6 1 1 5 2second intermediate99 6 6 1 1 5 2This leaves 103 and 99 as possible structures for thesecond intermediate. In figure 2-15 is shown the NMRspectrum of the second intermediate 99 and the chemicalshifts are summarized in table 2-32. The results of APTspectrum are summarized up in table 2-34. In table 2-35 areshown the results of two HETCOR experiments, one coveringthe aliphatic region the other the aromatic region.I-I.j I-I. (.Q01•1CD(Drt QJ. z (n CD C) r1 0 H CD U) CD C) 0 ci H rt CD ‘-1 CD ci F.-’.0) CD ‘.0 ‘.OI-’h754321115Table 2-32.spectrum ofChemical shifts observed in the NMRthe second intermediate 99aShift (ppm) # H Correlation2.12 (d) 3 H-182.60 (d) 1 H—152.67—2.82 (m) 2 H-15, H-172.84-3.01 (m) 2 H-14, H-173.42 (t) 1 H-63.67 (s) 3 H-233.71-3.82 Cm) 3 H-3, H-64.12 (t) 1 H-54.59 Cd) 1 H—57.15 (t) 1 H_lOb7.25 Ct) 1 H_llb7.39 (q) 1 H-197.46 (d) 1 H--127.73 (d) 1 H-99.10 (s) 1 N—2111.75(s) 1 N—Ha The assignment might have to be reversedb Solvent: Acetone-d6.116Table 2-33. Chemical shift correlations obtained in theC0SY spectrum of the second intermediate 99a•Signal (ppm) Correlation (ppm)2.60 (d) 2.77, 2.92, 3.77.2.77 (d) 2.60, 2.92, 3.77.2.87 (m) 3.77.2.92 (t) 2.60, 2.77.3.42 (t) 3.75, 4.12, 4.59.3.75 (d) 3.42, 4.12, 4.59.3.77 (s) 2.60, 2.77, 2.87.4.12 (t) 3.42, 3.75, 4.59.4.59 (d) 3.42, 3.75, 4.12.a Solvent: Acetone-d6.117Table 2-34. Chemical shifts observed in the ‘3C APT NMRspectrum of the second intermediate 99a•6(ppm) +/_b Correlation 6(ppm) 1_b Correlation16.71 C-18 110.83 ÷ C—726.56 + C-6 112.98 C—1227.46 C-14 115 + quartet; CF3O2-30.78 + C-15 119.52 - C-933.03 + C-17 121.19 - C_lob54.31 + C-3 124.55 - C_llb54.31 C-23 128.14 + C_2C60.87 + C-5 128.34 + C-885.89 + C-16 131.14 + C_2OD136.73 + C-13157 + quartet; CF3O2-162.03- C—19167.94 + C—22170.56 - C-21a Solvent: Acetone-d6.b The +/- sign indicate whether the sina1 is positive (>C<,-CR2) negative (>CH-, CR3-) in the ‘3C APT spectrum.118Table 2-35. Chemical shift correlations obtained in theHETCOR spectrum of the second intermediate 99a.l3cb ‘H correlation l3C0 correlation(ppm) (ppm)16.71 (C—18) 2.12 112.98 (C-12) 7.4626.56 (C-6) 3.42 and 3.75 119.52 (C-9) 7.7327.46 (C-14) 2.84 - 3.01 121.19 (C_1O)e 7.1530.78 (C-15) 2.60 and 2.77 124.55 (C_ll)e 7.2533.03 (C-17) 2.77 and 2.92 162.03 (C-19) 7.3954.31 (C-3,C-23) 3.77 and/or 3.67 170.56 (C-21) 9.1060.87 (C-5) 4.12 and 4.59a Solvent: Acetone-d5.b 1.8 ppm to 4.8 ppm correlated with 10 ppm to 100 ppmc 7.0 ppm to 11.4 ppm correlated with 105 ppm to 175 ppme The correlation might have to be reversedPosition 18, 19, 22, 23 and C-14From the 13C APT spectrum in table 2-34, carbons 14,18, 22 and 23 were easily identified as the signals at27.46, 16.71, 167.94 and 54.31 ppm respectively. The protonsattached to carbons 18, 19 and 23 were readily identified inthe 1H NMR spectrum listed in table 2-32 based on theirchemical shifts and multiplicity. The HETCOR spectrum listedin table 2-35 established the position of H-14 to be in themultiplet at 2.84-3.01 ppm and the position of C-19 to be at119162.03 ppm. The chemical shift assignment for position 18was also confirmed in the HETCOR spectrum.Position 5 and 6The positive signal at 60.87 ppm in the 13C APTspectrum must be one of the methylene groups alpha to thepositively charged nitrogen, N-4. The HETCOR spectrum intable 2-35 correlates the carbon signal at 60.87 ppm to theprotons at 4.12 ppm and 4.59 ppm. The COSY spectrum listedin table 2-33 establishes that these two protons couple toeach other, as expected, and to the protons at 3.42 ppm and3.75 ppm which in turn are correlated to the methylenecarbon at 26.56 ppm in the HETCOR spectrum. The onlypossible assignment of this -CH2 fragment is that theprotons at 4.12 ppm and 4.59 ppm and the carbon at 60.87 ppmbelong to 0-5, and protons 3.42 ppm and 3.75 ppm and thecarbon at 26.56 ppm to 0-6.Positions 3 15 17 and H-14The next highest methylene group in the 13C APTspectrum appears at 54.31 ppm and was assigned as 0-3. Thecorrelation of 0-3 to its protons in the HETCOR spectrum isunfortunately ambiguous in that 0-3 and C-23 have the samechemical shift and only one peak is seen in the HETCORspectrum which correlates to the area in the proton spectrumwhere the methoxy signal, as well as the H-3 protons,resonate. The remaining two methylene groups in the carbon-12013 spectrum resonate at 30.78 ppm and 33.03 ppm. The SINEPTTable 2-36. Observed enhancements in the SINEPT spectrum ofthe second intermediate 99. Irradiation at 7.39.ppm andObservation in the chemical shift region of 0 - 90 ppma.Irradiation (ppm) Observed enhancement (ppm)7.39 (H—19) 16.71 (100%), 30.78 (68%)a Solvent: CDC13.experiment in table 2-36 is optimized for detection of threebond couplings with a coupling constant of 7 Hz between Cand H by irradiation of the appropriate hydrogen.Irradiation of H-19 at 7.39 ppm resulted in enhancement ofC-18 at 16.71 ppm. In addition a 68% relative enhancementwas observed at 30.78 ppm. This signal could therefore beassigned as C-15 and the chemical shift at 33.03 ppm to C—17. The chemical shifts of the H-15 protons were found to beat 2.60 and 2.77 ppm in the HETCOR spectrum. As expected,the COSY spectrum showed the H-15 protons to be doubletscoupled to each other. The HETCOR spectrum established thatthe H-17 protons resonate at 2.77 ppm and 2.92 ppm. The COSYshowed the 1-1-17 protons, as expected, to be coupled to eachother and that the proton at 2.77 ppm is a doublet whereasthe proton at 2.92 ppm is a triplet. Unfortunately it is not121possible to see the coupling between the H-15 and the H-17protons with the H-14 since the COSY is too crowded in thisregion. By a process of elimination the remaining twoaliphatic protons are located in the broad singlet at3.77 ppm and must the H-3 protons. In the COSY the H-3protons at 3.77 ppm showed coupling to the H-15 proton at2.60 ppm and the H-15 (or H-17) proton at 2.77 ppm as wellas to a triplet centered at 2.87 ppm. The triplet at 2.87ppm was assigned to H-14.Position 21 and the Indole NHOf the two protons at 9.10 ppm and 11.75 ppm the HETCORspectrum correlated the proton at 9.10 ppm to the carbon at170.56 ppm, accounting for H-21 and C-21 respectively. Theindole NH proton must then be the signal at 11.75 ppm.Position 9, 10, 11, and 12The protons at 7.46 ppm and 7.73 ppm correlate in theHETCOR to the carbons at 112.98 ppm and 119.52 ppmrespectively, and corresponds either to position 9 orposition 12. Irradiation of the N-H proton at 11.75 ppmresulted in an NOE enhancement of the proton at 7.46 ppmwhich was therefore assigned to be H-12. The protons at7.15 ppm and 7.25 ppm correlate in the HETCOR to the carbonsat 121.19 ppm and 124.55 ppm respectively and correspondseither to position 10 or position 11.122Position 7In the SINEPT experiment listed in table 2-37 it isseen that irradiation of each of the H-5 protons at 4.59 ppmand 4.12 ppm gave, in both cases, maximum enhancement of thequaternary carbon at 110.83 ppm. This chemical shift valueis therefore assigned to be C-7. It can be unambiguouslyconcluded that the structure of the second intermediatecorresponds to structure 99 since the chemical shift of thequaternary aliphatic carbon is at 85.89 ppm.Table 2-37. Observed enhancements in the SINEPT experimentof the second intermediate 99. Irrdiation at 4.57 ppm and4.09 ppm and observation in the chemical shift regionof 80 - 175 ppma.Irradiation Observed enhancement(ppm) (ppm)4.59 (H—5) 110.83 (100%), 170.56 (11%)4.12 (H-5) 110.83 (100%), 128.14 (60%), 128.34 (25%)a Solvent: Acetone—d6.Position 8 and 13 and the tentative assignment of position2, 10, 11, and 20In the SINEPT experiment in table 2-38 are listed theenhancements observed in the aromatic/clef inic carbon regionby irradiation of the H-9 and H-12 protons. Irradiation at1237.73 ppm (H-9) results in maximum enhancement at 136.73 ppmand this chemical shift value is therefore assigned as C-13.Irradiation at 7.46 ppm (H-12) gave a maximum enhancement at128.34 ppm and is assigned as C-8. The irradiation atTable 2-38. Observed enhancements in the SINEPT experimentof the second intermediate 99. Irradiation at 7.73 ppm (H-9)and 7.46 ppm (H-12) and observation in the chemicalshift region of 83 - 175 ppma.Irradiation Observed enhancement(ppm) (ppm)7.73 (H—9) 136.73 (100%), 128.34 (44%), 124.55 (31%),110.83 (26%), 112.98 (18%), 121.19 (10%),119.52 (7%).7.46 (I-{—12) 128.34 (100%), 162•03b (65%), 170•56b110.83 (26%), 119.52 (18%), 112.98 (15%),a Solvent: Acetone—d6.b The enhancements observed at 170.56 ppm (C-21) and 162.03ppm (C-19) are due to the fact that the chemical shift ofH-19 at 7.39 ppm is too close to the chemical shift ofthe proton irradiated at 7.46 ppm, so that some irradiation of H-19 also occurred.7.73 ppm also gave a 10% relative enhancement for the carbonat 121.19 ppm and 31% relative enhancement for the carbon at124.55 ppm.124Again assuming a larger enhancement for the carbon threebonds away, the signal at 121.19 ppm was assigned as C-10and the signal at 124.55 ppm to C-il. The olefinic/aromaticquaternary carbons at 128.14 ppm and 131.14 ppm did not showany enhancement and are therefore assigned as position 2 or20.The numbering of the carbon atoms in the secondintermediate 99 is shown below and the assignments of thechemical shifts are summarized in table 2-39.CF3OO CO2H322 23125Table 2-39. Assignment of the ‘H and chemical shifts tothe second intermediate 99aPosition H ö(ppm) C 6(ppm)123 54.315 60.876 26.567 110.838 128.349 119.5210 121.9c11 12455c12 112.9813 136.7314 27.4615 30.7816 85.8917 33.0318 16.7119 162.032021 170.5622 167.9423 54.3111.753.774. 12/4. 593.42/3.757.737.157.257 . 462. 872. 60/2. 772.77/2.922. 127.399.103.67a Solvent: Acetone-d6.b,c c5 values might have to be reversed.1262.11. PRELIMINARY INVESTIGATION OF THE REACTIVITY OF THE19’,20’-DOUBLE BOND IN 19’,2O’-ANHYDROVINBLASTINE 91.The reactivity of the C—15’,C-20’-double bond in15’,20’-anhydrovinblastine 5 has been studied in somedetail, with the objective of synthesizing vinbiastine 1 orleurosidine 113, the C-20’ epimer of vinblastine 1. The0504Scheme 2-15. Transformation of the 15’,20’-double bond in15’ ,20’ -anhydrovinbiastine 5(CH3)C00HH NOOLINE88eH)I-I,2CH35H‘INOOLINE2CH31061) 8H32) H20/H0IHINDOLINE107127results of these investigations are outlined in scheme2_1522,23,67. Attempts to convert leurosine 88e or the diol106 to either vinblastine 1 or leurosidine 113 wereunsuccessful23.The reactivity of the 19’,20’-double bond in19’,20’-anhydrovinblastine 91 was found to be quitedifferent from that of the 15’,20’-double bond in 15’,20’-anhydrovinblastine 5. The 19’,20’-double bond in 91 was nottransformed into the corresponding epoxide when exposed toair. This is in sharp contrast to the reactivity of the15’,20’-double bond in 5, which, both as a solid and insolution, forms significant amounts of leurosine 88e whenexposed to air. Furthermore the 19’,20’-double bond in19’,20’-anhydrovinblastine 91 did not react with tert-butylhydroperoxide under the conditions described for 567•Treatment of 19’,20’-anhydrovinblastine 91 with mercuricacetate followed by reduction with sodium borohydride68 ledneither to hydroxylation nor epoxidation of the 19’,20’-double bond. Since the failure of the hydroxylation might bedue to the heterogeneous nature of the reaction conditionsor the fact that oxymercuration is potentially a reversiblereaction69’70,1H NMR was employed in order to determine ifany addition or coordination actually took place to 19’,20’-double bond. In order to ensure homogeneous reactionconditions mercuric trifluoroacetate in methanol was used71.The NMR spectrum of 19’,20’-anhydrovinblastine 91 indeuterated methanol and in deuterated trifluoroacetic acidrevealed that the position of the olefinic proton (H-19’)128did not change upon addition of an eight fold excess ofmercuric trifluoroacetate, establishing that the mercury iondid not coordinate to the 19’,20’-double bond. Attempts toadd palladium(II) acetate or palladium(II) trifluoroacetateto the 19’,20’-double in l9’,20’-anhydrovinblastine 91 werealso unsuccessful.Treatment of 19’,20’.-anhydrovinblastine 91 with benzylbromide gave an almost quantitative yield of benzylatec319’ ,20’ -anhydrovinblastine 108.1) 05042) H293) Et3NBrH2BrH INOOLINE.02CH391INDOLINE108OHCR3+H H30 HINDOLINE H INDOLINE:02CH3110109129Reaction of 108 with osmium tetroxide in aqueoustetrahydrofuran and subsequent debenzylation withtriethylamine gave two dihydroxy isomers 109 and 110 in 36%and 14% yield respectively. The dihydroxy compounds wereidentified by their molecular ion at 826 m/z. That this ionwas indeed the molecular ion was confirmed by thecharacteristic M + 14 and M + 28 ions64’73. The presenceof the ion fragment of 170 m/z confirmed thatdihydroxylation in the catharanthine half had taken place64.Comparison of the 13C chemical shifts for vinblastine 1 andleurosidine 11372 reveal that significant differences existfor C-3’, C-6’, C-19’, C-20’ and C-21’ (table 2-40). All theremaining chemical shifts are seen to be identical, within 1- 2 ppm. If the chemical shifts of aliphatic methylenes andquaternary carbons of vinbiastine 1 and leurosidine 113 arelisted in descending order (without correlation to theirrespective carbons) as shown in table 2-41 then thedifferences found in table 2-40 stand out. In the 13C APTspectrum of diols 109 and 110 the aliphatic methylenes andquaternary carbons are readily identified, and in table 2-41their chemical shifts are listed in descending order.Comparing the lists of chemical shifts of the diols 109 and110 with those of vinbiastine 1 and leurosidine 113 reveal,that the diol 109 best corresponds with vinbiastine 1 andthe diol 110 best corresponds with the list for leurosidine113. Based on this simple comparison the stereochemistry ofthe C-20’ hydroxy groups of 109 and 110 are tentatively130Table 2_40a• Assigent of th, 13c chemical shifts invinbiastine 1 and leurosidine 1132.R1.C H3Pc02CH31: R1: OH, R2: CH2—CH319’ 18’113: R1: CH2-C3, R2: OH19’ 18’9’6’ 5’111415CH3Position 1 113 Position 1 1132’ 130.9 130.23’ 47.5 43.95 55.5 555b6’ 28.7 21.47’ 115.9 116.88’ 129.0 128.99’ 118.1 117.910’ 122.2 122.011’ 118.8 118.612’ 110.4 110.22 83.0 83.13 50.2 50.25 50.2 50.26 44.3 44.57 52.9 53.18 122.6 123.09 123.1 123.410 120.4 120.411 157.8 157.612 93.9 94.0a The table is continued on the following page.b See the end of the table on the next page.131Table 2-40 continued.Position 1 113 Position 1 11313’ 134.7 134.514’ 29.2 29.815’ 40.0 40.c16’ 55.3 55.417’ 34.118’ 6.7 7.119’ 34.120’ 68.6 71.821’ f3..1 5qqbCH3OCO- 52.1 52.1CH3OQO- 174.6 173.913 152.5 152.814 124.3 124.315 129.7 129.716 79.3 79.517 76.2 76.218 8.1 8.319 30.4 30.720 42.6 42.621 65.2 65.5H30C0- 51.8 51.9CH3OQO- 170.6 170.7H3CO2- 21.7 21.0CH302- 171.4 171.4CH3O- 55.3 55.7CH3N- 38.0 38.2b,c The chemical shifts may be reversed.132respectively. In order to obtain direct evidence for thestereochemistry of the C-20’ hydroxy group in the two diolsan attempt was made to remove the 19’-hydroxy group. The1 113 109 11028.730.834.140.042.344.347.550.050.252 . 955.355. 563.168.679.321.430.735.540.442. 643.944.550.250.253. 155.455. 559.971.879. 5Table 2-41. The chemical shifts of the aliphaticmethylenes and quaternary carbons groups in vinbiastine 1,leurosidine 113 and the two diols 109 and 110aa Solvent: CDC13.correlated to vinbiastine 1 and leurosidine 11328.530.834.737.442. 744.647.850.350.353.355.755.865.871.379 . 718.830.635.538. 542. 543. 544.750.050.252.253. 154.455.171.979. 6133major did 109 did not react with tosyl chloride in pyridinebut treatment with mesyl chloride instead, under the sameconditions, gave one major product 111 on TLC. By additionof water to the reaction mixture the crude product could befiltered of f as a purple solid. This product (111) proved tobe unstable. One of the major decomposition products 112could however be readily obtained on column chromatography.The same material formed in significant amounts on attemptsto purify the crude mesylate by flash chromatography andcrystallized surprisingly easily as thin needles. The 1H NMRspectrum of 112, as well as the elemental analysis, clearlyshowed that a mesylate group still was present. The X-raydiffraction analysis of 112 in figure 2-16 confirmed thepresence of a mesylate group. From the structure of 112 itFigure 2-16. The structure of the mesylate decompositionproduct 112 determined by X-ray diffraction analysis.CH3:0HClCH3mcSO2CR3112134is clear that treatment of the did 109 with mesyl chlorideresulted in mesylation of the hydroxy group at C—16 as wellas C-19’. A subsequent cleavage of the C-0 bond at C-19’followed by a 1,2 shift then lead from the initially formeddimesylate 111 to the rearranged product 112 as outlined inscheme 2-16.R: 1—mesy1vindoNneCH3 CH3SO21CH3Scheme 2-16. Mesylation of the major diol 109 followed bydecomposition to 112.OHH INIJOLINE109S C H3H111 1121353. EXPERIMENTAL3.1. GENERAL EXPERIMENTAL CONDITIONSPhysical data:Melting points (with the solvent used forrecrystallization given in brackets) were determined using aNagle or a Herschberg melting point apparatus and areuncorrected. Infrared spectra were recorded on a PerkinElmer 157, 710B or 1710 spectrometer as KBr pellets.Ultraviolet spectra were recorded on a Cary 15 or aPerkiri Elmer lambda 17 spectrometer using 1 cm quartz cells.Mass spectra were recorded on an AEI-MS-9 (low resolution)or a KRATOS-MS-50 (high resolution) spectrometer employingthe electron impact ionization method. The temperature ofthe probe is given in parenthesis. 1H NMR spectra wererecorded on a Bruker WP-80, Bruker WH-250, Variari XL-300 orBruker WH—400 spectrometer and all chemical shifts arereported in ppm relative to tetramethylsilane as internalstandard. COSY spectra were recorded on a Bruker WH-400spectrometer and 13C APT, HETCOR and SINEPT spectra wererecorded on a Varian XL-300 spectrometer. The X-raystructures were recorded on a Rigaku AFC6S diffractometerwith graphite monochromated Cu Ka radiation. The elementalanalysis were determined using combustion analysis by Dr. P.Borda, Microanalytical Laboratory, University of BritishColumbia. Thin layer chromatography (TLC) was carried out oncommercial aluminum-backed silica plates (Merck art. 5554)136or commercial glass-backed aluminum oxide plates (Merck art.5731). The Rf values given should only be considered asa guide. Visualization was accomplished with ultravioletlight and/or by spraying with 5% ammonium molybdate in 10%aqueous sulfuric acid followed by brief heating, or in thecase of alkaloids, by spraying with 1% ceric ammoniumsulfate in phosphoric acid followed by brief heating.Solvents:Anhydrous ether, tetrahydrofuran, benzene and toluenewere prepared by distillation from sodium in the presence ofbenzophenone. Anhydrous dichloromethane was prepared bydistillation from phosphorous pentoxide. Anhydrous methanolwas prepared by distillation from magnesium. Anhydrouspyridine was prepared by distillation from potassiumhydroxide. Anhydrous acetonitrile was prepared bydistillation from calcium hydride and subsequently storedover 4A molecular sieves. Anhydrous acetone was prepared bydistillation from Drierite.Reagents:All the reagents used were reagent grade materialunless otherwise stated. Trifluoroacetic anhydride waspurified by distillation from phosphorous pentoxide. mchloroperbenzoic acid was purified by suspending 10.0 g ofcommercial m-chloroperbenzoic acid in 200 mL of phosphatebuffer (1.775 g of potassium dihydrogen phosphate and12.23 g of disodium hydrogen phosphate dissolved to 1000 mLin distilled water) and sonicated for 20 mm. The solution137was filtered and the filter cake washed three times with100 mL of the buffer followed by two times with 100 mL ofwater. The filter cake was transferred to a plastic beakerand dried over phosphorous pentoxide, in vacuo, overnight.The purified m-chloroperbenzoic acid was kept in a closedplastic container at 4°C. Kept in this way it will assay74as 95 - 97% peracid for 2- 3 months.Column chromatography:Unless otherwise stated column chromatography wasperformed using TLC grade silica gel (Merck art. 7730) orTLC grade aluminum oxide (neutral) (Merck art. 1090). Thecolumn was pressurized with nitrogen or argon gas pressureto obtain a suitable flow rate.3.2. CATHARANTHINE N-OXIDE 10Catharanthine 3 (400 mg, 1.19 mmol) was dissolved indichloromethane (70 mL) and cooled to -70°C under nitrogen.m-Chloroperbenzoic acid (98%) (210 mg, 1.19 mmol) dissolvedin dichloromethane (10 mL) was added dropwise over 15 mmunder vigorous stirring. After the addition was complete thereaction was stirred for another 10 mm at —70°C and the0H02CH3 H3138solvent was evaporate off at 0°C. The residue was dissolvedin ethyl acetate and loaded on a column (3 cm wide and 8 cmlong) made of silica gel (40 g) suspended in ethyl acetate.Elution successively with ethyl acetate (50 mL), ethylacetate/methanol (9:1) (200 mL) and ethyl acetate/methanol(8:2) (200 mL). The fractions containing the desiredmaterial was evaporated off at 0°C to give catharanthine N-oxide 10 (210 mg, 50%) and 7-hydroxycatharanthine N-oxide 92(70 mg, 16%).The NM?. and MS data of 10 was found to be inidentical to the data published in the iiterature2Ob,.3.3. N-BENZYLOXYCARBONYL-1, 2-DIHYDROPYRIDINE 30°r°-hA dry 1 liter three—necked flask was equipped with amechanical stirrer, a dropping funnel, a thermometer and abubbler. The apparatus was purged with argon and sodiumborohydride (95%) (13.1 g, 0.33 mol) was introduced andcooled to -78°C in a dry ice/acetone cooling bath. Drymethanol (115 mL) was added slowly. The reaction wasexothermic. Dry pyridine (25.8 mL, 0.32 mol) was then addedand the resulting white slurry was cooled to -78°C. Benzyl139chioroformate (45.0 mL, 0.30 mci) was dissoived in drydiethyl ether (35 mL) in a dropping funnel and the solutionadded dropwise at such a rate that the temperature was keptbelow -70°C (the addition took approximately 1 hr 30 mm).The reaction was stirred for another 3 hrs at -78°C. Thecold reaction mixture was poured into 1 liter of ice and thereaction vessel rinsed with diethyl ether (2 x 50 mL).Sodium chloride (10 g) was added in order to break theemulsion. The cold solution was kept under argon andextracted with diethyl ether (5 x 100 mL). The combinedextracts were washed successively with 1 M sodium hydroxide(50 mL), 1 M hydrochloric acid (50 mL), 2.5% sodiumcarbonate (100 mL) and water (2 x 100 mL). The organic phasewas dried over magnesium sulfate, filtered and the filtercake washed with diethyl ether (2 x 50 mL). The solvent wasremoved on rotary evaporator at 35°C under vacuum. Furtherdrying in vacuo gave 30 (56 g, 86%) as a slightly yellowoil, which could be kept for a few weeks when stored below -30°C and under argon.Physical data of 30:TLC (silica, benzene/EtOAc 9:1) Rf: 0.78.NMR (400 MHz, CDC13): Conformer A (60%) 6 (ppm): 7.36(broad s, 5 H, Ph-H), 6.71 (d, J = 7.0 Hz, 1 H, C6-H), 5.83(m, 1 H, C3-H), 5.53 (broad s, 1 H, olefinic H), 5.20 (s, 2H, -CH2O-), 5.10 (broad s, 1 H, olefinic H), 4.39 (m, 2 H, 2x C2-H).140Conformer B (40%) 6 (ppm): 7.36 (broad s, 5 H, Ph-H), 6.79(d, J = 7.0 Hz, 1 H, C6-H), 5.83 Cm, 1 H, C3-H), 5.45 (broads, 1 H, olefinic H), 5.20 (s, 2 H, -CH2O-), 5.20 (1 H,olefinic H), 4.39 Cm, 2 H, 2 x C2-H).3.4. a-CHLORO METHYL ACRYLATE 3138139:>=0cH3A 2 liter three-necked flask was equipped with amechanical stirrer, a dropping funnel, a thermometer and aref lux condenser. Trichloroethylene (180 mL, 2.01 mol), 98%sulfuric acid (400 mL) and basic copper carbonate (3.2 g,0.03 mol) were mixed and heated to 70°C. Aqueousformaldehyde (37%) (150 g, 1.85 mol) in methanol (80 mL,1.98 mol) was added to the reaction, while stirringvigorously, over a period of 35 mm. The temperature of thereaction was kept between 60 to 65°C during the addition.The temperature was raised to 90°C and methanol (80 mL, 1.98mol) added over a period of 15 mm, hydrochloric acid wasevolved. The reaction temperature was raised to 100°C andkept at this temperature for another hour. The reactionmixture was then steam distilled (paraffin oil (2 mL) wasadded to the reaction mixture in order to prevent foamingand hydroquinone was added to the receiver flask). The141temperature in the reaction vessel was kept at about 130°Cduring the steam distillation. The organic phase wasseparated from the aqueous phase and the latter extractedwith dichioromethane (2 x 50 mL). The combined organicphases were washed with 5% sodium bicarbonate solution(70 mL), dried over magnesium sulfate, filtered and thefilter cake washed with dichloromethane (2 x 10 mL). Thecrude product was fractionally distilled (57- 58°C at 57 -64 mm Hg). Compound 31 (77 g, 35%) was obtained as acolorless liquid.Physical data of 31:IR (neat on KBr) Vmax (cm): 2950 (C-H stretch), 1730 (C=0stretch), 1605 (C=C stretch), 1280 (C-0 stretch), 1120 (C-0stretch).NMR (80 MHz, CDC13) 6 (ppm): 6.53 (d, J = 1.5 Hz, 1 H,C3-H), 6.03 (d, J = 1.5 Hz, 1 H, C3-H), 3.85 (s, 3 H, -CO2H3).1423.5. N-BENZYLOXYCARBONYL-endo-7-METHOXYCARBONYL-7-CHLORO-2-AZABICYCLO [2,2,2] OCTAN-5-ENE 32 and N-BENZYLOXY-CARBONYL-exo---7-METHOXYCARBONYL-7-CHLORO-2-AZABICYCLO[2,2,2] OCTAN-5-ENE 33PA 500 mL flask was equipped with a magnetic stirringbar, a reflux condenser and a bubbler and the apparatus waspurged with a positive pressure of nitrogen. Compound 30(53.0 g, 0.229 mol) and compound 31 (45.0 g, 0.374 mol) wasdissolved in benzene (250 mL). The reaction mixture washeated under ref lux for 6 days. The solvent was removed byevaporation on a rotary evaporator at 45°C at fullvacuum giving a dark orange oil (90.5 g). This oil wasdivided into three portions and each portion was purified bychromatography on a column (7.5 cm wide and 5.5 cm long)made of TLC grade silica gel (150 g) suspended in benzene.The crude oil was loaded on the column and eluted withbenzene at flow rate of 6- 7 mL/min. A partly purifiedproduct (54.2 g, 70%) was obtained. 3y drying the partlypurified product on a vacuum line compound 33 slowlycrystallized out. Hexanes/ethyl acetate (8.5:1.5) (50 mL)CH 1C3202CH333143was added to the partly crystalline oil. The obtainedsuspension was filtered and the crystals washed withhexanes/ethyl acetate (8.5:1.5) (2 x 10 mL) and then withhexanes (2 x 10 mL). The mother liquor was evaporated todryness and the residue chromatographed, 10 - 12 g at atime, on a column (7.5 cm wide and 9 cm long) made ofTLC grade silica gel (200 g) suspended in benzene. Thecolumn was pressurized until a flow rate of 4 mL/min wasobtained. When mixed exo and endo adduct began to elute offthe column the eluent was changed to benzene/ethyl acetate(8:2). This partial crystallization and subsequent columnchromatography of the mother liquor was repeated until thetwo isomers had been separated. In this way compound 33(19.5 g, 26%) was obtained as a white solid and compound 32(27 g, 35%) as a yellowish oil.Physical data of 32:M.p.: 44- 49°C (the solidified oil).TLC (silica, benzene/EtOAc 9:1) Rf: 0.58.UV (CH3CN) ‘>max (nm) (log E): 207 (4.10), 250 (2.48), 257(2.46), 262 (2.36), 267 (2.22).IR (KBr) Umax (cm): 3005 (=C-H stretch), 2950 (C-Hstretch), 2855 (C-H stretch), 1730 (ester C=0 stretch), 1685(carbamate C=0 stretch).‘H NMR (400 MHz, CDC13): Conformer A (52%) 6 (ppm): 7.35 (m,5 H, Ph-H), 6.40- 6.60 (m, 2 H, C5-H, C6-H), 4.97 - 5.31(m, 3 H, Ph-CH2O-, C1-H), 3.70 (s, 3 H, -CO2CH3), 3.27 (d, J144= 10 Hz, 1 H), 3.02 (d, J = 14 Hz, 1 H), 2.94 (broad d, J =10 Hz, 1 H), 2.87 (broad s, 1 H, C4-H), 1.84 (broad d, J =14 Hz, 1 H).Conformer B (48%) 8 (ppm): 7.35 (m, 5 H, Ph-H), 6.40 - 6.60(m, 2 H, C5-H, C6—H), 4.97 - 5.31 (m, 3 H, Ph-CH2O-, C1-H),3.53 (s, 3 H, -CO2CI-{3), 3.20 (d, J = 10 Hz, 1 H), 3.02 (d, J= 14 Hz, 1 H), 2.94 (broad d, J = 10 Hz, 1 H), 2.87 (broads, 1 H, C4-H), 1.84 (broad d, J = 14 Hz, 1 H).MS (120°C) m/z (relative intensity 10%): 337/335 (M,<0.1/0.1), 215 (31.4), 170 (30.7), 92 (13.9), 91 (100), 80(18.1).Elemental analysis: Caic. for C17H81N04: C: 60.81, H:5.40, Cl: 10.56, N: 4.17. Found: C: 60.80, H: 5.58, Cl:10.56, N: 4.14.Physical data of 33:M.p.: 97 - 98°C (EtOAc/hexanes 3:17).TLC (silica, benzene/EtOAc 9:1) Hf: 0.52.UV (CH3CN) ‘Xmax (nm) (log E): 204(4.05), 250(2.25),256(2.32), 263 (2.21), 266 (2.02).IR (KBr) Umax (cm): 3020 (=C-H stretch), 2920 (C-Hstretch), 2850 (C-H stretch), 1745 (ester C=0 stretch), 1685(carbamate C=0 stretch).NMR (400 MHz, CDC13): Conformer A (57%) 8 (ppm): 7.36 (m,5 H, Ph-H), 6.44 (t, J = 7.2 Hz, 1 H, C5-H), 6.36 (t, J =7.2 Hz, 1 H, C6-H), 5.30 (d, J = 7.2 Hz, 1 H, C1-H), 5.20(d, J = 12 Hz, 1 H, Ph-CH2O ), 5.17 (d, J = 12 Hz, 1 H, Ph-145CH2O-), 3.77 (s, 3 H, CO2H3), 3.50 Cdt, J = 10, J’ = 2 Hz,1 H, C3-H), 3.07 (dt, J = 10, J’ = 2 Hz, 1 H, C3-H), 2.84(broad s, 1 H, C4-H), 2.73 (broad d, J = 14 Hz, 1 H, C7-H),1.98 (dd, J = 14, J’ = 2 Hz, 1 H, C8—H).Conformer B (43%) 5 (ppm): 7.36 (m, 5 H, Ph-H), 6.48 (t, J =7.2 Hz, 1 H, C5—H), 6.31 (t, J = 7.2 Hz, 1 H, C6-H), 5.16(d, J = 7.2 Hz, 1 H, C1—H), 5.20 (d, J = 12 Hz, 1 H, Ph—CH2O-), 5.17 (d, J = 12 Hz, 1 H, Ph-CH2O ), 3.76 (s, 3 H,CO2H3), 3.50 Cdt, J = 10, J’ = 2 Hz, 1 H, C3-H), 3.07 Cdt,J = 10, J’ = 2 Hz, 1 H, C3-H), 2.88 (broad s, 1 H, C4-H),2.73 (broad d, J 14 Hz, 1 H, C7-H), 1.99 (dd, J = 14, J’ =2 Hz, 1 H, C8-H).MS (150°C) m/z (relative intensity 10%): 337/335 (M,<0.1/<0.1), 215 (10.4), 170 (18.8), 92 (12.8), 91 (100), 80(14.3), 77(11.1), 65(12.9).Elemental analysis: Caic. for C17H81N04: C: 60.81, H:5.40, Cl: 10.56, N: 4.17. Found: C: 61.00, H: 5.49, Cl:10.42, N: 4.20.1463.6. N-BENZYLOXYCARBONYL-endo-7-METHOXYCARBONYL-7-CHLORO-2-AZABICYCLO [2,2,2] OCTAN-endo-6-OL 34 and N-BENZYLOXYCARBONYL-endo-7-METHOXYCARBONYL-7-CHLORO-2-AZABICYCLO[2,2,2] OCTAN-enc7o-5-OL 35A 100 mL three-necked flask was fitted with amechanical stirrer, a thermometer and a dropping funnelequipped with a bubbler. The apparatus was dried in the ovenovernight, assembled while still hot and cooled down under apositive nitrogen atmosphere. Compound 32 (5.77 g, 17.2mmol) was dissolved in dry tetrahydrofuran (55 mL) andborane dimethyl sulfide complex (1.70 mL, 17.8 mmol) wasadded dropwise, at such a rate that the temperature was keptat 26 - 28°C. The reaction mixture was then stirred for anadditional 4 hrs between 26 and 28°C. The reaction wasmonitored by TLC (silica) using benzene/ethyl acetate (9:1)as eluent. The heating bath was removed and methanol (10 mL)was slowly added with vigorous stirring. Then 3 N sodiumhydroxide (6.0 mL, 18.0 mmol) was added slowly (vigorous gasrelease) followed by the dropwise addition of 30% hydrogenperoxide (6.0 mL, 53 mmol) at such a rate that thePH34 35147temperature of the reaction was kept between 26 and 28°C.The reaction mixture was stirred for another 2 hrs at 26 -28°C, then extracted with ethyl acetate (3 x 20 mL). Theorganic phase was washed successively with 5% sodiumhydrosulfite (2 x 30 mL) (until a negative peroxide test),5% sodium bicarbonate (30 mL) and with brine (2 x 30 mL).The organic phase was dried over magnesium sulphate,filtered and the filter cake washed with ethyl acetate(20 mL). Evaporation of the solvent on a rotary evaporatorat 40°C under vacuum gave the crude product (5.6 g) as awhite solid. This crude product was suspended inhexanes/ethyl acetate (9:1) (20 mL) and filtered. The filtercake was then resuspended in hexanes/ethyl acetate (9:1)(20 mL) on the filter. Drying of the filter cake overphosphorus pentoxide gave 34 (5.15 g, 85%) as a white solid.Evaporation of the mother liquor gave a partly crystallineyellowish oil (0.42 g) which, according to TLC (silica,benzene/ethyl acetate 8:2), was a 1:1 mixture of the twoalcohols 34 and 35, together with other impurities. Compound35 could be isolated from the mother liquor residue byrepeated PLC on 0.5 mm silica plates eluted withdichioromethane/methanol (40:1).Physical data of 34:M.p.: 98 - 100°C (EtOAc/hexanes 3:20).TLC: (silica, CH21/C3O 40:1) Rf: 0.25.148UV (CH3CN)‘max (nm) (log ): 206 (4.00), 251 (2.53), 257(2.55), 260 (2.49), 263 (2.48), 267 (2.40).IR (KEr) Vmax (cm): 3445 (OH stretch), 3030 (=CH stretch),2950 (C-H stretch), 1735 (ester C=O stretch), 1670(carbamate C=0 stretch).NMR (250 MHz, CDC13): Conformer A (58%) 6(ppm): 7.32 (m,5 H, Ph—H), 5.19 (d, J = 15 Hz, 1 H, —CH2O—), 5.03 (d, J =15 Hz, 1 H, -CH2O-), 4.60 (dd, J = = 3.5 Hz, 1 H), 4.10 -4.20 (m, 1 H), 3.68 - 3.80 Cm, 1 H), 3.65 (s, 3 H, CH3O-),3.08 - 3.25 (m, 2 H), 2.70- 2.90 (m, 1 H), 2.03 —2.18 (m, 1H), 1.84— 1.98 (m, 1 H), 1.62— 1.80 (m, 1 H), 1.72 (s, 1H, OH).Conformer B (42%) 8(ppm): 7.32 (m, 5 H, Ph-H), 5.14 (d, J =15 Hz, 1 H, -CH2O-), 5.04 (d, J = 15 Hz, 1 H, -CH2O—), 4.55(ad, J = J’ = 3.5 Hz, 1 H), 4.10 - 4.20 (m, 1 H), 3.68 -3.80 (m, 1 H), 3.48 (s, 3 H, CH3O-), 3.08 - 3.25 (m, 2 H),2.70 - 2.90 (m, 1 H), 2.03 —2.18 (m, 1 H), 1.84 — 1.98 (m, 1H), 1.62— 1.80 (m, 1 H), 1.60 (s, 1 H, OH).MS (150°C) m/z (relative intensity 5%): 355/353 (M,0.3/0.8), 262 (5.6), 92(8.7), 91(100).Elemental analysis: Caic. for C17H201N05: C: 57.71, H:5.70, Cl: 10.02, N: 3.96. Found: C: 57.80, H: 5.74, Cl:9.90, N: 3.93.Physical data of 35:M.p.: 130- 131°C (EtOAc/hexanes 2:20).TLC (silica, CH21/CO 40:1) Rf: 0.20.149UV (CH3CN)‘max (nm) (log E): 205 (2.95), 250 (2.44), 256(2.43), 260 (2.32), 262 (2.29), 266 (2.12).IR (KBr) Umax (cm): 3460 (OH stretch), 3020 (=CH stretch),2940 (C—H stretch), 1735 (ester C=O stretch), 1695(carbamate C=O stretch).‘H NMR (300 MHz, CDC13): Conformer A (54%) 6(ppm): 7.35 (m,5 H, Ph-H, 5.18 Cd, J = 12.2 Hz, 1 H,-CH2O-), 5.07 (d, J =12.2 Hz, 1 H, -CH2O-), 4.58- 4.75 (m, 2 H, C1-H, C5-H),3.68 (s, 3 H, -CO2CH3), 3.33 - 3.42 (m, 1 H), 3.17- 3.30(m, 1 H), 2.97- 3.08 (m, 1 H), 2.08 - 2.28 (m, 2 H), 1.67 -1.97 (m, 2 H), 1.40— 1.52 (m, 1 H).Conformer B (46%) 6(ppm): 7.35 (m, 5 H, Ph-H, 5.21 (d, J =12.2 Hz, 1 H, —CH2O-), 5.07 (d, J = 12.2 Hz, 1 H, -CH2O-),4.58- 4.75 (m, 2 H, C1-H, C5-H), 3.53 (s, 3 H, —CO2CH3),3.33 - 3.42 (m, 1 H), 3.17- 3.30 (m, 1 H), 2.97- 3.08 (m,1 H), 2.08- 2.28 (m, 2 H), 1.67 - 1.97 (m, 2 H), 1.40 -1.52 (m, 1 H).MS (150°C) m/z (relative intensity 5%): 355/353 (M,2.7/6.5), 309 (6.4), 218 (5.6), 92 (11.5), 91 (100).Elemental analysis: Calc. for C17H201N05: C: 57.71, H:5.70, Cl: 10.02, N: 3.96. Found: C: 57.56, H: 5.76, Cl:9.85, N: 3.86.1503.7. N-BENZYLOXYCARBONYL-enc7o-7--METHOXYCARBONYL-7-CHLORO-2-AZABICYCLO [2,2,2] OCTAN-6-ONE 36 and N-BENZYLOXYCARBONYL-endo-7-METHOXYCARBONYL-7-CHLORO-2-AZABICYCLO[2,2,2] OCTAN-5-ONE 37Ph0CH3 oA 250 mL two-necked flask was equipped with amechanical stirrer and a reflux condenser fitted with acalcium chloride guard-tube. The apparatus was dried in theoven overnight. Pyridinium chioroformate (3.0 g, 13.9 mmol)was added to a mixture of the alcohols 34 and 35 (3.0 g,8.48 mmol) dissolved in dry dichioromethane (60 mL). Thereaction mixture was heated to reflux under vigorousstirring and monitored by TLC (silica, dichioromethane/ethylacetate 3:1). After 3 hrs more pyridinium chioroformate(0.42 g, 1.95 mmol) was added and the reaction was heatedunder reflux for another hour. The cold reaction mixture wascarefully filtered through a bed of florisil (4.5 cm wideand 3 cm long). Dichloromethane (40 mL) was added to thereaction vessel, heated to reflux for 15 mm under vigorousstirring and then filtered through the same florisil bed.The reaction vessel was similarly treated two times more136 37151with dichioromethane (40 mL) after which the florisil bedwas rinsed with dichioromethane (6 x 20 mL). It wasnecessary to thoroughly rinse the florisil bed in order toensure that all of the product was washed out. The solutionwas evaporated to dryness and the solid greenish residue wasdissolved in boiling ethyl acetate (5 mL). Hexanes (50 mL)was slowly added, keeping the solution at reflux. The hotsolution was decanted from a green oil which separated outof the solution and this in turn was rinsed with hexanes (2x 5 mL). The combined organics were evaporated to drynessand the residue redissolved in boiling ethyl acetate (6 mL).Hexanes (60 mL) was slowly added while keeping the solutionat reflux. The clear hot solution was seeded and allowed tocool to room temperature and then at 5°C overnight. The nextday the solution was cooled to 0°C and filtered. Thecrystals collected were rinsed with cold (0°C) hexanes(10 mL) and allowed to dry. Compound 36 (2.23 g, 77%) wasobtained as white crystals. Evaporation of the motherliquor gave a turbid oil (0.64 g). Compound 37 can beisolated from this oil by column chromatography. The oil isdissolved in a minimum of hexanes/ethylacetate (8:2) andloaded onto a column made of TLC grade silica gel suspendedin hexanes/ethylacetate (9:1). Elution with hexanes/ethylacetate (9:1) affords 37 (approximately 0.4 g).Physical data of 36:M.p.: 99 - 100°C (EtOAc/hexanes 1:10).152TLC (silica, hexanes/EtOAc 9:1, developed twice) Rf: 0.07.UV (CH3CN) ‘Xmax (nm) (log E): 207 (4.13), 250 (2.44), 256(2.45), 260 (2.34), 262 (2.34), 266 (2.20), 302 (2.20), 311(2.19), 323 (shoulder, 1.92).IR (KBr) Vmax(cm’): 3020 (=CH stretch), 2940 (C-H stretch),1740 (ester C=0 stretch), 1695 (carbamate C=0 stretch).1H NMR (300 MHz, CDC13): Conformer A (60%) 6(ppm): 7.39 (m,5 H, Ph-H), 5.27 (a, J = 12 Hz, 1 H, ArCH2O-), 5.01 (a, J =12 Hz, 1 H, PhCH2O-), 4.85 (s, 1 H, C1-H), 3.51 (s, 3 H,CH3O-), 3.3-3.48 (m, 2 H), 3.10 (broad d, J = 15 Hz, 1 H),2.60 (broad d, J = 16 Hz, 1 H), 2.58 (broad s, 1 H, C4-H),2.42 (broad d, J = 16 Hz, 1 H), 2.16 (broad d, J = 15 Hz, 1H).Conformer B (40%) 6(ppm): 7.39 (m, 5 H, Ph-H), 5.16 (d, J =12 Hz, 1 H, PhCH2O-), 5.09 (d, J = 12 Hz, 1 H, PhCH2O-),4.91 (s, 1 H, C1-H), 3.93 (s, 3 H, CH3O-), 3.3 - 3.48 (m, 2H), 3.10 (broad d, J = 15 Hz, 1 H), 2.60 (broad d, J = 16Hz, 1 H), 2.58 (broad s, 1 H, C4-H), 2.42 (broad d, J = 16Hz, 1 H), 2.16 (broad d, J = 15 Hz, 1 H).MS (140°C) m/z (relative intensity 4%): 353/351 (M, 0.6/1.6), 316 (5.6), 159 (10.1), (24.9), 92 (12.6), 91 (100).Elemental analysis: Caic. for C17H81N05: C: 58.04, H:5.16, N: 3.98, Cl: 10.08. Found: C: 57.86, H: 5.17, N: 4.00,Cl: 9.93.Physical data of 37:TLC (silica, hexanes/EtOAc 9:1, developed twice) Rf: 0.13.153UV (CH3CN) ‘>max (nm): 207, 251, 257, 262, 266 (shoulder).IR (KBr) 1max (cm): 3000 (=CH stretch), 2930 (C-Hstretch), 1735 (ester C=0 stretch), 1700 (carbamate C=0stretch).NMR (300 MHz, CDC13): Conformer A (61%) 6 (ppm): 7.35 (m,5 H, Ph-H), 5.16 (d, J = 12 Hz, 1 H, PhCH2O-), 5.07 (d, J =12 Hz, 1 H, PhCH2O-), 5.01 (dd, J = 3.5 Hz, J’ = 2 Hz, 1 H,C1-H), 3.68 (s, 3 H, CH3O-), 3.4- 3.7 (m, 3 H), 3.02 - 3.18(m, 1 H), 2.50 - 2.75 (m, 2 H), 2.28 (broad d, J = 16 Hz, 1H).Conformer B (39%) 6 (ppm): 7.35 (m, 5 H, Ph-H), 5.25 (d, J =12 Hz, 1 H, PhCH2O-), 5.06 (d, J 12 Hz, 1 H, PhCH2O-),4.88 (dd, J = 3.5 Hz, J’ = 2 Hz, 1 H, C1-H), 3.51 (s, 3 H,CH3O-), 3.4- 3.7 (m, 3 H), 3.02 - 3.18 (m, 1 H), 2.50 -2.75 (m, 2 H), 2.28 (broad d, J = 16 Hz, 1 H).MS (120°C) m/z (relative intensity 4%): 353/351 (M,0.3/0.91, 231 (16.8), 186 (4.6), 159 (4.7), 92 (8.0), 91(100), 65 (8.0).High resolution MS: Calc. for C17H81N05:351.0874, Found:351.0871.1543.8. N-BENZYLOXYCARBONYL-endo-7-METHOXYCARBONYL-7-CHLORO-2-AZABICYCLO [2,2,21 OCTAN-6-ONE 2’,2’-DIMETHYL-1’,3’-PROPANEDIYL ACETAL 38PCH3O ci 0 CH3CH3A 250 mL flask was equipped with a magnetic stirringbar and a Dean-Stark trap fitted with a ref lux condenser anda calcium chloride guard-tube. Compound 36 (4.00 g, 11.4mmol), 2,2-dimethyl-1,3-propanediol (6.00 g, 57.6 mmol) andp-toluene sulfonic acid monohydrate (0.60 g, 3.15 mmcl) wereadded to benzene (100 mL) and the mixture was heated underreflux for 18 hrs. The reaction was monitored by TLC(silica, hexanes/ethyl acetate 8:1). The reaction mixturewas cooled to room temperature, washed with saturated sodiumbicarbonate (30 mL), water (5 x 50 mL) and finally withbrine (50 mL). The organic phase was dried over magnesiumsulfate, filtered and the filter cake washed with benzene (2x 10 mL). Evaporation of the solvent at 40°C under vacuumfollowed by drying in vacuo gave 38 (4.92 g, 99%) as a whitesolid.Physical data of 38:M.p.: 124-126°C (hexanes).155TLC (silica, hexanes/EtOAc 7:3) Rf: 0.33.UV (CH3CN)‘Xmax (nm) (log E): 206 (4.05), 250 (2.38), 256(2.39), 260 (2.28), 262 (2.26), 267 (2.10).IR (KBr) Umax (cm): 2940 (C-H stretch), 1735 (ester C=0stretch), 1700 (carbamate C=0 stretch).NMR (300 MHz, CDC13): Conformer A (64%) 8(ppm): 7.35 (m,5 H, Ph-H ), 5.38 Cs, 1 H, C1-H), 5.19 Cd, J = 12.5 Hz, 1 H,PhCH2O-), 5.02 (d, J = 12.5 Hz, 1 H, PhCH2O-), 3.97 (d, 3 =10 Hz, 1 H), 3.56 (s, 3 H, CH3O-), 3.38 Cd, J = 10 Hz, 1 H),3.32 Cd, 3 = 10 Hz, 1 H), 3.03 - 3.40 (m, 4 H), 2.12 - 2.28(m, 1 H), 1.80 - 2.02 (m, 3 H), 1.22 (s, 3 H, CH3), 0.81(s, 3 H, CH3).Conformer B (36%) 6(ppm): 7.35 (m, 5 H, Ph-H ), 5.19 (d, J =12.5 Hz, 1 H, PhCH2O-), 5.18 Cs, 1 H, C1—H), 5.10 (d, 3 =12.5 Hz, 1 H, PhCH2O-), 3.97 (d, 3 = 10 Hz, 1 H), 3.61 (5, 3H, CH3O-), 3.38 Cd, 3 = 10 Hz, 1 H), 3.32 Cd, 3 = 10 Hz, 1H), 3.03- 3.40 (m, 4 H), 2.12- 2.28 Cm, 1 H), 1.80 -2.02 (m, 3 H), 1.13 (s, 3 H, CH3), 0.68 (s, 3 H, CH3).MS (140°C) m/z(relative intensity 5%): 402 (7.8) (M-Cl),228 (8.7), 154 (14.6), 138 (10.6), 129 (58.0), 128 (13.0),92 (13.3), 91 (100), 69 (26.7), 65 (10.6).Elemental analysis, Caic. for C22H81N0: C: 60.37, H:6.45, N: 3.20, Cl: 8.10. Found: C: 60.16, H: 6.30, N: 3.18,Cl: 7.95.1563.9. N- (3’ ‘-INDOLYLMETHYLENECARBONYL) -endo-7-METHOXY-CARBONYL-7-CHLORO-2-AZABICYCLO [2,2,2] OCTAN-6-ONE2’, 2’ -DIMETHYL-1’,3’ -PROPANEDIYL ACETAL 40CH3H3To a 250 mL flask equipped with a magnetic stirring barwas added 10% palladium on charcoal (0.50 g) followed by drytetrahydrofuran (150 mL). Compound 38 (7.60 g, 17.4 mmol)was added and the reaction stirred at room temperature underan atmosphere of hydrogen for 24 hrs. The solution wasfiltered through a 0.5 cm bed of celite and the celitewashed with tetrahydrofuran (3 x 15 mL). Evaporation of thesolvent at 35°C and subsequent drying in vacuo gave thecrude amine (4.86 g) as a white foam. The crude amine(4.86 g) was dissolved in dichloromethane (50 mL) underargon, 3-indole acetic acid (3.35 g, 19.1 mmol) added andthe reddish suspension was cooled to 0°C in an ice bath.From a dropping funnel was slowly added 1,3-dicyclo-hexylcarbodiimide (4.21 g, 20.4 mmol) dissolved indichloromethane (25 mL). After stirring at 0°C for 5 hrs thereaction mixture was filtered and the filter cake washedwith dichloromethane (2 x 10 mL). The mother liquor was03157washed successively with 1 N sodium hydroxide (2 x 15 mL), 1N hydrochloric acid (2 x 15 mL) and finally with water(15 mL). The combined organic phases were dried overmagnesium sulfate, filtered and the filter cake was washedwith dichloromethane (2 x 10 mL). Evaporation at 35°C undervacuum and further drying in vacuo resulted in the crudeproduct (8.77 g) as a reddish solid. Recrystallization frommethanol gave 40 (4.32 g, 54%) as white crystals.Physical data of 40:M.p.: 215- 216°C (MeOH/CH21/Et 5:2:5).TLC (silica, hexanes/acetone/Et3N10:10:0.6) Rf: 0.49.(neutral alumina, heptane/acetone 1:1) Rf: 0.63.(neutral alumina, CHC13/hexanes/EtN7:3:0.05) Rf: 0.26.(neutral alumina, CH21) Rf: 0.17.UV (CH3CN)‘max (nm) (log E) :221 (4.60), 272 (3.74), 279(3.77), 289 (3.71).IR (KBr) Vm8x (cm): 3250 (indole N-H stretch), 3045 (=C-Hstretch), 2940 (C-H stretch), 2850 (C-H stretch), 1730(ester C=0 stretch), 1640 (carbamate C=0 stretch).NMR (300 MHz, CDC13) 8(ppm): 8,15 (s, 1 H, indolic NH),7.61 (d, J = 7.5 Hz, 1 H, indolic C-H), 7.34 (d, J = 7.5 Hz,1 H, indolic C-H), 7.19 (t, J = 7.5 Hz, 1 H, indolic C-H),7.12 (t, J = 7.5 Hz, 1 H, indolic C-H), 7.07 (s, 1 H, C2,-H), 5.88 (s, 1 H, C1-H), 4.02 (t, J = 12 Hz, 2 H), 3.66 (5,2 H), 3.44 (s, 3 H, CH3OCO-), 3.26 - 3.39 (m, 4 H), 3.18(dt, J = 14 Hz, J’ = 3 Hz, 1 H), 2.19 (broad s, 1 H, C4-H),1581.97 (t, J = 12 Hz, 2 H), 1.81 (dt, J = 14 Hz, J 3 Hz, 1H), 1.21 (s, 3 H), 0.80 (s, 3 H).M.S. (160°C) m/z (relative intensity 5%): 460/462 (M,0.3/0.1), 426 (8.9), 425 (13.2), 424 (38.1), 270 (14.3), 269(6.7), 268 (37.6), 267 (26.8), 236 (9.3), 198 (6.6), 182(11.7), 168 (7.2), 157 (27.4), 154 (27.2), 151 (6.6), 150(6.0), 139 (6.4), 131 (23.1), 130 (100).Elemental analysis: Caic. for C22H91N5: C: 62.54, H:6.34, Cl: 7.69, N: 6.08. Found: C: 62.53, H: 6.27, Cl: 7.52,N: 6.08.3.10. N- (3’ ‘-INDOLYLMETHYLENETHIOCARBONYL) -endo-7-METHOXY-CARBONYL-7-CHLORO-2-AZABICYCLO [2,2,2] OCTAN—6-ONE2’ ,2’-DIMETHYL-l’ ,3’-PROPANEDIYL ACETAL 41•CH3C H3To a dry three-necked 1 liter flask equipped with athermometer, a bubbler and a mechanical stirrer was added 40(17.26 g, 37.44 mmol), Lawesson’s reagent76 (15.80 g, 39.06mmol) and dry toluene (600 mL). The reaction mixture washeated to 75°C in an oil bath for 10 hrs. The reactionmixture was then allowed to cool to room temperature, theCH3O159excess of Lawesson’s reagent filtered off and the filtercake washed with toluene (2 x 30 mL). The combined filtrateswere evaporated to dryness, the residue (25 g) dissolved indichioromethane and filtered through a bed (8.5 cm wide and2 cm thick) of TLC grade alumina (type 60/E) (200 g)suspended in dichioromethane. Evaporation of thefiltrate gave the crude product (20 g). This material waschromatographed three times on a column (8.5 cm high and8.5 cm wide) of TLC grade alumina (neutral without binder)(600 g). Compound 41 (9.43 g, 53%) was obtained as ayellowish foam.Physical data of 41:M.p.: 193- 195°C (MeOH).TLC (neutral alumina, CHC13/hexanes/EtN7:3:0.05) Rf: 0.45.(neutral alumina, CH21) Rf: 0.40UV (CH3CN)‘>‘max (nm) (log E): 196 (4.44), 221 (4.55), 282(4.20), 288 (4.16).IR (KBr) Umax (cm1): 3350 (indolic N-H stretch), 3040 (=C-Hstretch), 2910 (C-H stretch), 2845 (c-H stretch), 1730(ester C=0 stretch).‘H NMR (300 MHz, CDC13) 8(ppm): 8.19 (s, 1 H, indolic N-H),7.52 (d, J = 8 Hz, 1 H, indolic C-H), 7.32 (d, J = 8 Hz, 1H, indolic C-H), 7.18 (t, J = 8 Hz, 1 H, indolic C-H), 7.10(t, J = 8 Hz, 1 H, indolic C-H), 7.04 (s, 1 H, C2,,-H),6.88 (s, 1 H, C1-H), 4.42 (d, J 12 Hz, 1 H), 4.24 (d, J =15 Hz, 1 H), 4.10 (d, J = 15 Hz, 1 H), 4.01 Cd, J = 12 Hz, 1160H), 3.73 Cs, 3 H, CH3OCO-), 3.30 - 3.48 (m, 4 H,), 3.09 Cdt,J = 15 Hz, 3 = 3 Hz 1 H), 2.12 (broad s, 1 H, C4-H), 1.88(m, 2 H), 1.71 (dt, 3 = 15 Hz, J = 3Hz, 1 H), 1.20 Cs, 3 H),0.80 (s, 3 H).M.S. (150°C) m/z (relative intensity 10%): 478/476 (M,5.7/12.1), 443 (16.3), 442 (50.8),. 441 (18.3), 440 (45.2),312 (13.5), 270 (45.2), 268 (29.3), 195 (19.6), 174 (20.4),173 (58.4), 168 (10.4), 154 (11.8), 138 (15.5), 131 (17.6),130 (100).Elemental analysis: Caic. for C24H91N04S: C: 60.43, H:6.13, Cl: 7.43, N: 5.87, S: 6.72. Found: C: 60.15, H: 6.08,Cl: 7.40, N: 5.59, S: 6.62.3.11. N-BENZYLOXYCARBONYL-exo-7-METHOXYCARBONYL-7- -CHLORO-2-AZABICYCLO [2,2,2] OCTAN-endo-6-OL 44 and N-BENZYLOXYCARBONYL-exo-7-METHOXYCARBONYL-7-CHLORO-2-AZABICYCLO[2,2,2] OCTAN-enc7o--5-OL 45 and N-BENZYLOXYCARBONYL-7-METHOXYCARBONYL-2-AZATRICYCLO [2,2,2, o6, 7 OCTANE 46.0 0PhON PC1OHOHCO2H344A 100 mL three-necked flask was fitted with aClCO2H345C46mechanical stirrer, a thermometer, a bubbler and a septum.161The apparatus was dried in the oven overnight, assembled hotand cooled down under a nitrogen atmosphere. Compound 33(1.15 g, 3.42 mmol) was dissolved in dry tetrahydrofuran(10 mL). Borane dimethyl sulfide complex (0.4 mL, 4.2 mmol)was added and the reaction mixture stirred overnight at roomtemperature. The reaction was monitored by TLC (silica,benzene/ethyl acetate 9:1). Methanol (5 mL) was added slowlywith vigorous stirring, and the reaction was stirred foranother hour. The reaction mixture was then cooled to -5°Con a methanol/salt bath and 3 N sodium hydroxide (1.8 mL,5.2 mmol) was added slowly (vigorous release of gas),followed by the dropwise addition of 30% hydrogen peroxide(0.8 mL, 7.8 mmol) at such a rate that the temperature ofthe reaction was kept below 0°C. The reaction mixture wasthen stirred for another 2 hrs at 0°C and at 4°C for another18 hrs. A 5% (w/w) solution of sodium bisulfite (10 mL, 4.8mmol) was added and the organic solvents evaporated underreduced pressure. The remaining aqueous phase was extractedwith ethyl acetate (5 x 25 mL) and the combined organicphase washed successively with 5% sodium bicarbonate (2 x10 mL) and brine (15 mL), dried over magnesium sulphate,filtered and the filter cake washed with ethyl acetate(10 mL). Evaporation of the filtrate on a rotary evaporatorat 40°C under vacuum gave the crude product (1.07 g, 93%) asa clear glass. The crude product was purified on TLC gradesilica gel (50 g) suspended in benzene (column diameter 4cm). The crude product was loaded onto the column dissolved162in benzene and eluted initially with benzene/ethyl acetate(9:1) (200 mL) and then with benzene/ethyl acetate (8:2) ata flow rate of 4 mL/min. A mixture of 44 and 45 (878 mg,73%) was obtained as well as pure compound 46 (62 mg, 6%).Compound 44 and 45 could be separated by PLC on 0.5 mmsilica plates eluted with chloroform/methanol (20:1).Physical data of 44:M.p.: 97- 98°C (EtOAc/hexanes 1:6).TLC (silica, benzene/EtOAc 8:1) Rf: 0.05.(silica, chloroform/methanol 20:1) Rf: 0.24.UV (CH3CN)‘max (nm) (log E) : 204 (3.99), 251 (2.36), 257(2.39), 260 (2.28), 263 (2.28), 267 (2.11).IR (KBr) Vmax (cm1): 3440 (OH stretch), 2940 (C-H stretch),2860 (C-H stretch), 1740 (ester C=0 stretch), 1690(carbamate C=0 stretch).NMR (300 MHz, CDC13) Conformer A (54%) 8(ppm): 7.34 (m, 5H, Ph—H), 5.25 (d, J = 13 Hz, 1 H, —CH2O—), 5.15 (d, J 13Hz, 1 H, -CH2O-), 4.67 (d, J = 2 Hz, 1 H, C1-H), 3.77 - 3.95(m, 1 H), 3.82 (s, 3 H, CH3O2-), 3.50 (s, 2 H), 2.64 - 2.80(m, 1 H), 2.54 (d, J = 4 Hz, 1 H, -OH), 2.08 - 2.20 (m, 1H), 1.90— 2.08 (m, 2 H), 1.31— 1.43 (m, 1 H).Conformer B (46%) ö(ppm): 7.34 (m, 5 H, Ph-H), 5.21 (s, 2 H,-CH2O-), 4.58 (d, J = 2 Hz, 1 H, C1-H), 3.77 — 3.95 (m, 1H), 3.82 (s, 3 H, CH3O2-), 3.47 (s, 2 H), 2.64 - 2.80 (m, 1H), 2.21 (d, J = 6 Hz, 1 H, —OH), 2.08 - 2.20 (m, 1 H), 1.90- 2.08 (m, 2 H), 1.31- 1.43 (m, 1 H).163MS (160°C) m/z (relative intensity 5%): 353/355 (1.5/0.5),141 (6.1), 108 (10.1), 107 (9.5), 105 (6.1), 92 (13.2), 91(100), 80 (7.3), 79 (21.9), 77 (28.2), 73 (7.9), 70 (12.0).Elemental analysis: Caic. for C17H201N05: C: 57.71, H:5.70, N: 3.96, Cl: 10.02. Found: C: 58.00, H: 5.71, N: 3.93,Cl: 10.00.Physical data of 45:M. p.: 112 - 113°C (EtOAc/hexanes 4:11).TLC (silica, benzene/EtOAc 8:1) Rf: 0.05.(silica, chloroform/methanol 20:1) Rf: 0.32.UV (CH3CN) ‘Xmax (nm) (log E): 206 (3.92), 251 (2.30), 257(2.35), 262 (2.22), 267 (2.02).IR (KEr) Vmax (cm): 3400 (0-H stretch), 3000 (=C-Hstretch), 2930 (C-H stretch), 2860 (C-H stretch), 1780(ester C=0 stretch), 1660 (carbamate 0=0 stretch).NMR (300 MHz, CDC13): Conformer A (54%) 8(ppm): 7.38 (m,5 H, Ph-H), 5.24 (d, J = 13 Hz, 1 H, CH2O-), 5.17 (d, J = 13Hz, 1 H, CH2O-), 4.65 (dd, J = 4 Hz, J’ = 2 Hz, 1 H, C1-H),3.90- 3.99 (m, 1 H), 3.72- 3.90 (m, 1 H), 3.82 (s, 3 H,CH3O2—), 3.42 (d, J = 11 Hz, 1 H), 2.78 - 2.88 (m, 1 H),2.05- 2.20 (m, 2 H), 1.92 - 2.04 (m, 1 H), 1.73 - 1.82 (m,1 H), 1.69— 1.73 (m, 1 H).Conformer B (46%) 6(ppm): 7.38 Cm, 5 H, Ph-H), 5.20 (s, 2 H,CH2O-), 4.54 (dd, J = 4 Hz, J’ = 2 Hz, 1 H, C1-H), 3.90 -3.99 (m, 1 H), 3.72- 3.90 (m, 1 H), 3.82 (s, 3 H, CH3O2-),3.42 (d, J = 11 Hz, 1 H), 2.78 - 2.88 (m, 1 H), 2.05 - 2.20164(m, 2 H), 1.92— 2.04 (m, 1 H), 1.73 - 1.82 (m, 1 H), 1.69 —1.73 (m, 1 H).MS (160°C) m/z (relative intensity 5%): 353/355 (0.7/0.4),105 (5.5), 92 (11.2), 91 (100), 77 (14.6).Elemental analysis: Caic. for C17H201N05: C: 57.71, H:5.70, N: 3.96, Cl: 10.02. Found: C: 57.75, H: 5.64, N: 3.87,Cl: 9.95.Physical data of 46:M.p.: 67 - 69°C (EtOH/H20).TLC (silica, benzene/EtOAc 8 : 1) Rf: 0.35.UV (CH3CN)‘max (nm) (log E) : 204 (4.24), 251 (2.25), 256(2.32), 260 (2.22), 262 (2.23), 266 (2.06).IR (KBr) Vmax (cm): 3075 (=CH stretch), 3020 (=CHstretch), 2975 (CH stretch), 2940 (CH stretch), 2870 (CHstretch), 1680 (ester C=0 stretch).‘H NMR (300 MHz, CDC13): Conformer A (55%) 8(ppm): 7.35 (m,5 H, Ph—H), 5.21 (d, J = 12 Hz, 1 H, -CH2O-), 5.08 (d, J12 Hz, 1 H, -CH2O-), 4.26 (d, J = 8 Hz, 1 H, C1-H), 3.66 (s,3 H, CH3O2-), 3.12- 3.33 (m, 2 H), 2.08 - 2.35 (m, 3 H),1.78— 1.99 (m, 2 H), 1.70 (m, 1 H).Conformer B (45%) 6(ppm): 7.35 (m, 5 H, Ph—H), 5.15 (s, 2 H,-CH2O-), 4.12 (d, J = 8 Hz, 1 H, C1-H), 3.66 (s, 3 H,CH3O2-), 3.12 - 3.33 Cm, 2 H), 2.08 - 2.35 (m, 3 H), 1.78 -1.99 (m, 2 H), 1.70 (m, 1 H).MS (150°C) rn/z (relative intensity 10%): 302 (13.1), 301(M, 68.9), 257 (50.0), 242 (14.1), 199 (17.8), 198 (10.7),166 (35.2), 158 (33.2), 139 (36.4), 134 (13.7), 107 (30.0),165106 (11.2), 93 (10.0), 92 (69.7), 91 (100), 81 (16.8), 80(18.5), 79(80.6), 78(11.0), 77(23.2).Elemental analysis: Caic. for C17H91N04: C: 67.76, H:6.36, N: 4.65. Found: C: 67.65, H: 6.25, N: 4.60.3.12. N-BENZYLOXYCARBONYL-exo-7-METHOXYCARBONYL-7-CHLORO-2-AZABICYCLO [2,2,2] OCTAN-6-ONE 47 and N-BENZYLOXYCARBONYL-exo-7-METHOXyCARBONYL--7-CHLORO-2-AZABICyCLO[2,2,2] OCTAN-5-ONE 48 and N-BENZYLOXYCARBONYL-7-METHOXYCARBONYL-2-AZATRICYCLO [2,2,2, o6, OCTANE 46Ph FhCO3H47A 250 mL two-necked flask was equipped with amechanical stirrer and a reflux condenser fitted with acalcium chloride guard-tube. Pyridinium chlorochromate(4.00 g, 18.6 mmol) was added to a solution of crudealcohols (3.86 g) dissolved in dichioromethane (75 mL). Thereaction mixture was heated to reflux with vigorous stirringfor 5 hrs and was monitored by TLC (silica, dichioromethane/ethyl acetate 3:1). The cold reaction mixture wasfiltered carefully through a bed of florisil (3.5 cm wideand 2.5 cm long). Dichioromethane (40 mL) was added to the0 0PhC’48 C H3 4166reaction vessel and heated to reflux for 15 mm undervigorous stirring and then filtered through the sameflorisil bed. This was repeated twice. Then the florisil bedwas rinsed with dichioromethane (6 x 20 mL). The greenishsolution was evaporated to dryness and further drying invacuo gave a greenish oil (3.86 g). This crude product wasadsorbed on silica gel (230 - 400 mesh) (200 g) and loadedonto a column (7.5 cm wide) made of TLC grade silica gel(200 g) suspended in hexanes. Elution with hexanes/ethylacetate (7:3) at a flow rate of 25 mL/min. The columnchromatography had to be performed twice in order to achievecomplete separation, giving 47 (1.412 g, 34% based on 33)as a colorless oil, 48 (576 mg, 14% based on 33) as a whitesolid and 46 (397 mg, 11% based on 33).Physical data of 47:TLC (silica, hexanes/EtOAc 7:3) Rf: 0.10(silica, CH21/EtOAc 3:1) Rf: 0.70UV (CH3CN)‘Xmax (nm): 208, 251, 257, 262, 267.IR (KBr) Umax (cm): 3005 (=C-H stretch), 2875 (C-Hstretch), 1740 (ester C=0 stretch), 1700 (carbamate C=0stretch).NMR (300 MHz, CDC13): Conformer A (57%) 8(ppm): 7.38 (m,5 H, Ph-H), 5.24 (d, J = 13 Hz, 1 H, CH2O-), 5.14 (d, J = 13Hz, 1 H, CH2O-), 4.72 (s, 1 H, C1-H), 3.82 Cs, 3 H, CH3O2-), 3.45 - 3.70 (m, 2 H), 2.78 - 2.90 (m, 1 H), 2.50 - 2.62(m, 1 H), 2.38 (broad s, 2 H), 2.15 - 2.28(m, 1 H).167Conformer B (43%) 6(ppm): 7.38 (m, 5 H, Ph-H), 5.24 (d, J =13 Hz, 1 H, CH2O-), 5.18 Cd, J = 13 Hz, 1 H, CH2O-), 4.84(s, 1 H, C1-H), 3.82 (s, 3 H, CH3O2-), 3.45 - 3.70 Cm, 2H), 2.78 - 2.90 (m, 1 H), 2.50 - 2.62 (m, 1 H), 2.38 (broads, 2 H), 2.15— 2.28 (m, 1 H).MS (150°C) m/z (relative intensity 5%): 351/353 (1.0/0.5),159 (6.1), 158 (18.7), 107 (5.3), 92 (9.6), 91 (100), 79(9.7), 77 (6.7).High resolution MS: Caic. for C17H8105: 351.0874. Found:351.0879.Physical data of 48:M.p.: 133- 135°C (EtOAc/hexanes 1:3)TLC (silica, hexanes/EtOAc 7:3) Rf: 0.15(silica, dichioromethane/EtOAc 3:1) Rf: 0.70UV (CH3CN) max (nm) (log C): 205 (4.00), 246 (2.27), 250(2.32), 256 (2.38), 260 (2.28), 262 (2.30), 266 (2.16), 288(shoulder, 1.38).IR (KBr)‘umax (cm): 3005 (=CH stretch), 2945 (CH stretch),2875 (CH stretch), 1740 (shoulder, ester C=0 stretch), 1705(carbamate C=0 stretch).‘H NMR (300 MHz,CDC13) Conformer A (52%) o(ppm): 7.38 Cm, 5H, Ph—H), 5.24 (d, J = 12 Hz, 1 H, CH2O-), 5.18 (d, J = 12Hz, 1 H, CH2O—), 5.06 (dd, J = 2 Hz, J = 5 Hz, 1 H, C1—H),3.86 (s, 3 H, CH3O2-), 3.67 - 3.82 (m, 2 H), 3.07 - 3.18(m, 1 H), 2.60- 2.78 (m, 2 H), 2.43 (m, 1 H), 2.32 Cm, 1H).168Conformer B (48%) ô(ppm): 7.38 (m, 5 H, Ph-H), 5.23 (s, 2 H,CH20-), 4.94 (dd, J = 2 Hz, J = 5 Hz, 1 H, C1-H), 3.85 (s, 3H, CH3O2-), 3.67 - 3.82 (m, 2 H), 3.07 - 3.18 Cm, 1 H),2.60- 2.78 (m, 2 H), 2.43 (m, 1 H), 2.32 (m, 1 H)MS (120CC) rn/z (relative intensity 5%): 351/353 (1.2/0.4),232 (5.3), 231 (27.2), 186 (9.0), 159 (8.5), 96 (5.8), 92(15.4), 91 (100).Elemental analysis: Calc. for C17H81N05: C: 58.04, H:5.16, N: 3.98, Cl: 10.08. Found: C: 58.23, H: 5.15, N: 3.89,Cl: 9.89.Physical data of 46:As described in the synthesis of 44.3.13. N-BENZYLOXYCARBONYL-exo- 7-METHOXYCARBONYL--7-CHLORO-2-AZABICYCLO [2,2,2] OCTAN-6-ONE 2’ ,2’-DIMETHYL-l’ ,3’-PROPANEDIYL ACETAL 490PCH3CO2H3 CH3A 25 mL flask was equipped with a magnetic stirring barand a Dean-Stark trap fitted with a reflux condenser and acalcium chloride guard-tube. Compound 47 (206 mg, 0.59mmol), 2,2-dimethyl--1,3-propanediol (68 mg, 0.65 mmol) andC’169p-toluene sulfonic acid monohydrate (26 mg, 0.14 mmol) wereadded to benzene (10 mL) and the mixture was heated toreflux for 4 hrs. The reaction mixture was diluted withbenzene (50 mL) and washed with saturated sodium bicarbonate(10 mL) and water (3 x 15 mL). The organic layer was driedover magnesium sulfate, filtered and the filter cake washedwith benzene (25 mL). Evaporation of the solvent and dryingof the residue in vacuo overnight gave an oily white solid(241 mg). This crude product was stirred in hexanes/ethylacetate (5:1) until a fine powder had formed. Filtration andwashing of the filter cake with hexanes (2 x 2.5 mL) gave,after drying in vacuo, 49 (146 mg). The mother liquor wasevaporated to dryness and the residue dissolved indichloromethane (2 mL) and loaded onto a column (1.2 cmwide) made of TLC grade silica gel (2 g) suspended inhexanes. Elution with hexanes/ethyl acetate (7:3) at a flowrate of 5 mL/min gave another 42 mg of 49, giving a totalyield of 188 mg (73%) of 49.Physical data of 49:M.p.: 141- 143°C (EtOAc/hexanes 1:2).TLC (silica, hexanes/EtOAc 7:3) Rf: 0.22/0.38. Double spotdue to the two conformers.UV (CH3CN)‘“max (nm) (log E) : 207 (3.86), 252 (2.46), 257(2.51), 262 (2.46), 266 (2.36, shoulder).IR (KBr) Umax (cm): 2925 (CH stretch), 2860 (CH stretch),1745 (ester C=0 stretch), 1690 (carbamate C=0 stretch).170NMR (300 MHz, CDC13): Conformer A (68%) o(ppm): 7.35 Cm,5 H, Ph-H), 5.26 (d, J 12 Hz, 1 H, PhCH2O-), 5,22 (s, 1 H,C1-H), 5.18 (d, J 12 Hz, 1 H, PhCH2O-), 3.81 Cs, 3 H,CH3O2-), 3.58 (d, J 12 Hz, 1 H, C—CH2O-), 3.10- 3.50 (m,5 H), 2.80— 3.03 Cm, 1 H), 2.15 — 2.28 (m, 1 H), 1.92 -2.07 (m, 1 H), 1.78 (broad s, 2 H), 1.07 (s, 3 H, CH3),0.85 (s, 3 H, CH3 ).Conformer B (32%) 6(ppm): 7.35 Cm, 5 H, Ph-H), 5.40 (d, J =12 Hz, 1 H, PhCH2O-), 5.12 (d, J = 12 Hz, 1 H, PhCH2O-),4.99 (s, 1 H, C1-H), 3.79 (s, 3 H, CH3O2-), 3.73 (d, J = 12Hz, 1 H, C-CH2O-), 3.10- 3.50 (m, 5 H), 2.80 - 3.03 Cm, 1H), 2.15— 2.28 Cm, 1 H), 1.92— 2.07 Cm, 1 H), 1.61 (broads, 2 H), 0.87 Cs, 3 H, CH3), 0.75 (s, 3 H, CH3 ).M.S. (150°C) m/z (relative intensity 5%): 402 (M - Cl,28.6), 154 (5.6), 129 (10.0), 92 (9.4), 91 (100).Elemental analysis: Calc. for C22H81N06: C: 60.34, H:6.45, N: 3.20, Cl: 8.10. Found: C: 60.35, H: 6.45, N: 3.02,Cl: 8.00.1713.14. N- (3’’ -INDOLYLMETHYLENECARBONYL) -exo-7-METHOXY-CARBONYL-7-CHLQRQ-2-AZABICYCLO [2,2,2] OCTAN-6-ONE2’,2’ -DIMETHYL-1’,3’ -PROPANEDIYL ACETAL 51•C H3CH3Compound 49 (3.418 g, 7.81 mmol) was dissolved in drybenzene (50 mL) under argon. Anhydrous hydrogen bromide wasbubbled through the solution for 45 mm, after which thesolution was purged with argon. Dry diethyl ether (200 mL)was slowly added to the vigorously stirred solution until afine white precipitate had formed. The solution was filteredand the filter cake was stirred twice with dry diethyl ether(20 mL). Drying in vacuo gave the crude amine salt (2.730 g,91%) as a white powder.Method A:The crude amine salt (439 mg, 1.14 mmol) was suspendedin dry acetonitrile (10 mL) under argon and triethylamine(210 p1, 1.51 mmol) was added, giving a clear solution. Then3-indole acetyl chloride (290 mg, 150 mmol) wasadded, giving a clear brownish solution which was stirredfor 4.5 hrs.. Water (0.5 mL), followed by triethylamineC’5172(0.2 mL) was added to the now clear yellow solution, givinga dark greenish solution which was evaporated to dryness.The residue was dissolved in a minimum volume ofdichioromethane and loaded onto a flash column (3 cm wideand 14 cm long) packed with silica gel (230 - 400 mesh) inhexanes/ethyl acetate/triethylamine (12:8:0.6). Elution withthe above solvent system at a flow rate of 5 cm/mm gave 51(377 mg, 71% based on 49) as a white powder.Method B:The crude amine salt (800 mg, 2.08 mmol) was suspendedin dry acetonitrile (6 mL) under argon and triethylamine(290 iii, 2.08 mmol) was added to give a thick suspension.Then 3-indole acetic acid (365 mg, 2.08 mmol) was added,giving a clear brownish solution which was cooled to 0°C inan ice bath. 1,3-Dicyclohexylcarbodiimide (472 mg, 2.29mmol) was added and the reaction was stirred for 19 hrs at0°C. The reaction mixture was filtered and the filter cakewashed with dichloromethane (3 x 5 mL). After evaporation ofthe solvent the residue was redissolved in a minimum ofdichioromethane. This solution was loaded onto a flashcolumn (4.5 cm wide and 14 cm long) packed with of silicagel (230- 400 mesh) in hexanes/ethyl acetate/triethylamine(14:6:0.6). Elution, first with hexanes/ethylacetate/triethylamine (14:6:0.6) (1000 mL) then withhexanes/ethyl acetate! triethylamine (12:8:0.6) at a flow173rate of 5 cm/mm gave 51 (559 mg, 58% based on 49) as awhite solid.Physical data of 51:M.p.: 174 — 179°C (MeOH/CH2C1 6:1).TLC (silica, hexanes/acetone/Et3N10:10:0.6) Rf: 0.40.UV (CH3CN)‘max (nm) (log E) : 202 (4.40, shoulder) , 220(4.58), 273 (3.75), 279 (3.77), 2.89 (4.79).IR (KBr) Vmax (cm1): 3200 (N-H stretch), 2900 (C-Hstretch), 1735 (ester C0 stretch), 1630 (amide C=0stretch).‘H NMR (250 MHz, CDC13) ö(ppm): 8.06 (s, 1 H, N-H), 7.59 (d,J = 7.5 Hz, 1 H, indolic C-H), 7.37 (d, J 7.5 Hz, 1 H,indolic C-H), 7.07 - 7.24 (m, 3 H, indolic C-H), 5.84 (s, 1H, C1—H), 3.78- 3.88 Cm, 2 H), 3.80 (s, 3 H, CH3OCO), 3.66(d, J = 10 Hz, 1 H), 3.16 — 3.59 (m, 5 H), 3.01 (dt, J’ = 15Hz, J’’ = 2 Hz, 1 H), 2.13 - 2.20 (m, 1 H), 1.87 - 2.00 (m,1 H), 1.75 (broad s, 2 H), 1.08 (s, 3 H, CH3—C), 0.77 (s, 3H, CH3-C).M.S. (150°C) m/z (relative intensity 5%): 460/462 (M,1.9/0.8), 426 (19.4), 425 (47.8), 424 (39.5), 307 (5.5)270(10.7), 268 (16.1), 236 (11.0), 198 (5.9), 182 (11.1), 168(5.9), 167 (8.5), 157 (15.2), 154 (20.4), 131 (14.7), 130(100), 129 (23.4), 128 (6.2), 103 (5.6), 77 (5.7), 69(19.6), 55 (6.5), 41 (18.6), 36 (5.5).174Elemental analysis: Calc. for C24H91N05: C: 62.54, H:6.34, Cl: 7.69, N: 6.08. Found: C2: 62.73, H: 6.40, Cl:7.55, N: 6.10.3.15. N- (3’ ‘-INDOLYLMETHYLENETHIOCARBONYL) -exo-7-METHOXY-CARBONYL-7-CHLORO-2-AZABICYCLO [2,2,2] OCTAN-6-ONE2’ ,2 ‘-DIMETHYL-l 1,3T -PROPANEDIYL ACETAL 52.CO2H3•CH3Dry benzene (12 mL) was added to 51 (331 mg, 0.718mmol) and 2, 4-bis(thiophenyl)-1, 3-dithia-2, 4-diphosphetane-2,4-disulfide (293 mg, 0.717 mmol) in a dry 25 mL flaskunder argon. The reaction mixture was heated to 60°C in anoil bath for 3.5 hrs and monitored by TLC (neutral alumina,dichioromethane). The solvent was removed at 35°C undervacuum and the residue dissolved in a minimum ofdichioromethane and loaded onto a column (3 cm wide) of madeTLC grade alumina G (neutral, type 60/E) (30 g) suspended indichioromethane. Elution with dichioromethane at a flow rateof 6 mL/min gave the crude product (261 mg) which wasdissolved in a minimum volume of dichioromethane and loadedH’ 5175equally onto two PLC plates (neutral alumina, 1 mm, 20 x20 cm) and eluted with dichioromethane. The partly separatedproducts so obtained were loaded onto new alumina plates andre-eluted with dichioromethane. In this way 52A (120 mg) and52B (89 mg) were isolated.Physical data of 52A:TLC (neutral alumina, CH21) Rf: 0.27.UV (CH3CN) ‘>max (nm): 195, 220, 280, 288 (shoulder).IR (KBr) 1max (cm1): 3320 (N-H stretch), 2910 (C-Hstretch), 2840 (C-H stretch), 1735 (ester C=0 stretch).‘H NMR (300 MHz, CDC13) 6(ppm): 8.36 (s, 1 H, indole N-H),7.60 (d, J = 8 Hz, 1 H, indolic C-H), 7.38 Cd, J = 8 Hz, 1H, indolic C-H), 7.25 Cs, 1 H, C2,,-H), 7.22 (t, J = 8 Hz, 1H, indolic C-H), 7.14 (t, J 8 Hz, 1 H, indolic C-H), 7.05(s, 1 H, C1-H), 4.30 (s, 2 H, -CH2O O-), 4.24 (d, J = 12 Hz,1 H), 3.85 (s, 3 H, CH3O-), 3.72 (d, J = 11.2 Hz, 1 H), 3.64Cd, J = 12 Hz, 1 H), 3.54 Cdt, J = 12 Hz, J’ = 2 Hz, 1 H),3.36 (dd, J = 11.2 Hz, J’ = 2 Hz, 1 H), 3.27 (dd, J = 11.2Hz, J’ = 2 Hz, 1 H), 3.06 Cdt, J = 16 Hz, J’ = 2 Hz, 1 H),2.15 (broad s, 1 H), 1.91 (broad d, J = 16 Hz, 1 H), 1.75(s, 2 H), 1.11 (s, 3 H), 0.81 (s, 3 H).M.S. C220°C) m/z (relative intensity): 476/478 (M,3.1/1.1), 442 (11.8), 441 (17.6), 440 (50.8), 354 (21.2),268 (13.2), 195 (15.6), 182 (12.8), 174 (12.6), 173 (46.9),154 (28.8), 139 (10.7), 138 (22.1), 131 (12.6), 130 (84.7),129 (22.4), 128 (17.4), 77 (11.7), 69 (50.7), 68 (25.5), 59176(12.0), 57 (16.1), 56 (19.3), 55 (29.8), 45 (11.4), 44(19.7), 43(21.8), 42(13.5), 41(70.6).Physical data of 52B:TLC (neutral alumina, CH21) Rf: 0.15.UV (CH3CN)‘>%max (nm): 221, 270, 279 (shoulder).IR (KEr) umax (cm): 3350 (N-H stretch), 2900 (C-Hstretch), 2825 (C-H stretch), 1730 (ester C=0 stretch).NMR (300 MHz, CDC13) o(ppm): 8.11 Cs, 1 H, indole N-H),7.75 Cd, J = 8 Hz, 1 H, indolic C-H), 7.37 (d, J = 8 Hz, 1H, indolic C—H), 7.32 (s, 1 H, C27,-H), 7.19 Ct, J = 8 Hz, 1H, indolic C-H), 7.14 (t, J = 8 Hz, 1 H, indolic C-H), 5.05Cs, 1 H, C1-H), 4.57 (q, J = 15.2, 2 H, -CH2O O-), 3.94 (dt,J = 14 Hz, J’ = 2 Hz, 1 H), 3.75 ((d, J = 14 Hz, 1 H), 3.46(s, 3 H, CH3O-), 3.38 Cd, J = 11.2 Hz, 1 H), 3.32 (d, J =11.2 Hz, 1 H), 3.12 (t, J = 11.2 Hz, 2 H), 2.80 Cd, J = 15.2Hz, 1 H), 2.43 (broad s, 1 H), 1.90 - 2.08 Cm, 2 H), 1.70(broad d, J = 15.2 Hz, 1 H), 0.87 (s, 3 I-I), 0.73 (s, 3 H).M.S. (220°C) m/z (relative intensity): 476/478 (M,3.8/1.4), 442 (13.3), 441 (14.2), 440 (36.5), 354 (12.1),268 (15.1), 195 (11.7), 182 (23.0), 174 (14.3), 173 (53.4),168 (10.7), 166 (12.1), 155 (12.1), 154 (28.9), 140 (16.2),139 (24.9), 138 (38.7), 131 (14.2), 130 (100), 129 (22.0),128 (16.3), 124 (10.6), 80 (18.9), 77 (11.0), 71 (11.4), 69(51.1), 68 (25.8), 59 (10.4), 59 (14.1), 57 (19.7), 56(19.1), 55 (33.4), 45 (11.4), 44 (17.0), 43 (26.0), 42(11.7), 41 (72.1).1773.16. 20—DESETHYL-15, 20-ANHYDRO-5-OXO-CATHARANTHIN-20-ONE2’, 2’ -DIMETHYL-1’ , 3’ -PROPANEDIYL ACETAL 55 AND THEISOMERIC BY-PRODUCT 56C H3CH3L0C H3•CH3CH3Sodium hydrogen carbonate (541 mg, 6.4 mmol) dissolvedin water (180 mL) was added to 40 (300 mg, 0.65 mmol)dissolved in methanol. The mixture was degassed by bubblingwith argon for 6 mm. The solution slowly became a thinwhite suspension which was photolysed for 20 mm using aVycor filter and with so little cooling that the heat fromthe lamp heated the solution to reflux. The methanol wasevaporated from the resulting clear yellowish solution andthe remaining aqueous phase was extracted five times with40 mL of ethyl acetate. Sodium chloride had to be added inorder to prevent formation of an emulsion. The combinedorganic extracts were dried over magnesium sulfate, filteredand the filter cake washed with ethyl acetate (2 x 20 mL).Evaporation of the solvent at 40°C gave yellow-brown oil(258 mg) which was dissolved in dichioromethane (2 mL) andHCO2H55 56178loaded equally onto four 2 mm silica PLC plates (20 x20 cm). The plates were eluted once with each of thefollowing solvent systems; hexanes/ethyl acetate/diethylamine (30:60:4), hexanes/ethyl acetate (30:60) and finallywith hexanes/ethyl acetate/diethylamine 30:60:10. The secondband from the top (Rf = 0.7) gave 55 (11 mg, 4%) as a whitesolid. The last band (Rf 0.3) gave partly purified 56(44 mg) which was dissolved in dichloromethane (1 mL) andloaded equally on to two 0.5 mm PLC silica plates (20 x20 cm) . The plates were developed twice with hexanes/ethylacetate 1:3. Compound 56 (and 41 mg, 15%) was obtained as awhite solid.The compounds 55 and 56 were not synthesized on largescale from the amide 51. That the amide 51 does lead to 55and 56 under the photochemical conditions was established bycomparative TLC and HPLC with isolated 55 and 56.Physical data of 55:TLC (silica, hexanes/EtOAc 1:2) Rf: 0.54.UV (CH3CN)’Xmax (nm): 192, 219, 282, 290.IR (KBr) Vmax (cm): 3400 (indole N-H stretch), 3280(indole N-H stretch), 3050 (=C-H stretch), 2945 (C-Hstretch), 2925 (C-H stretch), 2865 (C-H stretch), 2850 (C-Hstretch), 1735 (ester C=0 stretch), 1645 (lactame C=0stretch).NMR (400 MHz, CDCN) 8(ppm): 9.06 (s, 1 H, N-H), 7.54 Cd,J 8.0 Hz, 1 H, indolic C-H), 7.28 (d, J = 8.0 Hz, 1 H,179indolic C-H), 7.12 Ct, J = 8.0 Hz, 1 H, indolic C-H), 7.07Ct, J = 8.0 Hz, 1 H, indolic C-H), 5.25 Cs, 1 H, C21-H),4.27 Cd, J = 16 Hz, 1 H), 3.74 Cd, J = 16 Hz, 1 H), 3.55 -3.65 (m, 2 H), 3.58 Cs, 3 H, CH3OCO), 3.42 - 3.53 (m, 2 H),3.32 (d, J = 12 Hz, 1 H), 2.98 - 3.07 Cm, 2 H), 1.75 - 1.87Cm, 2 H), 1.46 (a, J = 14 Hz, 1 H), 1.11 Cs, 3 H, CH3—C),0.85 (s, 3 H, CH3-C).MS C180°C) m/z (relative intensity): 424 (M, 89.0), 296C17.2), 267 (16.6), 255 (12.0), 241 C10.3), 223 (23.5), 214(18.4), 209 (10.3), 207 (10.5), 195 (14.7), 194 C15.4), 182(21.7), 181 (15.6), 180 (16.9), 169 C13.5), 168 (29.9), 168C100), 155 (18.5), 154 (41.6), 130 (28.7), 129 (14.3), 128C17.1), 127 C12.5), 81 C12.3), 69 (45.4), 68 C14.2), 54(16.5)High resolution MS: Caic. for C24H8N05 : 424.1998. Found:424.1998.Physical data of 56:TLC (silica, hexanes/EtOAc 1:2) Rf: 0.15.UV (CH3CN)‘>‘max (nm): 198, 226, 288, 296.IR (KBr) Vmax (cm): 3400 (indole N-H stretch), 3260(indole N-H stretch), 3040 (=C-H stretch), 2940 (C-Hstretch), 2865 (C-H stretch), 1732 (ester C=0 stretch), 1635(lactame C=0 stretch).NMR (400 MHz, CDC13) ö(ppm): 8.72 (s, 1 H, indolic N-H),7.25 (d, J = 8.0 Hz, 1 H, indolic C-H), 7.09 (t, J = 8.0 Hz,1 H, C11-H), 6.94 (s, 1 H, C2-H), 6.84 Cd, J = 8.0 Hz, 1 H,180indolic C-H), 5.38 Cs, 1 H, C2-H), 4.25 (d, J = 14.0 Hz, 1H), 3.47- 3.60 (m, 5 H, C-CH2-0), 3.45 (s, 3 H, CH3OCO-),3.33 - 3.41 (m, 2 H), 3.24 (d, J = 12 Hz, 1 H), 2.26 (broads, 1 H), 1.93— 2.04 (m, 2 H), 1.75 Cd, J = 14 Hz, 1 H),1.06 (s, 3 H, CH3-C), 0.90 (s, 3 H, CH3-C).MS (180°C) m/z (relative intensity): 424 (M, 100), 392(7.7), 365 (6.7), 338 (12.1), 294 (17.6), 251 (10.2), 223(11.5), 208 (11.8), 207 (13.3), 196 (12.8), 195 (10.7), 194(13.6), 182 (15.4), 181 (15.0), 180 (12.2), 169 (15.0), 168(28.5), 167 (71.8), 155 (16.9), 154 (59.2), 149 (11.4), 141(10.2), 130 (10.2), 129 (10.0), 128 (11.5), 127 (16.7), 115(11.5), 70(22.3), 69(54.2).High resolution MS: Caic. for C24H8N05 : 424.1998. Found:424.2016.3.17. 20-DESETHYL-15, 20-ANHYDRO-5-THIOXO-CATHARANTHIN-20-ONE2’, 2’ -DIMETHYL-1’ ,3! -PROPANEDIYL ACETAL 61 AND THEISOMERIC BY-PRODUCT 62C•C H3C H3S 0 SHC02H61 62181Sodium hydrogen carbonate (146 mg, 1.74 mmol) dissolvedin water (300 mL) was added to 41 (319 mg, 0.67 mmol)dissolved in methanol (450 mL). The mixture was degassed bybubbling with nitrogen for 10 mm and the solution was thenphotolysed for 50 mm at room temperature using a Vycorfilter. The methanol was evaporated from the clear yellowishsolution and the remaining aqueous phase was extracted withchloroform (5 x 50 mL). Sodium chloride had to be added inorder to prevent formation of an emulsion. The combinedextracts were dried over magnesium sulfate, filtered and thefilter cake washed with chloroform (2 x 20 mL). Evaporationof the combined filtrates at 40°C gave a brown oil (352 mg).The crude product was partly purified by columnchromatography on deactivated neutral alumina usingchioroform/hexanes (2:1) as eluent. The partly purifiedmaterial so obtained was then loaded on 0.5 mm PLC plates(20 cm x 20 cm) and eluted three times withdichloromethane/methanol (200:1). Compound 61 (31.4 mg, 11%)and compound 62 (28.6 mg, 10%) were obtained as whitesolids.The compounds 61 and 62 were not synthesized onlarge scale from the thioamides 52A and 52B. That thethioamides 52A and 52B do lead to 61 and 62 under the photochemical conditions was established by comparative HPLC withisolated 61 and 62.182Physical data of 61:TLC (neutral alumina, CHC13/hexanes 2:1) Rf: 0.15.UV (CHCN) ‘>max (nm): 197, 223, 272, 289 (shoulder)IR (KBr) Umax (cm): 3340 (indole N-H stretch), 2945 (C-Hstretch), 2920 (C-H stretch), 2865 (C-H stretch), 2850 (C-Hstretch), 1735 (ester C=0 stretch), 1718 (ester C=0stretch).1H NMR (300 MHz, CDC13) 8(ppm): 7.86 (s, 1 H, indolic N-H),7.58 (d, J = 7.4 Hz, 1 H, indolic C-H), 7.28 (d, J = 7.4 Hz,1 H, indolic N-H), 7.17 (t, J = 7.4 Hz, 1 H, indolic N-H),7.13 Ct, J = 7.4 Hz, indolic C-H), 5.52 Cs, 1 H, C21-H),4.50 (dd, J = 16 Hz, 2 H, C6-H), 4.05 (broad d, J = 13 Hz),3.62 (s, 3 H, CH3OC-), 3.35 - 3.65 (m , 5 H), 3.06 (broad d,J = 13 Hz), 2.36 (broad s, 1 H), 2.15 (broad d, J = 13 Hz, 1H), 1.85 (d, J = 13 Hz, 1 H), 1.56 (d, J = 13 Hz, 1 H), 1.08(s, 3 H), 0.92 Cs, 3 H)MS (120°C) m/z (relative intensity): 440 (M, 100), 408(9.2), 354 (61.1), 321 (15.9), 295 (15.9), 280 (10.3), 267(34.9).High resolution MS: Calc. for C24H8N04S: 440.1770. Found:440.1757.Physical data of 62:TLC (neutral alumina, CHC13/hexanes 2:1) Rf: 0.10.UV (CH3CN)‘max (nm): 193, 228, 272, 295 (shoulder).NMR (300 MHz, CDC13) 8(ppm): 8.50 (s, 1 H, indolic N-H),7.29 (d, J 7.0 Hz, 1 H), 7.22 Cs, 1 H, C2,-H), 7.11 (t, J183= 7.0 Hz, 1 H, C11-H), 6.81 (d, J = 7.0 Hz, 1 H), 5.78 Cs, 1H, C2—H), 4.65 (d, J = 14, 1 H, C5—H), 4.28 (d, J = 14, 1 H,C6-H), 3.47 Cs, 1 H, CH3O2—), 3.38 - 3.70 (m, 7 H), 2.34(broad s, 1 H), 2.07 (broad d, J = 14 Hz, 1 H), 1.90 (d, J =14 Hz, 1 H), 1.75 (d, J = 14 Hz, 1 H), 1.02 (s, 3 H), 0.93(s, 3 H).MS (150°C) m/z (relative intensity): 440 (M, 4.7), 353(0.6), 326 (1.0), 130 (20.5), 83 (100).High resolution MS: Caic. for C24H8N04S: 440.1770. Found:440.1763.3.18. EXOCATHARANTHINE 89-CH3In a 1000 mL three-necked flask equipped with athermometer, a ref lux condenser, a magnetic stirrer and arubber septum was placed 10% palladium on carbon (700 mg).The apparatus was flushed with nitrogen and toluene(spectro grade) (600 mL) containing 0.05% thiophene wasadded followed by catharanthine 3 (1509 mg). The reactionvessel was attached to a hydrogenation apparatus through thetop of the reflux condenser, evacuated and refilled withhydrogen without stirring. The reaction was heated to 70°CH184and stirred at a moderate rate for 3 hours. The hot solutionwas filtered and the filter paper was washed with hottoluene (2 x 100 mL). The solvent was removed and the crudeproduct was purified by flash chromatography usinghexanes/ethyl acetate/triethyl amine 10:10:0.6 as eluent.Compound 89 (1249 mg, 83%) was obtained as a white solid. Itshoul be noted that compound 89 is a mixture of the E andthe Z isomer in a typical ratio of 5:1.Physical data of 89.TLC (silica, hexanes/acetone/Et3N10:10:0.1) Rf: 0.47.UV (CH3CN) Xmax (nm): 223 (4.49), 232 (3.89), 241 (3.84).IR (KBr) Vmax (cm): 3370 (indole N-H stretch ), 3050 (=CH stretch), 2935 (C-H stretch), 2845 (C-H stretch), 1750(ester C0 stretch).NMR (300 MHz, CDC13) 8(ppm): 7.65 (s, 1 H, indole N-H),7.50 (d, J = 8 Hz, 1 H, C9-H), 7.25 (d, J 8 Hz, 1 H, C12-H), 7.16 (t, J = 8 Hz, 1 H, C11—H), 7.07 (t, J = 8 Hz, 1 H,C10-H), 5.36 (q, J = 6 Hz, 1 H, C19-H), 4.00 (s, 1 H, C21-H), 3.70 (s, 3 H, CH3O), 3.43- 3.57 (m, 1 H, C5-H), 3.22 -3.41 (m, 2 H, C5-H, C6-H), 3.08 - 3.16 Cm, 1 H, C3-H), 2.90- 3.05 (m, 2 H, C3-H, C6-H), 2.78 (dt, J = 14 Hz, J = 2 Hz,1 H, C17-H), 2.32 (broad s, 2 H, C15-H), 2.16 (broad s, 1 H,C14—H), 1.81 (d, J = 14 Hz, 1 H, C17—H), 1.56 (a, J = 6 Hz,3 H, C18-H).MS (150°C) m/z (relative intensity 10%): 337 (23.6), 336(M, 100), 335 (19.3), 321 (M— CH3, 6.2), 277 (12.5), 249185(10.4), 214 (43.3), 195 (12.9), 170 (11.6), 168 (20.2), 167(12.0), 154 (22.7), 122 (46.6).Elemental analysis: Caic. for C21H4N0:C: 74.97, H: 7.19,N: 8.33. Found: C: 75.17, H: 7.20, N: 8.26.3.19. EXOCATHARANTHINE N-OXIDE 90CH3To a solution of exocatharanthine 89 (500 mg, 1.49mmol) in dry dichioromethane (5 mL) cooled to -30°C underargon was added m-chloroperbenzoic acid (98%) (265 mg, 1.46mmol) in one portion. After stirring for 20 mm at -30°Cthe solvent was evaporated off at 10°C. The residue wasdissolved in ethyl acetate and loaded on a column (3 cm wideand 8 cm long) made of silica gel (40 g) suspended in ethylacetate. Elution successively with ethyl acetate (50 mL),ethyl acetate/methanol (9:1) (200 mL) and ethylacetate/methanol (8:2) (200 mL). Compound 90 (370 mg, 71%)was obtained as a white solid.Physical data of 90:M.p. (vacuo): 170-171°C (CH3OH/Et201:4)0186UV (CHCN) (nm) (log E): 221 (4.57), 274 (3.87,shoulder), 280 (3.89), 2.89 (3.81).IR (KBr) vmax (cm1): 3000 - 3500 (H20 and indole N-Hstretch), 2950 (C-H stretch), 1740 (ester C=0 stretch).‘H NMR (400 MHz, CD3O ) 5(ppm): 7.46 Cd, J = 8 Hz, 1 H, C9-H), 7.29 (d, J = 8 Hz, 1 H, C12-H), 7.10 (t, J = 8 Hz, 1 H,C10-H or C11-H), 7.04 Ct, J = 8 Hz, 1 H, C10-H or C11-H),5.69 (q, J = 7.2 Hz, 1 H, C19—H), 4.30 (s, 1 H, C21—H), 3.92- 4.02 (m, 3 H), 3.67 Cs, 3 H, CH3OCO-), 3.30 - 3.45 (m, 2H), 3.04- 3.20 (m, 2 H), 2.30 - 2.51 Cm, 3 H), 1.75 (d, J =14 Hz, 1 H), 1.68 (a, J = 7.2 Hz, 3 H, C18-H).MS (200°C) m/z (relative intensity 10%): 352 (M, 3.0), 337(12.3), 336 (52.9), 334 (13.3), 277 (12.9), 249 (13.6), 235(11.9), 229 (18.4), 218 (12.2), 214 (33.4), 205 (13.1), 182(17.6), 170 (33.5), 169 (10.3) 168 (17.0), 167 (19.7), 156(10.2), 154 (30.2), 141 (21.5), 139 (64.9), 127 (10.0), 123(11.0), 122 (65.3), 121 (11.3), 113 (11.1), 111 (38.1), 108(10.5), 86 (24.9), 84 (43.1), 77 (11.7), 75 (20.3), 73(26.6), 51 (30.3), 50 (16.9), 49 (78.2), 47 (12.6), 44(100), 43 (27.7), 42 (15.2), 41 (27.1).High resolution MS: Caic. for C21H4N03: 352.1787. Found:352.1779.Elemental analysis: Caic. for C21H4N03, 1.25 H20: C:67.29, H: 7.01, N: 7.47. Found: C: 67.37, H: 7.15, N: 7.19.1873.20. 191 ,20’-ANHYDROVINBLASTINE 91 and epi-19’ ,20’-ANHYDROVINBLASTINE 943A 50 mL three-necked flask, dried in the oven overnightand cooled down under argon, was equipped with a magneticstirrer, a gas bubbler, a thermometer and a glass stopper.To a stirred solution of exocatharanthine 89 (100 mg, 0.297mmol) in dichioromethane (1 mL) was added m-chloroperbenzoicacid (96%) (54 mg, 0.297 mmol), in one portion, at -20 to -30°C under argon. The reaction was exothermic and atemperature increase of 5 to 10°C was observed. Thetemperature was kept at -10 to -15°C for 10 mm and theformation of exocatharanthine N-oxide 90 was monitored byHPLC (reverse phase C18, MeOI-{/H20 23:77 containing 0.3%Et3N, 1.5 mL/min). After cooling to -30 to -40°C vindoline 4(137 mg, 0.300 mmol) was added as a solid. When all thevindoline 4 had dissolved the mixture was cooled to -78°Cand trifluoroacetic anhydride (0.2 mL, 1.42 mmol) was addedHCH3CH3 CH3 ICH3CH391Pc94188in one portion. The reaction was very exothermic and thetemperature rose 10 to 15°C within 10 to 15 sec. After 3 mmthe temperature had returned to -78°C. The reaction wasfollowed by HPLC and kept at -78°C until the coupling wascomplete according to MPLC (that is; the N—oxide 90 peak haddisappeared, and the peaks corresponding to acetylated N-oxide 99 and the iminium salt 100 respectively remained thesame (hight wise) in two successive samples taken 30 mmapart). The reaction time was 3 - 4 hrs and the color of thereaction changed from yellowish to deep red within the firsthour. The bubbler was replaced by a stopper and thethermometer by an adapter fitted with a stopcock. Thereaction vessel was then connected to a vacuum line and thesolvent evaporated for 5 mm at -78°C. The temperature ofthe cooling bath was allowed to rise to -40°C and theevaporation continued until a sticky foam had formed. Normalpressure was restored with argon. The temperature of thecooling bath was lowered to —60°C and dry dichioromethane(1 mL) was added. The temperature of the cooling bath wasallowed to rise to -40°C and was kept at this temperatureuntil the sticky foam had dissolved, which took about 15mm. The reaction vessel was swirled in order to wash downany material sitting on the sides. When all had dissolvedthe reaction vessel was again connected to the vacuum lineand the solvent evaporated of f at -40°C (this secondevaporation is necessary in order to remove the remainingtraces of trifluoroacetic anhydride). When a sticky foam had189once again formed the temperature was allowed to rise to -30°C and the reaction kept at this temperature for anadditional 15 mm. The foam should at this point have quitea dry appearance. Normal pressure was then restored withargon and the cooling bath removed. The foam should slowlycollapse to give very viscous oil, barely stirable. 1 hrafter removing the cooling bath the oil was dissolved in drydichloromethane (1 mL) and cooled to -25°C. Degassedmethanol (15 mL) was added and the temperature was restoredto -25°C. An orange-red solution was obtained. Sodiumborohydride (115 mg, 304 mmol) was added in portions over aperiod of 3 - 4 mm, the temperature being kept at -20 to -25°C during the addition. The reaction was exothermic andtended to foam. The solution became yellow and the pHincreased to greater than 8. The reaction was kept at -20 to-25°C for 5 mm and was then heated up to 8 - 10°C in 5 mm.The reaction mixture was evaporated to dryness on rotaryevaporator at 20- 25°C and the yellow residue obtaineddried further on a vacuum line for 5 - 10 mm to give ayellow foam which subsequently was dissolved in degassed(with argon) ethyl acetate (30 mL). This solution was washedwith water (3 x 10 mL) (the pH of the washings fell from >10to 8 to 7), dried over magnesium sulfate for 5 - 10 mm andfiltered. The filter cake was washed with ethyl acetate(10 mL) and the combined organics evaporated to dryness invacuo at 35°C. Further drying on a vacuum line for 2 to 4hrs gave the crude product (255 mg) as a yellowish foam.190This material was dissolved hexanes/acetone/ triethylamine(10:10:0.6) and loaded onto a flash column (4.5 cm wide and14 cm long) packed with silica gel (230 - 400 mesh). Elutionwith the above solvent system at a flow rate of 5 mL/mingave 19’,20-anhydrovinblastine 91 (165 mg, 70%).Physical data of 91:TLC (silica, hexanes/EtOAc/Et3N10:10:0.6) Rf: 0.26.UV (CH3CN) Xmax (nm) (log E): 212 (4.69), 258 (4.17), 288(shoulder, 4.07).IR (KBr)Vmax (cm-1): 3467 (indole N-H stretch), 2950 (C-Hstretch), 1741 (ester C=0).NMR (400 MHz, CDC13): 8 (ppm): 9.82 (s, 1 H, C16-OH),8.00 (s, 1 H, N-H), 7.50 (d, J = 8 Hz, 1 H, C9,—H or C12,-H), 7.08 - 7.19 (m, 3 H, Cg,-H or C12,-H, C101—H, C111-H),6.59 (s, 1 H, C9—H or C12—H), 6.13 (s, 1 H, C9—H or5.85 (dd, J = 10 Hz, J = 4 Hz, 1 H, C14-H), 5.49 (q, J = 6.5Hz, 1 H, C19,—H), 5.46 (s, 1 H, C17-H), 5.30 (d, J = 10 Hz,1 H, C15-H), 3.83 (s, 3 H, CO2H3), 3.80 (s, 3 H, CO2H3),3.74 (s, 1 H, C2-H), 3.62 (s, 3 H, OCH3), 3.17 - 3.52 (m, 9H), 3.04 (dd, J = 15 Hz, J = 6 Hz, 1 H), 2.77 - 2.89 (m, 2H), 2.73 (s, 3 H, CH3N), 2.67 (s, 1 H, C21—H), 2.39 - 2.49(m, 1 H), 2.29- 2.39 (m, 2 H), 2.14 - 2.22 (m, 1 H), 2.11(s, 3 H, CH3O2), 2.00 - 2.08 (m, 1 H), 1.72 - 1.90 (m, 4H), 1.68 (d, J = 6.5 Hz, 3 H, C181-H), 1.19 — 1.40 (m, 2 H),1.02- 1.12 (m, 1 H), 0.81 (t, J = 8 Hz, 3 H, C18-H).191Physical data of 94:TLC (silica, hexanes/EtOAc/Et3N10:10:0.6) Rf: 0.34.UV (CH3CN)‘Xmax (nm): 214, 261, 291.‘H NMR (400 MHz, CDC13): 6 (ppm): 9.05 Cs, 1 H, C15-OH),7.36 (d, J = 8 Hz, 1 H, C91—H or C12,-H), 7.25 (d, J = 8 Hz,1 H, Cg,—H or C121-H), 7.10 (t, J = 8 Hz, 1 H, C10,—H or28, 2.2),MS (260°C) m/z (relative intensity 5%): 820 (M +806 (M + 14, 1.8), 794 (6.2), 793 (21.1), 792 (M, 42.5),791 (6.5), 790 (9.4), 777 (M - CH3, 2.4), 761 (M - OCH3,1.8), 732 (7.9), 733 (12.8), 732 (7.9), 633 (6.8), 631(5.2), 525 (9.5), 524 (6.6), 469 (10.9), 337 (11.7), 336(24.3), 335 (7.0), 323 (7.1), 283 (5.0), 282 (20.9), 278(5.2), 277 (7.7), 265 (6.1), 263 (6.1), 252 (5.6), 251(5.8), 250 (9.0), 249 (17.6), 214 (14.8), 202 (8.4), 200(5.8), 194 (5.0), 188 (8.6), 185 (5.2), 182 (5.5), 170(7.8), 169 (5.2), 168 (8.2), 167 (5.9), 156 (6.1), 154(8.8), 144 (12.0), 143 (5.0), 138 (7.7), 137 (8.9), 136(55.6), 135 (54.8), 134 (9.7), 130 (6.6), 124 (14.9), 123(18.0), 122 (56.0), 121 (22.6), 120 (7.2), 110 (5.3), 109(6.4), 108 (19.4), 107 (21.1), 106 (11.1), 93 (14.8), 92(7.4), 91 (6.7), 82 (5.4), 79 (7.8), 77 (7.7), 67 (6.0), 55(5.7), 44 (100).Elemental analysis: Caic. for C46H5N08,CH3O : C: 68.62,H: 7.47, N: 6.56. Found: C: 68.45, H: 7.28, N: 6.79.Cab, for C46H5082HS04: C: 52.62, H: 6.50, N: 5.26, S:6.35. Found: C: 52.97, H: 6.33, N: 5.37, 5: 6.14.192C11,—H), 6.99 (t, J = 8 Hz, 1 H, C10,—H or C11,—H), 6.96 (s,1 H, Cg-H or C12-H), 6.02 (s, 1 H, C9-H or C12-H), 5.88 (dd,J = 10 Hz, J = 4 Hz, 1 H, C14-H), 5.51 (q, J 6.5 Hz, 1 H,C191—H), 5.44 (s, 1 H, C17-H), 5.26 (d, J = 10 Hz, 1 H, C15—H), 3.85 Cs, 3 H, CH3O2), 3.78 (s, 3 H, CH3O2), 3.76 Cs, 3H, CH3O), 2.60 (s, 3 H, CH3N), 2.08 Cs, 3 H, CH3O2), 1.66Cd, J = 6.5 Hz, 1 H, C18,-H), 0.57 Ct, J 8 Hz, 3 H, C18-H).MS (260°C) m/z (relative intensity 5%): 806 (M + 14, 1.0),790 (5.5), 777 (M —794 (5.5), 793 (17.9), 792 (M, 33.3),CH3, 1.4), 761 (M(5.3), 633 (7.6),(8.9), 337 (7.5),(6.0), 222 (5.7),(11.9), 187 (6.1),(5.3), 156 (6.3),(48.4), 135 (58.9)(8.7), 122 (40.9),(22.8), 107 (23.4),(10.0), 91- OCH3, 1.3), 733 (8.5), 732 (7.2), 716631 (5.6), 526 (5.2), 525 (14.1), 524336 (13.6), 282 (15.9), 272 (5.4), 249214 (6.2), 202 (5.5), 200 (5.5), 188174 (6.6), 170 (5.7), 168 (5.9), 157144 (10.1), 138 (7.2), 137 (7.7), 136, 134 (7.9), 130 (6.0), 124 (7.4), 123121 (32.4), 120 (13.7), 109 (6.2), 108106 (25.7), 105 (5.1), 93 (16.3), 92(8.6), 82 (6.0), 79 (13.7), 77 (13.6),65 (7.7), 60 (7.3), 58 (5.7),53 (5.2), 51 (7.8), 45 (10.5),High resolution MS: Caic for792.4088.67 (7.7),57 (5.8), 56 (5.2), 55 (8.2),44 (100).C46H55N08: 792.4098. Found:1933.21. 7-HYDROXYCATHARANTHINE N-OXIDE 920To a solution of catharanthine 3 (400 mg, 1.19 mmol) indichioromethane (5 mL) was cooled to -20°C under argon. mChloroperbenzoic acid (420 mg, 2.38 mmol) was added in oneportion and the reaction was stirred for another 20 mm at -20°C before the solvent was evaporated off at 0°C in vacuo.The residue was dissolved in ethyl acetate and loaded on acolumn (3 cm wide and 8 cm long) made of silica gel (40 g)suspended in ethyl acetate. Elution successively with ethylacetate (50 mL), ethyl acetate/methanol (9:1) (200 mL) andethyl acetate/methanol (8:2) (200 mL). The fractionscontaining the desired material was evaporated off at 0°C.Compound 92 (318 mg, 70%) was obtained as a white solid.Physical data of 92.M.p. (in vacuo): 137- 138°C (EtOAc/MeOH/Et207:1:3).UV (MeOH)‘Xmax (nm) (log E): 221 (4.27), 266 (3.75), 281(3.73).IR (KBr) Vmax (cm): 3590 (free 0-H stretch), 3360 (0-Hstretch), 2960 (C-H stretch), 1745 (ester C=0 stretch).HOCO2H3194NMR (400 MHz, CD3OD), 400 MHz) 8(ppm): 7.48 Cd, J = 8 Hz,1 H, C9—H or C12-H), 7.41 (d, J = 8 Hz, 1 H, Cg-H or C12-H),7.38 (t, J = 8 Hz, 1 H, C10-H or C11-H), 7.32 Ct, J = 8 Hz,1 H, C10-H or C11-H), 6.15 Cm, 1 H, C15-H), 5.26 Cs, 1 H,C21-H), 4.39 (broad s, 1 H, C7-OH), 3.80 - 3.92 Cm, 1 H),3.70 Cs, 3 H, CH3OCO), 3.65 (d, J = 12 Hz, 1 H), 3.32 Cm, 1H), 3.02- 3.12 Cm, 2 H), 2.99 (broad s, 1 H), 2.34 - 2.55Cm, 2 H), 2.00— 2.15 Cm, 1 H), 1.75 — 1.95 (m, 2 H), 1.10(t, 3 H, C18-H).MS (200°C) m/z (relative intensity 10% (from m/z 91 - 200relative intensity 20%)): 368 (M, 37.6), 352 (M- 0,3.4), 350 (M- H20, 9.0), 194 (28.8), 188 (27.4), 188(32.0), 175 (22.4), 174 (32.0), 173 (25.3), 171 (24.0), 164(20.8), 163 (21.9), 160 (37.1), 158 (22.1), 157 (32.0), 156(51.2), 155 (20.8), 146 (28.8), 135 (73.6), 133 (48.0), 130(41.6), 129 (25.9), 128 (21.1), 119 (32.0), 117 (25.6), 107(43.2), 105 (54.4), 103 (31.2), 102 (21.0), 93 (25.6), 92(29.6), 91(100).Elemental analysis: : Calc. forC21H4N04,1 H20: C: 65.20,H: 6.37, N: 7.24. Found: C: 65.22, H: 6.92, N: 7.06.1953.22. 7-HYDROXY-EXOCATHARANTHINE N-OXIDE 93CH3To a solution of exocatharanthine 89 (400 mg, 1.19mmol) in dry dichioromethane (5 mL) cooled to -30°C underargon was added nz-chloroperbenzoic acid (96%) (450 mg, 2.49mmol) in one portion. After stirring for 20 mm at -30°Cthe solvent was evaporated off at room temperature. Theresidue was dissolved in ethyl acetate and loaded on acolumn (3 cm wide and 8 cm long) made of silica gel (40 g)suspended in ethyl acetate. Elution successively with ethylacetate (50 mL), ethyl acetate/methanol (9:1) (200 mL) andethyl acetate/methanol (8:2) (200 mL). The fractionscontaining the desired material was evaporated off at 10°Cin the dark. Compound 93 (298 mg, 68%) was obtained as awhite solid.Physical data of 93:M.p. (vacua): 181 -184°C (EtOAc/MeOH/Et204:1:5).UV (MeOH)‘>\max (nm) (log E): 218 (shoulder, 4.26), 223(4.31), 231 (shoulder, 4.20), 270 (3.67).HO 0196IR (KBr) V (cm): 3600 (free 0-H stretch), 3360 (0-Hstretch), 2950 (C-H stretch), 1742 (ester C=0 stretch).NMR (400 MHz, CDC13) 6(ppm): 7.47 (d, J = 8 Hz, 1 H, C9-Hor C12-H), 7.40 (t, J = 8 Hz, 1 H, C10-H or C11-H), 7.39 (d,J = 8 Hz, 1 H, C9-H or C12-H), 7.33 (t, J = 8 Hz, 1 H, C10-Hor C11-H), 5.50 (broad s, 1 H, C19-H), 5.05 (s, 1 H, C21-H),4.21 (broad s, 1 H, C7-OH), 3.76 Cd, J = 11.2 Hz, 2 H), 3.65Cs, 3 H, CH3OCO), 3.26 - 3.35 Cm, 2 H), 3.18 (d, J = 15.2Hz, 1 H), 2.30— 2.63 Cm, 4 H), 2.00 (d, J = 11.2 Hz, 1 H),1.85 Cm, 1 H), 1.61 Cd, J = 7.2 Hz, 3 H, C18—H).MS (200°C) m/z (relative intensity 10%): 368 (M, 2.0), 353(M— CH3, 7.1), 352 (26.2), 351 (M— OH, 4.4), 335 (11.2),293 (15.2), 188 (27.3), 186 (12.4), 164 (11.6), 163 (10.3),160 (33.4), 157 (10.1), 156 (11.7), 149 (11.1), 146 (15.0),145 (10.4), 136 (57.8), 135 (20.7), 134 (13.8), 133 (32.7),132 (12.4), 130 (23.5), 129 (10.1), 123 (12.7), 122 (15.8),121 (18.9), 120 (20.8), 119 (17.5), 117 (13.7), 109 (15.2),108 (66.0), 107 (21.2), 106 (22.2), 105 (35.7).High resolution MS: Caic. for C21H4N04: 368.1736. Found:368.1738.Elemental analysis: Calc. for C21H4N04,1.5 H20: C: 64.12,H: 6.77, N: 6.92. Found: C: 63.79, H: 6.83, N: 7.09.1973.23. 4’ -BENZYL-19’,20’ -ANHYDROVINBLASTINE 10819’,20’-Anhydrovinblastine 94 (535 mg, 0.675 mmcl) wasdissolved in dry benzene (10 mL) under argon. Benzyl bromide(2 mL, 16.8 mmcl) was added and the reaction was stirred inthe dark at room temperature overnight. The next day thesolvent was removed at 40°C in vacuo. Hexanes (10 mL) wasadded and the resulting suspension stirred untilhomogeneous. The resulting precipitate was filtered off andwashed successively with hexanes (10 mL), benzene (2 x 1 mL)and hexanes (10 mL). Compound 108 (513 mg, 79%) was obtainedas white solid.Physical data of 108:UV (CH3CN)‘Xmax (nm): 216, 269, 309.IR (KBr) Vmax (cm): 3360 (indole N-H stretch), 2880 (C-Hstretch), 2820 (C-H stretch), 1715 (ester C=0).Br cl-I3H198NMR (250 MHz, DMSO-d6): 6 (ppm): 173.58, 171.14,157.60, 152.98, 135.71, 133.49, 131.38, 130.46,128.59, 128.31, 127.91, 126.30, 123.83, 122.06,118.90, 117.94, 112.03, 111.35, 94.31, 82.08, 79.47,70.42, 64.00, 62.34, 60.83, 57.80, 56.35, 54.83,52.50, 51.77, 49.86, 49.25, 44.86, 42.50, 42.21,37.00, 30.86, 20.96, 19.38, 13.24, 7.84.3.24. 19’-HYDROXY VINBLASTINE 109 and 19’-HYDROXYLEUROSIDINE 110Compound 108 (513 mg, 0.532 mmol) was dissolved intetrahydrofuran/water (1:1) (10 mL) under argon. Pyridine(94 p1, 1.16 mmol) was added, followed by osmium tetroxide(298 mg, 1.17 mmol). The reaction vessel was covered withaluminium foil. After stirring for 1 hr and 30 mm. Methanol(20 mL) was added and hydrogen sulfide bubbled through the170. 13,129.03,119.58,75. 90,52.94,37.90,H3HCH3.CH3 ncCH3 O2CH3109110199reaction mixture for 15 mm. The solvent was evaporated offat 40°C under reduced pressure and the resultant blackresidue dried in vacuo. The residue was suspended in amixture of dry acetonitrile (7 mL) and dry trie-thylamine(7 mL) and heated to ref lux for 2 hrs and 30 mm. Thereaction mixture was filtered and the black filter cakewashed with methanol (10 mL). Evaporation of thesolvents left a brown residue which was dissolved indichloromethane and loaded on to a column (3 cm wide) ofTLC grade silica gel (17 g) suspended in hexanes andinitially eluted with hexanes/acetone/triethylamine(10:10:0.6) (1.5 1) and then with acetone/triethylamine(20:0.3) (500 mL) at a flow rate of 19 mL/min Compound 109(160 mg, 36%) and compound 110 (60 mg, 14%) were obtained aswhite solids.Physical data of 109:TLC (neutral alumina, CH21/C3O 20 : 1) Rf: 0.40.(silica, acetone/hexanes/Et3N10 : 10 : 0.6) Rf: 0.11.UV (CH3CN) ‘>max (nm) (log E): 214.8 (4.70), 261.4 (4.22),288.6 (4.06), 296.0 (4.05), 306.2 (shoulder) (3.95).IR (KBr) umax (cm): 3469 (OH stretch), 3009 (=CH stretch),2971 (CH stretch), 2586, 2472, 1741 (ester C=O stretch).‘H NMR (400 Mz, CDC13) o(ppm): 8.05 (s, 1 H, N-H), 7.54 Cd,J = 8 Hz, 1 H, C9,-H or C127-H), 7.09 - 7.20 (m, 3 H, C10,-H, C111-H and C9,-H or C121-H), 6.65 Cs, 1 H, C9-H or C12-200H), 6.11 (s, 1 H, C9-H or C12—H), 5.85 (dd, J = 10 Hz, J = 4Hz, 1 H, C14—H), 5.48 Cs, 1 H, C17-H), 5.30 Cd, J = 10 Hz, 1H, C15_H), 3.96 (t, J = 14 Hz, 1 H), 3.81 (s, 6 H, CH3OCO-),3.73 (s, 1 H, C2-H), 3.65 — 3.72 (m, 1 H), 3.62 (s, 3 H,CH3O-), 3.41 (q, J = 6 Hz, 1 H, C191-H (COSY)), 3.25 - 3.40(m, 5 H), 3.13 (broad d, J = 12 Hz, 2 H), 3.06 (d, J = 14Hz, 1 H), 2.80 — 2.90 (m, 2 H), 2.71 Cs, 3 H, CH3N-), 2.67Cs, 1 H), 2.64 (s, 1 H), 2.36 — 2.49 Cm, 2 H), 2.29 (broadd, J = 14 Hz, 1 H), 2.16 - 2.22 (m, 1 H), 2.11 (s, 3 H,CH3O2-), 1.75 - 1.85 (m, 7 H), 1.40 (broad d, J = 14 Hz, 2H), 1.33 (q, J = 7.2 Hz, 2 H, C19-H), 1.12 (d, J = 6 Hz, 3H, C18,—H), 0.82 (t, J = 6 Hz, 3 H, C18—H).MS (260°C) m/z (relative intensity 5%): 841 (6.6), 840(12.0), 826 (M, 8.0), 824 (7.6), 795 (7.0), 768 (6.2), 767(7.8), 766 (6.3), 765 (8.1), 469 (10.5), 371 (14.5), 341(5.2), 327 (5.7), 325 (6.5), 313 (5.1), 311 (7.0), 310(5.1), 309 (5.6), 299 (5.1), 298 (5.0), 297 (8.4), 295(5.4), 284 (6.0), 283 (6.1), 282 (18.3), 272 (6.9), 240(5.3), 222 (9.4), 215 (5.2), 214 (10.9), 212 (5.2), 202(7.8), 200 (8.7), 188 (12.5), 184 (6.3), 183 (5.0), 172(10.0), 171 (16.0), 170 (100), 169 (7.1), 168 (8.7), 157(16.9), 156 (10.5), 154 (7.7), 152 (7.6), 144 (11.3), 138(6.4), 136 (15.2), 135 (48.9), 134 (5.1), 124 (18.9), 123(6.4), 122 (26.7), 121 (20.5), 110 (6.5), 108 (13.9), 107(19.4), 106 (5.7), 94 (5.4), 93 (13.2), 92 (5.6), 91 (5.1),84 (5.1) 83 (5.5), 82 (8.6), 81 (5.8), 70 (7.3), 60 (6.2),20158 (24.4), 45 (8.0), 44 (40.7), 43 (27.7), 42 (12.7), 41(5.4), 30 (6.6), 28 (9.4).High resolution MS: Calc for C46H58N010: 826.4153. Found:826.4171.Elemental analysis: Calc. forC46H58N010,CH21: C: 61.94,H: 6.64, N: 6.15, Cl: 7.79. Found: C: 61.95, H: 6.79, N:6.10, Cl: 7.50.Physical data of 110:PLC (neutral alumina, CH21/C3O 20:1) Rf: 0.2(silica, acetone/hexanes/Et3N10:10:0.6) Rf: 0.00UV (CH3CN) ‘>\max (nm) (log E): 215.0, 263.0, 288.2, 295.6,307.2 (shoulder).IR (KBr) Vmax (cm): 3465 (sharp, 01-I stretch), 3360 (broad,OH stretch), 3010 (CH stretch), 2970 (CH stretch), 2879,2635, 1741 (ester C=O stretch).‘H NMR (400 Mz, CDC13) 8(ppm): 8.08 (s, 1 H, N-H), 7.47 (d,J = 8 Hz, 1 H, C9,-H or C127-H), 7.13- 7.27 (m, 3 H, C107-H, C111—H and C97-II or C121-H), 6.43 (s, 1 H, C9-H or C12-H), 6.13 (s, 1 H, C9-H or C12-H), 5.91 (dd, J 10 Hz, J = 4Hz, 1 H, C14-H), 5.46 (s, 1 H, C17—H), 5.35 (d, J = 10 Hz, 1H, C15-H), 3.81 (s, 3 H, CH3OCO-), 3.81 (s, 3 H, CH3OCO-),3.78 (s, 1 H, C2-H), 3.64— 3.78 (m, 5 H), 3.64 (s, 3 H,CH3O-), 3.27- 3.44 (m, 4 H), 3.15 (broad d, J = 12 Hz, 1H), 2.83 (d, J = 18 Hz, 1 H), 2.75 (s, 3 H, CH3N-), 2.65 (s,1 H), 2.45 - 2.55 (m, 1 H), 2.39 (broad d, J = 14 Hz, 1 H),2.15- 2.25 (m, 1 H), 2.11 Cs, 3 H, CH3O2-), 1.87 - 1.98202H), 1.11(m, 1 H), 1,72 — 1.87 (m, 3 H), 1.54 - 1.65 Cm, 2 H), 1.34(q, J = 7.2 Hz, 2 H, C19—H), 1.16 (d, J = 6 Hz, 3 H, C18,-(d, J = 14 Hz, 1 H), 0.80 (t, J = 6 Hz, 3 H, C18-H).MS (260°C) m/z (relative intensity 5%): 826 (M, 0.6), 768(6.6), 766 (5.0), 607 (5.8), 393 (5.0), 353 (5.0), 351(8.0), 343 (5.6), 327 (5.5), 325 (5.0), 323 (5.3), 313(7.0), 312 (5.2), 311 (9.3), 310 (15.9), 309 (7.0), 308(5.0), 299 (6.5), 298 (5.4), 297 (11.4), 295 (5.5), 284(7.6), 283 (8.0), 282 (18.2), 265 (5.1), 240 (6.8), 226(5.0), 224 (5.1), 222 (9.7), 215 (8.6), 214 (11.9), 212(6.0), 210 (5.5), 202 (9.1), 201 (5.1), 200 (13.4), 198(6.6), 197 (6.4), 196 (5.2), 194 (5.8), 188 (13.6), 186(7.1), 185 (7.2), 184 (9.7), 183 (5.5), 182 (6.1), 180(5.8), 174 (7.3), 172 (10.4), 171 (13.8), 170 (65.3), 169(9.9), 168 (10.2), 167 (5.5), 158 (11.0), 157 (13.0), 156(15.6), 154 (7.5), 152 (7.0), 144 (14.9), 143 (5.2), 138(9.0), 136 (28.0), 135 (47.9), 134 (6.2), 130 (6.9), 124(17.1), 123 (8.1), 122 (26.5), 121 (21.4), 110 (7.2), 108(12.8), 107 (22.7), 106 (10.7), 96 (18.1), 94 (20.4), 93(14.8), 92 (9.8), 91 (16.3), 83 (5.2), 82 (8.5), 81 (6.4),79 (6.2), 77 (6.5), 74 (6.3), 70 (6.0), 60 (6.2), 67 (5.0),65 (5.6), 60 (10. .0), 58 (37.8), 55 (7.3), 50 (5.3), 45(22.9), 44 (100), 43 (79.4), 42 (17.1), 41 (10.64), 30(8.6).High resolution MS: Calc for C46H58N010: 826.4153. Found:826.4180.2033.25. THE BIS-INDOLIC MESYLATE 112 FROM THE MESYLATION OF19’- HYDROXY VINBLASTINE 109H3flcIS CH3Methanesulfonyl chloride (200 p1, 2.58 mmol) was addedto a solution of 109 (160 mg, 0.194 mmcl) in dry pyridine(2 mL). After stirring at room temperature under argon for 3hrs water (20 mL) was added. The precipitate thus formed wasfiltered off and washed with water (5 x 10 mL). Drying invacuo over phosphorous pentoxide gave a purple solid(97 mg). On standing in chloroform this material was partlyconverted to 112 and attempts on flash chromatography onsilica similarly gave decomposition to 112.Physical data of 112:TLC (silica, acetone/hexanes/Et3N10:10:0.6) Rf: 0.54UV (CH3CN)’Xmax (nm): 214, 279, 288, 296.IR (KBr) vmax (cm1): 3360 (indole N—H stretch), 2850 (C-Hstretch), 2750 (C-H stretch), 1730 (0=0), 1750 (ester C=0),1700 (ester 0=0).H204NMR (400 Mz, CDC13) 5(ppm): 8.02 Cs, 1 H, N—H), 7.50 (d,J = 8 Hz, 1 H, Cg,-H or C12,-H), 7.05 - 7.20 Cm, 3 H, C10,-H, C11,—H and Cg,-H or C12,-H), 6.47 (s, 1 H, C9-H or C12-H), 6.10 (s, 1 H, C9-H or C12-H), 5.86 (dd, J = 10 Hz, J = 4Hz, 1 H, C14-H), 5.71 Cs, 1 H, C17-H), 5.18 (d, J = 10 Hz, 1H, C15-H), 3.95 (S, 1 H, C2-H), 3.88 (s, 3 H, CH3OCO—), 3.77(s, 3 H, CH3OCO-), 3.60 (s, 3 H, CH3O-), 3.42 - 3.52 (m, 2H), 3.24 - 3.39 (m, 2 H), 3.01 - 3.22 (m, 5 H), 3.10 (s, 3H, CH3SO-), 2.89- 3.00 (m, 2 H), 2.75 (s, 3 H, CH3N-),2.70- 2.75 (m, 1 H), 2.43 (broad d, J = 16 Hz, 1 H), 2.13 -2.30 Cm, 6 H), 2.10 Cs, 3 H, CH3O2-), 1.90 - 2.00 Cm, 1H),1.67— 1.85 (m, 4 H), 1.39 - 1.48 (m, 1H), 1.12 — 1.23 Cm, 1H), 0.88 Cd, J = 6 Hz, 3 H, C187-H), 0.60 (t, J = 6 Hz, 3 H,C18-H).Elemental analysis: Caic. for C47H60N012S, CH3OCH: C:62.35, H: 6.91, N: 5.82, S: 3.33. Found: C: 62.31, H: 6.80,N: 5.84, S: 3.44.2053.26. GENERAL EXPERIMENTAL CONDITIONS FOR THE OPTIMIZATIONOF THE PHOTOCHEMICAL CYCLIZATION OF THE AMIDES 40 AND41 AND THE THIOAMIDES 51 AND 52abCdeI:.Figure 3-1. Apparatus used in the photolysis.I) This apparatus was used for “high” light intensity andwas wrapped in aluminum foil. For reactions at roomtemperature compressed air was used for cooling.II) This apparatus was used for “low” light intensity. Inboth setups magnetic stirring was used and a mantle used forheating the reaction. (a) Quartz immersion well. (b) Glassfilter (Corex, Vycor or Pyrex). Cc) Low-pressure mercuryHanovia lamp. (d) Stainless steel needle. (e) Condenser. (f)Quartz test tube. (g) Nitrogen inlet.A solution (0.5 mg/mL) of the appropriate amide (40,51) or thioamide (41, 52A, 52B) was purged with nitrogen for5 mm and then irradiated using the appropriate filter.Aliquots were withdrawn every 15 mm, without stopping theii)4cmphotolysis, through the stainless steel needle by using a206positive nitrogen pressure as the driving force. The sampleswere diluted in methanol and analyzed by HPLC.HPLC conditions used for the analysis:Column: Waters Reverse Phase Radial Pak C8 lOp, 8 mm x 10cm.Solvent: CH3N/H20 (60:40).Flow: 1.00 mL/min.Detection: 254 nm.In the photochemical cyclization of the amides 40 and51 the yields of 55 and 56 were determined by HPLC using thepure lactams 55 and 56 as external standards. In case of thethioamides 41, 52A and 52B the pure thiolcatams 61 and 62were used as external standards.3.27. GENERAL EXPERIMENTAL CONDITIONS USED IN THEOPTIMIZATION OF THE YIELD OF EXOCATHARANTHINE 89Method A:Methanol (2.0 mL) was added to 10% palladium on carbon(19.3 mg) under argon. The argon atmosphere was replacedwith hydrogen and the catalyst stirred vigorously for 10mm. The reaction vessel was evacuated and then refilledwith argon. Catharanthine 3 (9.3 mg) in methanol (2.1 mL)was added to the catalyst suspension and the mixture wasstirred under argon at room temperature. The reaction wasmonitored by TLC (silica briefly exposed to ammonia,toluene/ethyl acetate/ MeOH 10:10:2).207Method B:Benzene (10.0 mL) was added to 10% palladium on carbon(10.0 mg) and catharanthine 3 (10.0 mg) under argon. Thereaction vessel was evacuated and then refilled withhydrogen and the mixture was stirred at the appropriatetemperature. The reaction was monitored by TLC (silicabriefly exposed to ammonia, toluene/ethyl acetate/methanol10:10:2).3.28. GENERAL EXPERIMENTAL CONDITIONS USED IN THEINVESTIGATION OF THE FORMATION OF CATHARANTHINEN-OXIDE 10 AND EXOCATHARANTHINE N-OXIDE 90.A dry 50 mL three necked flask was equipped with amagnetic stirrer, a gas bubbler, a thermometer and a glassstopper. catharanthine 3 (200 mg, 0.594 mmol) orexocatharanthine 89 was dissolved in dry solvent (2 mL) togive a clear, often yellowish solution (depending on thequality of the catharanthine 3 or exocatharanthine 89),which was cooled to -10 or -30°C under argon. m-Chloroperbenzoic acid (107 mg, 0.595 mmcl) was added in oneportion (it is crucial for the reproducibility that all theperacid is introduced into the solution and dissolve asquickly as possible). The reaction was exothermic and thetemperature increased 5 - 10°C in less than 30 sec.. Itreturned to the temperature of the cooling bath with in 2mm. The reaction was complete in less than 10 mm. After 10208mm a sample was withdrawn and diluted with methanol forHPLC analysis.HPLC conditions for monitoring and analysis of thecatharanthine N-oxide reaction:Column: Waters reverse phase Radial Pak C18 lOp, 8 mm x10 cm.Solvent: MeOH/H20 (23:77) containing 0.1% Et3N.Flow: 1.5 mL/min.Detection: 254 nm.HPLC conditions for monitoring and analysis of theexocatharanthine N-oxide reaction:Column: Waters Reverse Phase Radial Pak C18 lOu, 8 mm x10 cm.Solvent: MeOH/H20 (23:77) containing 0.3% Et3N.Flow: 1.5 mL/min.Detection: 254 nm.TLC conditions for monitoring the formation of N-oxide:Silica plates using dichioromethane/methanol (10:1) aseluent.3.29. GENERAL EXPERIMENTAL PROCEDURE USED IN THE SYSTEMATICINVESTIGATION OF THE MODIFIED POLONOVSKI REACTIONSetup: A dry 50 mL three-necked flask was equipped witha magnetic stirrer, a gas bubbler, a thermometer and a209rubber septum. Dry glass syringes were used to transfersolvents and trifluoroacetic anhydride and the reaction wasperformed under argon.Steps in the procedure:la) Exocatharanthine 3 (200 mg, 0.594 mmol) was dissolvedin dry solvent (1.0 mL).ib) Exocatharanthine 89 (200 mg, 0.594 mmol) was dissolvedin dry solvent (2.0 mL).2) Cooling to -20 to -30°C.3) 96% m-Chloroperbenzoic acid (112.5 mg, 0.626 mmol) wasadded in one portion. The reaction was exothermic. Thetemperature rises 5 to 10°C less than 30 sec. and thenreturn to the temperature of the cooling bath.4) The temperature was kept at -10 to -15°C for 10 mm.5) The N-oxide formation was monitored on HPLC or TLCusing the condition described in chapter 3.28.6) Cooling to -30 to -40°C.7a) Vindoline 4 (272 mg, 0.596 mmol) was placed in a vial,and dissolved in dry solvent (0.8 mL) and added to thereaction. The vial was rinsed with dry solvent (0.2 mL)which was also added to the reaction.Alternative, and probably better, methods of addingvindoline 4 when hygroscopic solvents are used are:7b) Vindoline 4 (272 mg, 0.596 mmol) was added to thereaction as a solid followed by the dry solvent(1.0 mL).2107c) The N-oxide was generated in 2.0 mL of dry solventinstead of 1.0 mL and vindoline 4 (272 mg, 0.596 mmcl)was added as a solid at this stage. The temperaturewas kept at -20 to -10°C until all the vindoline hasdissolved and a clear solution was obtained.8) Cooling to the desired reaction temperature (-60°C)9) Addition of trifluoroacetic anhydride (0.4 mL. 2.83mmcl) in one portion. The reaction was very exothermic.The temperature rose 10 to 15°C in 10 to 15 sec. After3 mm the temperature had returned to the desiredtemperature (-60°C).10) The reaction was kept at the desired reactiontemperature (-60°C) for 3 hrs.11) The bubbler and the septum are replaced by stoppers andthe thermometer by an adapter. The reaction vessel wasconnected to a vacuum-line and evacuated for 5 mm atthe reaction temperature (-60°C).12) Removal of the cooling bath and evacuation for another15 mm.13) The sticky foam was dissolved in dry degassed (withargon) methanol (5 mL) at room temperature under Ar.The residue dissolved in 3 mm and an orange solutionwas obtained.14) Cooling to -25 to -30°C.15) The temperature was kept at -20 to -25°C during theportion-wise addition of sodium borohydride (225 mg,595 mmol). The reaction was exothermic and tended to211foam. The addition takes 3 - 4 mm. The reaction turnedyellow and pH rose above 8.16) The reaction was kept at -o to -25°C for 5 mm afterthe addition of tlxe sodium borohydride was complete.17) The reaction was heated up by the palm of the hand to 8- 10°C over 5 mm.18) The reaction mixture was transferred to a 100 mL roundbottom flask and the solvent evaporated off on arotary evaporator at 20 - 25°C. This took 15 - 20 mm.19) The yellow residue was dried further on a vacuum-linefor 5 - 10 mm.20) The yellow foam was dissolved in degassed (with argon)ethyl acetate (50 mL).21) Extraction once with water (25 mL). The pH of thewashing was > 10. Then with water (2 x 15 mL). ThepH of the washing fell from about 8 to 7.22) Drying of the organic phase over magnesium sulfate for5 to 10 mm.23) Filtration, washing of the filter cake with ethylacetate (2 x 10 mL).24) Evaporation of the solvent on a rotary evaporator at35°C to dryness.25) Further drying on a vacuum-line for 2 to 4 hrs.26) The crude product (520 - 550 mg) was obtained as afoam, often yellow in color.27) Isolation of 19’,20’-anhydrovinblastine 91 wasperformed as described in section 3.20.2123.30. PREPARATION OF THE SECOND INTERMEDIATE 99.In a dry 50 mL three-necked flask equipped with amagnetic stirrer, a gas bubbler, a thermometer and a glassstopper acetone-d6 (0.8 mL) (from a freshly opened ampule)was added to exocatharanthine N-oxide 90 (60 mg, 0.170 mmol)under argon. The reaction was stirred until a fine suspension was obtained and then cooled to -65°C in an iso—propanol/dry ice cooling bath. Trifluoroacetic anhydride(0.3 mL, 2.1 mmol) was added in one portion and the progressof the reaction was followed by HPLC (using the conditionsdescribed in section 3.28). When the reaction was completethe thermometer was replaced with an adapter fitted with astopcock. Evaporation on a vacuum line for about 30 mm at—50°C removed the majority of the excess of trifluoroaceticanhydride. Normal pressure was restored with argon and thereaction was cooled to -65°C. The reaction was transferredas quickly as possible with a dry syringe to a precooled (-.65°C) NMR tube kept under argon. The syringe was notprecooled in order to prevent introduction of moisture. TheNMR tube was then transferred to the NMR probe which hadalso been precooled to -40°C.Comments:Dichloromethane-d2was also used as solvent. It seemedthat the transformation of the N-oxide to the secondintermediate 99 was faster in acetone than indichioromethane. Compound 99 seemed to be equally stablebelow -40°C in both solvents but more stable in acetone213above —40°C. In one experiment, after completion of the NMRinvestigation, the reaction mixture was transferred back tothe reaction vessel and 1.1 equivalents of vindoline 4 wasadded. Subsequent evaporation and heating gave 19’,20’-anhydrovinbiastine 91 and epi-19’,20’-anhydrovinblastine 94in a 3.5 to 1 ratio as the only dimers formed.2144. REFERENCES1. Noble, R. L.; Beer, C. T.; Cutts, J. H.: Ann. N. Y.Acad. Sd., 1958, 76, 882.2. a) Neuss, N.; Gorman, M.; Svoboda, G. M.; Maciak, G.;Beer, C. T.: J. Am. Chem. Soc., 1959, 81, 4754. b)Johnson, I. S.; Wright, H. F.; Svoboda, M. A.; Svoboda,G. H.: J. Lab. Clin. Med., 1959, 54, 830.3. Svoboda, G. H.: Lloydia, 1961, 24, 173.4. Steam, W. T., in: The Catharanthus Alkaloids; Taylor,W. T., Farnsworth, N. R. (Eds.). New York: MarcelDekker, 1975; Chapter 1.5. Svoboda, G. H., Blake, D. A., in: The CatharanthusAlkaloids; Taylor, W. T., Farnsworth, N. R. (Eds.). NewYork: Marcel Dekker, 1975; Chapter 2.6. Javanovics, K.; Bittner, E.; Dezseri, E.; Eles, J.;Szasz, K.: Ger. Qffen. 2124023, 1971.7. Gerzon, K.; Med. Chein., Ser. Monogr., 1980, 16, 271.8. DeConti, R. C., Creasey, W. A., in: The CatharanthusAlkaloids; Taylor, W. T., Farnsworth, N. R. (Eds.). NewYork: Marcel Dekker, 1975; Chapter 8.9. Carter, S. K.; Livingston, R. B: Cancer Treat. Rep.,1976, 60, 1141.10. Creasey, W. A., in: The Catharanthus Alkaloids; Taylor,W. T., Farnsworth, N. R. (Eds.). New York: MarcelDekker 1975; Chapter 7.11. a) Moncrief, J. W.; Lipscomb, W. N.: J. Am. Chem. Soc.,1965, 87, 4963. b) Moncrief, J. W.; Lipscomb, W. N.:Acta Cryst., 1966, 21, 322.12. Neuss, N.; Gorman, M.; Boaz, H. E.; Cone, N. J.: J. Am.Chem. Soc., 1962, 84, 1509.13. Atta-Ur-Rahman: Pakistan J. Sc md. Res., 1971, 14,487.14. a) Scott, A. I.; Gueritte, F.; Lee, S.-L.: J. Am. Chem.Soc., 1978, 100, 6253. b) Hassam, S. B.; Hutchinson, C. R.: Tetrahedron Lett., 1978, 19, 1681.21515. a) Stuart, K. L.; Kutney, J. P.; Honda, T.; Worth, B.R.: Heterocycles, 1978, 9, 1419. b) Kutney, J. P.;Choi, L. S. L.; Honda, T,; Lewis, N. G.; Sato, T.;Stuart, K. L.; Worth, B. R.: Helv. Chizn. Acta., 1982,65, 2088. c) Stuart, K. L.;Kutney, J. P.; Honda, T.;Worth, B. R.: Heterocycles, 1978, 9, 1391. d) Baxter,R. L.; Hasan, M.; MacKenzie, N. E.; Scott, A. I.: J.Chem. Soc., Chem. Commun., 1982, 791. e) McLauchlan, W.R.; Hasan, M.; Baxter, L.L.; Scott, A. I: Tetrahedron,1983, 39, 3777.16. Langlois, N.; Potier, P: J. Chem. Soc., Chem. Commun.,1979, 582.17. a) Kutney, J. P.; Boulet, C. A.; Choi, L. S.;Golinski, J.; McHugh, M.; Nakano, J.; Nikaido, T;Tsukamoto, H.; Hewitt, G. M.; Suen, R.: Heterocycles,1988, 27, 613. b) Kutney, J. P.; Botta, B.; Boulet, C.A.; Buschi, C. A.; Choi, L. S.; Golinski, J.; Gumulka,M.; Hewitt, G. M.; Lee, G.: Heterocycles, 1988, ZZ,629. c) Kutney, J. P.; Choi, L. S.; Nakano, J.;Nikaido, T; Tsukamoto, H.: Heterocycles, 1988, 21,1837.18. Kutney, J. P.; Beck, J.; Byisma, F.; Cook, J.; Cretney,W. J.; Fuji, K.; Imhof, R.; Treasurywala, A. M.: Helv..Chim. Acta., 1975, 58, 1690.19. Kutney, J. P.; Cook, J.; Treasurywala, A. M.; Clardy,J.; Fayos, J.; Wright, H.: Heterocycles, 1975, 3, 205.20. a) Kutney, J. P.; Ratcliffe, A. H.; Treasurwala, A. M.;Wunderly, S.: Heterocycles, 1975, 3, 639. b) Kutney, 3.P.; Hibino, T; Jahngen, E; Okutani, T; Ratcliffe, A.H.; Treasurwala, A. M.; Wunderly, S.: Helv. Chim.Acta., 1976, 59, 2858.21. a) Potier, P.; Langlois, N.; Langlois, Y.; Gueritte,F.: J. Chem. Soc., Chem. Coinmun., 1975, 670.b) Langlois, N.; Gueritte, F.; Langlois, Y.; Potier,P.: J. Am. Chem. Soc., 1976, 98, 7017.22. Kutney, 3. P.; Balsevich, J.; Bokelman, G. H.; Hibino,T.; Honda, T.; Itoh, I.; Ratcliffe, A. H.; Worth, B.R.: Can. J. Chem., 1978, 56, 62.23. Kutney, 3. P.; Honda, T.; Kazamaier, 3. L.; Lewis, N.L.; Worth, B. R.: Helv. Chim. Acta., 1980, 63, 366.24. Atta-Ur-Rahman; Basha, A.; Ghazala, M.: TetrahedronLett., 1976, 27, 2351.25. Atta-Ur-Rahman; Waheed, N.; Ghazala, M: Z. Naturforshung, 1976, 31b, 264.21626. Kutney, J. K.; Honda, T.; Joshua, A. V.; Lewis, N. G.;Worth, B. R.: Helv. Chim. Acta., 1978, 61, 690.27. Mangeney, P.; Andriamialisoa, R. Z.; Langlois, N.;Langlois, Y.; Potier, P: J. Am. Chem. Soc., 1979, 101,2243.28. Kutney, J. P.; Choi, L. .S. L.; Nakano, J.; Tsukamoto,H.; McHugh, M.; Boulet, C. A.: Heterocycles, 1988, 27,1845.29. Buchi, G.; Kulsa, P.; Ogasawara, K.; Rosati, R. L.: J.Am. Chem. Soc., 1970, 92, 999.30. Sundberg, R. J.; Bloom, J. D.: J. Org. Chem., 1980, 45,3382.31. Marazano, C,; Fourrey, J-L.; Das, B. C.: J. Chem. Soc.,Chem. Commun., 1981, 37.32. Kuehne, M. E.; Bornmann, W. G.; Earley, W. G.; Marko,I.: J. Org. Chem., 1986, 51, 2913.33. Szãntay, C.; Keve, T.; BOlcskei., H.; Acs, T:Tetrahedron Lett., 1983, 24, 5539.34. Raucher, S.; Bray, B. L.; Lawrence, . F.: J. Am. Chem.Soc., 1987, 109, 442.35. a) Imanishi, T.; Shin, H; Yagi, N; Hanaoka, M:Tetrahedron Lett., 1980, 21, 3285. b) Imanishi, T.;Yagi, N; Shin, H; Hanaoka, M.: Chem. Pharm. Bull.,1982, 30, 4052.36. Fowler, F. W.: J. Org. Chem., 1972, 37, 1321.37. Cook, N. C.; Lyons, J. E.: J. Am. Chem. Soc., 1966, 88,3396.38. Roberts, A. M.: U.S. Ptent 2 379 104, 1945. Chem.Abstr., 1945, 39, 462139. Pastushak, N. 0.; Dombrovskii, A. V.; Mukhova, A. N.:Zhurnal Organicheskoi Khimii, 1965, 1, 572.40. Since the brigdes in the Diels-Alder adducts 32 and 33are of the same length, the prefixes endo and exo aredefined as follows:The prefix endo is used when the substituent with thehighest priority points towards the bridge with thehighest priority. The prefix exo is used when thesubstituent with the highest priority points away fromthe bridge with the highest priority.21741. Raucher, S; Bray, B.L.; Lawrence, R. F.: J. Am. Chern.Soc., 1987, 109, 442.42. Olah, G. A.; Narang, S. C.; Gupta, B. G. B.; Maihotra,R.: J. Org. Chem., 1979, 44, 1247.43. Yokoyama, M; Hasegawa, Y; Hatanaka, H; Kawazoe, Y;Imamoto, T: Synthesis, 1984, 827.44. See the appendix for a short introduction to fractionalfactorial design.45. Sundberg, R. J.: Organic Photocheznistry, 1983, 6, 121.46. Honty, K.; Szabo, L.; Kolonits, P., Kaitar, M.,Szantay, Cs.: Exocatharantine, a new isomer ofcatharanthine. The revised structure of exocatharanthine. A lecture on the Annual Meeting of the WorkingCommittee for Alkaloid Chemistry of the HungarianAcademy of Science. Balatonfured 1984, Hungary.47. Kutney, J. P.; Honty, K.: Unpublished results.48. Sundberg, R. J.; Luis, J. G.; Parton, R. L.; Schreiber,S.; Scrinivasan, P. C.; Lamb, P.; Forcier, P.; Bryan,R. F.: J. Org. Chem., 1978, 43, 4859.49. Naruto, S; Yonemitsu, 0: Chem. Chem. Pharm. Bull.,1980, 28, 900.50. Polonovski, M; Polonovski, M: Bull. Soc. Chim. Fr.,1927, 41, 1190.51. a) Lounasmaa, M; Koskinen, A: Heterocycles, 1984, 22,1591. b) Volz, D. H: Kontalcte, 1984, 14.52. a) Huisgen, R.; Kolbeck, W.: Chem. Ber.., 1959, 92,3223. b) Michelot, R: Bull. Soc. him. Fr., 1969, 4377.53. Gartner, H.: Acetylierte Aminoxide. Eine Untersuchungzum Mechanismus der Polonovski-Reaction. UniversitätKarisruhe (TH), 1981. Dissertation.54. Volz, H; Ruchti, L: Liebigs Ann. Chern., 1972, 763, 184.55. Cave, A.; Michelot, R.: C. R. Acad. Sci. Paris, 1967,265, 669.56. a) Cave, A.; Kan-Fan, C.; Potier, P.; Le Men, J.:Tetrahedron, 1967, 23, 4681. b) Ahond, A.; Cave, A.;Kan-Fan, C.; Husson, H. P.; De Rostolan, J Potier, P.:J. Am. Chem. Soc., 1968, 90, 5622.57. Grob, C. A.: Angew. Chern., 1969, 81 , 543.21858. Kutney, J. P.; Joshua, A. V.; Liao, P.-H.; Worth, B.R.: Can. J. Chem., 1977, 55, 3235.59. Langlois, Y.; Langlois, N.; Mangeney, P.; Potier, P.:Tetrahedron Lett., 1976, 44, 3945.60. Honma, Y.; Ban, Y.: Tetrahedron Lett., 1978, 155.61. Honty, K.; Szabo, L.; Kolonits, P.; Katjar, M.;Szantay, C.: “Exocatharanthine, a new isomer ofcatharanthine. The revised structure of isocatharanthine”. A lecture on the Annual Meeting of the WorkingCommittee for Alkaloid Chemistry of the HungarianAcadamy of Science, Balatonfured 1984, Hungary.62. Langlois, Y.; Gueritte, F.; Andriamialisoa, R. Z.;Langlois, N.; Potier, P.; Chiaroni, A.; Riche, C:Tetrahedron , 1976, 32, 945.63. a) Kutney, J. P.; Horinaka, S.; Ward, R. S.; Worth, B.R.: Can. J. Chem., 1980, 58, 1829. b) Kutney, J. P.;Cretney, J. H.; Hall, E. S.; Nelson, V. R: 3. Am. Chem.Soc., 1970, 92, 1704.64. Mak, M.; Tamas, J.; Honty, K., Szabo, L.; Szantay, C.:Biommed. Environ. Mass. Spectrom., 1989, 18, 576.65. Miller, J. P.; Gustowski, G. E.; Poore, G. A.; Boder,G. B.: 3. Med. Chem., 1977, 20, 409.66. Kutney, J. P.; Gregonis, D. E.; Imhof, R.; Itoh, I.;Jahngen, E.; Scott A. I.; Chan, W. K.: J. Am. Chem.Soc., 1975, 97, 5013.67. Kutney, J. P.; Balsevich, J.; Worth, B. R: Can. 3.Chem., 1979, 57, 1682.68 Brown, H. C.; Geoghegan, P. J.: J. Org. Chem., 1970,35, 1844.69 a) Brown, H. C.; Rei, M.-H.; Liu, K.-T.: 3. Am. Chem.Soc., 1970, 92, 1760. b) Brown, H. C.; Rei, M.-H.: 3.Chem. Soc., Chem. Comrnun., 1969, 1296.70 Mangeney, P.; Langlois, Y.: Tetrahedron Lett., 1978,3015.71 Brown, H. C.; Kurek, J. T.; Rei, M.-H.; Thompson, K.L.: 3. Org. Chem., 1985, 50, 1171.72 Wenkert, E; Hagaman, E. W.; Lal, B.; Gutowski, G. E.;Katner, A. S.; Miller, J. C.; Neuss, N.: Helv. Chim.Acta., 1975, 58, 1560.21973 Biemann, K.: L,loydia 1964, 27, 397.74 Vogels Textbook of Quantitative Inorganic Analysis. 4thed. Longman Scientific & Technical, 1978, pp 375 - 377and p 386.75 a) Kempthorne, 0.: The Design and Analysis ofExperiments. New York: John Wiley & Sons, 1952.b) Davies, 0. L. Design and Analysis of IndustrialExperiments. New York: Hafner Publishing Co., 1954.c) Box, G. E. P.; Hunter, W. G.; Hunter, J. S.:Statistics for Experimenters. New York: John Wiley &Sons, 1978. d) Bayne, C. K.; Rubin, I. B.: PracticalExperimental Designs and Optimization Methods forChemists. Weinheim: VCH Publishers, 1986.76 a) Pedersen, B. S.; Scheibye, S.; Lawesson, S.-0.:Bull. Soc. Chim. Belg., 1978, 87, 223. b) Scheibye, S.;Pedersen, B. S.; Lawesson, S.-0.: Bull. Soc. Chim.Beig., 1978, 87, 229.2205. APPENDIX5.1. FACTORIAL DESIGN75Experiments carried out by chemists, physicists, andengineers are in enera1 intended to determine the effectsof one or more factors on the yield or quality of a product,the performance of a machine or measuring instrument, theresistance of a material to chemical attack, and so on. Thetraditional strategy is to investigate one factor whilekeeping all other factors at a constant value and then toselect another factor for the next set of experiments.However, this one-factor-at-a-time strategy has been shownto be inefficient and lacks the ability to detectinteractions among factors. Increased efficiency can begained by studying several factors simultaneously usingfactorial design as experimental strategy.The term factor (or variable) is used in a generalsense to denote any feature of the experimental conditionwhich may be deliberately varied from experiment toexperiment. A factor may be either quantitative (e. g.temperature, stirring rate, concentration etc.) orqualitative (e. g. solvent, the person carrying out theexperiment, magnetic versus mechanical stirring etc.). Thevarious values of a factor examined in a design are known aslevels. The effect of a factor is the change observed in achosen observable (the yield, the product quality, the221.purity, the rate of reaction and so on) produced by a changein the levels of the factor in question.Example 1:Hydrogenation of an olefin required an examination ofthe influence of the temperature and the pressure on theyield of product as well as the interaction betweentemperature and pressure. A factorial design of two factorsFACTOR LEVEL VALUE1 1 atm.AC pressure)2 2 atm.1 10°CB(temperature), 2 40°CScheme 5-1. The factors and their levels used in thefactorial design in scheme 5-2.each on two levels was carried out. Scheme 5-1 shows the twofactors and their levels and scheme 5-2 shows the factorialdesign. The design consists of four experiments and for eachexperiment the design dictates the combination and the222levels of the factors, according to the mathematical theoryon which factorial design is founded. The four experimentsFACTOR INTERACTIONEXP A B AB Yield(atm) (°C) (%)1 1 1 2 502 2 1 1 823 1 2 1 614 2 2 2 92Z1 55.5 66.0 71.5½E2 87.0 76.5 71.0E 31.5 10.5 0.5Scheme 5-2. The factorial design.in scheme 5-2 are performed by varying the pressure and thetemperature as dictated by the design while keeping allother factors constant. The yield in experiment 1 was foundto be 50%. The levels given in the column representing the223interaction, AB, between the two factors relates only to thecalculation of the effect of the interaction between thefactors on the yield, after the experiments have beencarried out.In the ros ½E1 and ½E2 are calculated the averageyields for factor A, factor B and their interaction AB atlevel 1 and level 2 respectively. In the last row, E, iscalculated the effect, the absolute difference between theaverage yield at level 1 and 2, for each of the factors andtheir interaction. For factor A the effect of changing thepressure from 1 atm. (level 1) to 2 atm. (level 2) resultedin an increase of the yield of 31.5% whereas the effect ofchanging factor B from 10°C (level 1) to 40°C (level 2) onlyresults in an increase in the yield of 10.5%. Since theeffect of factor A is larger than the effect of factor B,the variation in the pressure has the largest influence onthe yield. The effect of the interaction of the two factorsat level 1 compared to level 2 is only 0.5% and theinfluence of the interaction on the yield can be said to beinsignificant. At this point it should be stressed that theimportance of the observed effects is not implicit in thedesign, but depends entirely on the interpretation of thechemist. In the example above, the effect of the interactionbetween the two factors is insignificant, but it might be avery important result for the chemist to know that nointeraction exists between the pressure and temperature inthe range examined for the two factors. It is also important224to realize that the conclusions reached on the basis of thedesign can only be expected to have validity as long as theindividual factors are kept within their upper and lowerlevels. In the example above it can not be assumed thatchanging the pressure to 3 atm. will result in a furtherincrease in the yield. On basis of the results obtained inthe design it is possible to construct a new experimentwhich should result in a further improvement of the yieldcompared to the ones obtained in the design. This isaccomplished by combining the best levels for each factor ina new experiment. In the example, above that would cor;espondto an experiment where both factors are kept at level 2. Inthis case that happens to correspond to experiment 4 whichalso gave the highest yield. Mathematically, whencalculating the effects of factor A, in scheme 5-2, theeffects of factor B and the interaction AB are cancelledout;Effect of A = ½[tA(level 1) + B(level 1) + AB(level 2)) -fA(level 2) + B(level 1) + AB(level 1)) + {A(level 1) +B(level 2) + AB(level 1)) - tA(level 2) + B(level 2) +AB(level 2)) = A(level 1) - A(level 2). The same is true forthe effect of factor B and the interaction AB.5.2. FRACTIONAL FACTORIAL DESIGNIn scheme 5-3 is shown a factorial design with threefactors each on two levels. Dividing the factorial design inscheme 5-3 in two and assuming that no interactions exist225between the factors, the fractional factorial design inscheme 5-4 is obtained. The fractional factorial design inscheme 5-4 could also have been constructed from thefactorial design in scheme 5-2 by assigning factor C to thecolumn containing the interaction AB.The calculation of the effects in the fractionalfactorial design in scheme 5-4 is performed in the same wayas described in example 1.For the interpretation of the observed effects to bereliable, it is important that the initial assumption holdstrue, namely that no interactions exist between thedifferent factors. The necessity for this assumption to holdtrue is due to the fact that the effects observed for thedifferent factors are the sum of the effect of the factoritself and the effect of one set of interactions. In table5-1 are shown how the effects of the factors are confoundedwith the effect of the interactions between the factors forthe fractional factorial design in scheme 5-4. It is clearfrom the table that the worst mistake which might be made,due to interactions in a fractional factorial design, isthat a factor might be attributed an importance it does nothave. Normally such misinterpretations are revealed insubsequent experiments. However, by carrying out the otherhalf (experiment 5 - 8) of the factorial design in scheme 5-3, and thereby completing the full factorial design, it ispossible to resolve the factors from the interactions.226FACTORS INTERACTIONSEXP A B C AB AC BC ABC1 1 1 2 1 2 2 12 2 1 1 1 1 2 2.3 1 2 1 1 2 1 24 2 2 2 1 1 1 15 1 1 2 2 1 1 26 2 1 1 2 2 1 17 1 2 1 2 1 2 18 2 2 2 2 2 2 2Scheme 5-3. A three factor two level factorial design.227FACTOREXP A B C1 1 1 22 2 1 13 1 2 14 2 2 2Scheme 5-4. Fractional factorial design.Table 5-1. Confounding of main effect and two-factorinteractions in a fractional factorial design of threefactors each on two levels.observed effect effect + effectfactor A = factor A + interaction BCfactor B = factor B + interaction ACfactor C = factor C + interaction ABThe possibility of expanding a fractional factorialdesign, into the parent factorial design for investigationof the interactions, makes fractional factorial design thepreferred strategy in order to obtain the most informationin as few experiments as possible.

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061752/manifest

Comment

Related Items