PREPARATIVE RADICAL REARRANGEMENT REACTIONS FOR ORGANIC SYNTHESIS
A thesis submitted by JOHN DAVID HARLING
in partial fulfilment of the requirements for the award of
DOCTOR OF PHILOSOPHY OF THE UNIVERSITY OF LONDON
BARTON LABORATORY DEPARTMENT OF CHEMISTRY IMPERIAL COLLEGE OF SCIENCE TECHNOLOGY & MEDICINE LONDON SW7 2AY OCTOBER 1989 ABSTRACT
This thesis is divided into two sections. In the first section a review of free radical rearrangement reactions is presented under the headings of ring opening processes, ring closure processes and group transfer reactions. The review illustrates how free radical rearrangement reactions; once only studied by physical chemists, have during the last few years been exploited by synthetic chemists.
The second section describes the development of a novel, free radical tandem cyclopropyl carbinyl rearrangement-cyclisation strategy for regio- and stereoselective carbon-carbon bond formation. Initial studies in the development of this process are directed towards the synthesis of spiro[ 5 .4 Jdecanes. The extension of this methodology to the synthesis of tricyclo[ 3. 3. 1. 0 ] nonanes and hydrindanes, the latter via two different radical triggers, is then desribed. This is followed by a brief re-examination of the regiochemistry of stereoelectronically controlled radical cyclopropyl carbinyl ring opening reactions. The section is concluded by a formal presentation of the experimental results.
2 CONTENTS
ABSTRACT 2 DEDICATION 4 ACKNOWLEDGEMENTS 5 ABBREVIATIONS 6 A REVIEW OF RADICAL REARRANGEMENTREACTIONS 8 Introduction 8 Ring Opening Reactions 11 Ring Closure Reactions 30 Group Transfer Reactions 33 Review References 39 RESULTS AND DISCUSSION 43 Introduction 43 Regioselective Synthesis of Spiro[4.5 ]Decanes 47 Stereoselective Synthesis of Spiro[4.5 ]Decanes 57 Attempted Synthesis of Spiro-Ethers 67 Synthesis of Hydrinane Derivatives 73 Attempted Synthesis of Bridged Bicyclo Compounds 97 A Reappraisal of Regiochemistry 105 EXPERIMENTAL SECTION 111 Appendix 171 References 176
3 To my parents
4 ACKNOWLEDGEMENTS
First and foremost, I would like to thank my supervisor Willie Motherwell for his encouragement, infectious enthusiasm and constant stream of ideas.
To all my friends, past and present, of the Barton, Whiffen and Perkin labs., I would like to say a big thank you for all their help and advice and for providing a pleasant working atmosphere. Particular thanks go to my contempories, Matt and Graham for their close friendship over the last six years. Thanks are also due to Mike and Botty for their proof reading, often carried out at short notice.
I am grateful to John Bilton and Geoff Tucker for mass spectra, Paul Hammerton and Dick Sheppard for assistance with high field nmr and Ken Jones for microanalyses.
Finally, I would like to thank Quest International for their financial support of this case award and also everyone in the Research Section who made my three month stay there so enjoyable.
5 ABBREVIATIONS
5N a, a'- Azobisisobutyronitrile Ar Aryl b.p. boiling point Bu Butyl t-Bu tertiary-butyl DCM Dichloromethane d doublet dd double doublet ddd double double doublet DIBAL Di-iso-butylaluminium hydride DMAP Dimethylamino pyridine DME Dimethoxy ethane DMSO Dimethyl Sulphoxide Et Ethyl g.c. gas chromatography I.r. Infra-red LDA Lithium diisopropylamide m miltiplet M Methyl m.p melting point nmr nuclear magnetic resonance NOE Nuclear Overhauser Effect Ph Phenyl q quartet s singlet
6 t triplet tert tertiary THF Tetrahydrofuran TBDMS tert-butyl-dimethylsilyl TMS Trimethylsilyl t.I.c. thin layer chromatography
7 A REVIEW OF PREPARATIVE FREE RADICAL REACTIONS
INTRODUCTION
The comparatively recent resurgence in interest in preparative free radical chemistry can be attributed, at least in part, to the development of a wide variety of new and mild procedures for the generation of carbon centred radicals1-4 ( Scheme 1 ).
R—Cl ( Br, I, HgX, CoX )
( X = -SMe, -OPh,— N ^ N ) \ = J
R-Z
Y# = n-Bu3Sn Z = H or Z = H, OH, halogen for thiohydroxamates
Scheme 1 V. Substitution or functional group interconversion reactions, by free radical means, such as the deoxygenation of alcohols, or the radical variant of the Hunsdfecker reaction in which carboxylic acids are transformed into bromides using acylthiohydroxamates2 are already well developed and widely used in synthesis. Similarly, intermolecular addition reactions5 have been extensively studied and subsequently applied to synthesis ( Scheme 2 ).
Bu3SnH CfiHnBr + HnC6 CN ‘CN 95%
Bu3SnH C6HnBr + Hn C6 C02Me C02Me 85%
Intermolecular Addition Reactions
S ch e m ^2 j
Radical rearrangements, although well known and thoroughly investigated from a physical standpoint, have, with the obvious exception of the now ubiquitous free radical intramolecular addition or cyclisation process, not been exploited to nearly the same extent.
This point may be succinctly illustrated by the excellent reviews of Wilt6 and Beckwith and Ingold7, both of which provide exhaustive coverage of radical rearrangements with the Beckwith and Ingold review containing a large amount of useful kinetic data. It is 9 immediately apparent on inspection however, that both reviews contain very few examples of radical rearrangements which have been specifically designed and applied to organic synthesis.
Very recently however, and in particular during the course of this thesis, considerable interest has been aroused by the potential for the incorporation of well understood and predictable radical rearrangements into useful synthetic sequences. Accordingly, this Group Transfer • • A-B-Cn-D w B-Cn-D—A
Ring Opening and Ring Closure
Isomerisation
H H \c—cn—c • ^ c—cn—c• /
Inversion Atom Transfer
Scheme 3 review sets out to update the reviews of Wilt and Beckwith and Ingold and in so doing, to illustrate and emphasise how the foundation work
10 of the Physical chemist is now being successfully exploited by his synthetic organic counterpart.
Beckwith and Ingold have classified radical rearrangements into four categories ( Scheme 3 ) ring opening, ring closures, group transfers and isomerisations which include inversions, rotations and atom transfers. This entirely valid classification is also used in this present review though isomerisations have seen little development over the last decade and will not be discussed further.
11 RING OPENING PROCESSES a) The cvclopropvl carbinvl radical rearrangement
P>— Me
k = 108 M'1 s'1
Cyclopropyl Carbinyl Radical Rearrangement
t
Scheme 4
11 This rearrangement was first observed by Roberts and Mazur8. Ironically, they were interested in the cyclopropyl carbinyl-cyclobutyl carbocation rearrangement but when they tried to prepare cyclopropyl carbinyl chloride by photochlorination of methylcyclopropane, they isolated as the major product, allyl carbinyl chloride ( Scheme 4).
It was some time however before the classical radical mechanism of the rearrangement was established. The kinetics of both the ring opening and ring closure processes have been extensively studied, generally by steady state e.p.r. techniques, even to the present day9'12.
Scheme 5 Because of the different pathway from the cyclopropyl carbinyl- cyclobutyl carbocation rearrangement and the rapid but precisely measured rate constant for ring opening, the cyclopropyl carbinyl radical rearrangement has been widely used as a mechanistic probe13'15. Thus, by way of illustration, Baldwin and co-workers16 have used this rearrangement recently to probe for radical intermediates during the biosynthesis of Penicillin ( Scheme 5). The product obtained from this transformation suggested the intermediacy of a radical species or labile Fe-C intermediate during C-S bond formation.
In principle, cyclopropyl carbinyl radical rearrangements with polysubstituted cyclopropanes may give rise to both regio- and geometrical olefin isomers through homolysis of either of the available cyclopropane bonds. Thus, considerable attention17-21 has been focussed on whether regioselectivity can be obtained in such circumstances. Generally however, unless there is an element of symmetry17 in the molecule, mixtures of compounds are obtained although somewhat surprisingly, the ratios of these compounds can vary quite considerably depending upon the exact nature of the starting material.
For disubstituted cyclopropanes, Davies and Blum18-19 have found that product ratios are not only sensitive to the radical precursor but also whether the two groups are gis. or trans to each other ( Scheme 6 ). Thus, the oi§. substituted cyclopropanes give predominantly products resulting from homolysis of the cyclopropyl bond generating the lower energy secondary radical regardless of reaction conditions.
13 ,B r VV *
cis 83 : 17 (kinetic trapping) trans (kinetic trapping) 34 : 66
trans 92 : 8 (thermodynamic trapping)
Scheme 6 )
However, the trans cyclopropane opens predominantly via homolysis of the cyclopropyl bond that gives the higher energy primary radical under kinetic trapping conditions while under equilibrating thermodynamic conditions the major products arise from homolysis of the other cyclopropyl bond giving the secondary radical. These observations have not yet been satisfactorily explained.
The radical ring opening of gem-difluorocyclopropanes does provide an example of both regioselectivity of bond homolysis and control of olefin geometry20 ( Scheme 7 ). The regioselectivity arises from the bond strengthening of the two cyclopropane bonds attached to the aem-difluoro unit so that cleavage of the other, weaker cyclopropane bond occurs exclusively while the olefin geometry is claimed to be controlled by a conformation preference of the initially formed cyclopropyl carbinyl radical.
14 83% Ph Bu3SnH F F AlBN F F I (E-olefin)
E-olefin Z-olefin
Scheme 7
Oshima22-23 has reported a facile synthesis of vinylcyclopentenes proceeding via a cyclopropyl carbinyl radical rearrangement. Thus, thiyl or stannyl radicals are added to a diene system which in turn sets up a cyclopropyl carbinyl radical system which will rapidly undergo ring opening. This radical can then undergo a 5-exo cyclisation expelling the thiyl radical as shown in Scheme 8 to provide the vinyl cyclopentene. This process only proceeds in good yield when one or more electron withdrawing groups are attached to the cyclopropane.
15 5-exo cyclisation 11
Scheme 8 C02Me C02Me
Regiochemistry in this rearrangement can be conveniently controlled by fusing the cyclopropane to another ring system allowing stereoelectronic factors to control bond homolysis ( vide infra).
SePh
....C02Et
Ph3SnH AIBN
1 r
,C02Et
Scheme 9 Although Clive24 does not discuss stereoelectronic factors in a recent communication on the radical ring opening of cyclopropanes, they are clearly the reason why products proceeding via the higher energy primary radical are exclusively seen ( Scheme 9 ). Thus, the overall process provides a means of attaching alkyl groups to five and six membered rings in a stereoselective manner. In contrast to our previously published work, the radical ring opening involves no subsequent carbon-carbon bond formation.
bl The cvclobutvl carbinvl radical rearrangement
k = 5.6 x 102 s'1 at 25° C
Scheme 10 i
Cyclobutyl carbinyl radicals undergo ring opening several orders of magnitude slower than corresponding cyclopropyl carbinyl radicals (Scheme 10). As a result, intermolecular hydrogen atom abstraction is usually competitive with ring opening, a factor which has seen this rearrangement largely neglected in synthetic chemistry.
A notable exception lies in the recent work of Crimmins25 who has used this rearrangement in the final steps of a concise, regio- and stereoselective synthesis of (+)-Silphinene ( Scheme 11). In order to
17 Bu3SnH, AIBN I ------► added via syringe pump (+) silphinene
(Bu3Sn)2N AIBN hv Bu3SnH AIBN
Scheme 11
circumvent problems with hydrogen atom capture taking place before ring opening, the tri-n-butylstannane was added via a syringe pump or alternatively a Curran type iodine atom transfer was effected using hexabutylditin.
18 c) The epoxv carbinvl radical rearrangement
.0 R R R O
R = phenyl, vinyl R = alkyl, carbonyl
* Schem e 12 V,
This rearrangement has been known for many years and is essentially analogous to the cyclopropyl carbinyl radical rearrangement, though the presence of the oxygen simplifies the problem of regioselectivity. Several early examples27*28 of this rearrangement indicated that alkyl substituted epoxides underwent fragmentation exclusively by carbon- oxygen bond homolysis to give an alkoxy radical. Purely thermodynamic considerations would suggest that the a-alkoxyalkyl radical might be favoured over the alkoxy radical and it is interesting to note that no explanation has been put forward for the apparent kinetic preference. Stogryn29 was able to demonstrate however, that a vinyl or phenyl substituted epoxide opened by carbon-carbon bond homolysis to give vinyl ether products ( Scheme 12 ). This work has recently been reproduced by Murphy30-31, who, however has subsequently shown32 that ketoepoxides undergo carbon-oxygen bond cleavage followed by rapid fragmentation to an acyl radical.
19 Nv J o
OH 65% Addition of substrate Addition of Bu3SnH to Bu3SnH to substrate via
Barton and Motherwells33 have demonstrated that radical induced fragmentation of epoxides may be used as an alternative to the Wharton rearrangement ( Scheme 13 ). While this procedure invariably worked under high stannane concentrations; exposure of the
20 thiocarbonylimidazolide derivative of 4p, 5p-epoxycholestanol to low concentrations of stannane thereby increasing radical life-times, the initially formed alkoxy radical underwent a further apparent 1,2 alkyl shift via a second p-scission mechanism.
O
Schem e^^J
Bowman and Marples34 have subsequently shown that both 4p, 5p and 4a, 5a epoxycholestanol derivatives will, on radical cleavage give 5p-hydroxycholestene. The 4p, 5p-epoxide proceeds to the 5p- alcohol as outlined in Scheme 13 while the 5a-alkoxy radical formed initially on the opening of the 4a, 5a-epoxide must undergo a p- scission followed by a reclosure to give the favoured cis decalin system ( Scheme 14 ).
In studies directed towards a synthesis of Rifamycin S, Murphy35 has developed a tandem epoxide ring opening-cyclisation strategy
21 leading to the synthesis of tetrahydrofuran and pyran derivatives (Scheme 15). This work is conceptually similar to the work presented in this thesis on the cyclopropyl carbinyl radical rearrangement (vide infra). B r
10 : 1 cis : trans
dl The Aziridinvlcarbinvl radical rearrangement
In comparison with the cyclopropyl and epoxy carbinyl radical rearrangements, the aza variant has received scant attention36. In fact, only the Stamm37*38 group have provided a preliminary investigation into this rearrangement and even then, only with the nitrogen atom oriented adjacent to the initially formed radical ( Scheme 16 ) so that nitrogen radicals are precluded. The evidence presently available suggests that regiochemistry is much more dependent on the stabilities of the ring opened radicals ( i.e. bond homolysis occurs virtually
22 exclusively in favour of formation of a tertiary radical over a secondary radical) than is the case with the cyclopropyl carbinyl rearrangement.
Schem e 16
Since the initial radical has only been generated by the reversible addition of stannyl radicals to the carbonyl group of an N-acylaziridine and subsequent work up generates an amide group via loss of the double bond; no information is available regarding selectivity in double bond formation. However, this ring opening process has been extended to provide a subsequent cyclisation in good yield via the equilibria shown in Scheme 17.
23 Uncyclised products (0-17%)
H
Schem e 17 \
e) Rina opening bv B-scission of an alkoxv radical
Tertiary alkoxy radicals readily undergo p-scission processes to yield ketones and carbon centred radicals. These alkoxy radicals are usually generated by photolysis or thermolysis of hypohalites, or the reduction of hydroperoxides by cuprous or ferrous salts6. Fragmentation usually proceeds regioselectively to produce the most stable carbon centred radical (Scheme 18).
24 Schem e 18
Fairly recently however, both the Beckwith39 and Macdonald40 groups have found exceptions to this generality ( Scheme 19 ). While photolysis at low temperature resulted in the expected C 9,10 bond cleavage to provide the more stable secondary radical; thermolysis at elevated temperatures ( 80° C ) resulted in almost exclusively C 1,9 bond cleavage to produce the higher energy primary radical. The tentative and rather complex explanation for these observations
Schem e 19
proposed by Beckwith involves the subtle interplay of the rate constants of all the equilibria involved and how they are affected both by temperature and the concentration of trapping agent. One salient
25 feature is that C 9,10 bond cleavage is reversible whereas C 1,9 cleavage does not appear to be reversible.
A general synthesis of medium ring lactones via the p-scission of alkoxy radicals generated from catacondensed lactols has been developed independently by two groups41*42. The alkoxy radicals are generated by photolysis of the lactols in the presence of mercury (II) oxide / iodine and iodosobenzene diacetate respectively. Both groups have applied this methodology to total synthesis42-43 while the Suarez group have extended the reaction further by incorporation of further oxygen functionality by carrying out the reactions under an oxygen atmosphere44.
Suarez' sythesis of A and A" rings of Limonin
S c h e m ^ 2 ^
26 Schreiber45 has introduced an alternative means of generating the required alkoxyl radical to Suarez and Yamada which involves treating hydroperoxy compounds with iron(ll)sulphate and copper(ll)acetate. This methodology ( Scheme 21 ) which also selectively introduces a trans olefin into the ring by a svn-ft hydrogen elimination from an organocopper(lll) intermediate, has been applied to the synthesis of fourteen membered rings.
t OMe OMe
20 : 1 trans: cis
Scheme 21 |
27 Those p-scissions of alkoxy radicals discussed so far have all employed pre-formed, bicyclic, often steroidal structures. The application of these ring enlargement processes to total synthesis is clearly limited by the requirement of matching substitution patterns of starting materials and products. Baldwin46-47 and co-workers have developed an elegant approach to both and trans cyclononenones and decenones via p-scission of alkoxy radicals which is completely regiospecific, controls olefin geometry and only requires simple
Scheme 22 j
starting materials, making the overall process extremely flexible. This strategy relies on the fact that alkyl radicals will add to a ketone albeit in an unfavourable equilibrium in order to set up the required bicyclic system with the alkoxy radical ( Scheme 22).
28 Both the £is and trans alkylstannylcyclohexanones in Scheme 22 are readily available via standard conjugate addition chemistry of enones. The stannyl radical leaving group completely controls the regiochemistry of bond cleavage as it provides a low energy pathway and its orientation with respect to the alkoxy radicals, controls olefin geometry via a preferred anti-elimination ( Scheme 23). Thus, the trans substituted cyclohexanones give rise to a trans olefin while the ds cyclohexanone gives a sis olefin.
cis cyclohaxanone cis olefin
Schem e 23 2) INTRAMOLECULAR FREE RADICAL ADDITIONS
Scheme 24 \
The intramolecular addition, or cyclisation reaction (Scheme 24 ) has been the most significant development in radical chemistry in recent years. Attributes such as regio- and stereoselectivity and a tolerence for a wide range of functional groups has ensured considerable past, and continuing interest in this process. The recent reviews of Ramaiah48 and Curran49 covering the now substantial volume of literature in this area preclude the necessity of a long exposition on this aspect of free radical rearrangements in this present review.
By way of a summary, the "state of the art" in intramolecular radical addition reactions may be expressed by the two examples in Scheme 25. Curran50 has used a central five membered ring to act as a template for a tri-n-butylstannane mediated tandem radical cyclisation which generates a tricyclopentanoid skeleton in good yield. Beckwith51 has started with a completely acyclic precursor and using the less reactive tri-n-butylgermanne, has managed to obtain reasonable yields of tricyclic compounds arising from three consecutive cyclisations.
30 o
Curran's synthesis of ( 9-epi) silphiperfol-6-ene
Beckwith
Schem e 25
Current research is focussing on two perceived weaknesses of radical cyclisations; namely, that functionality is lost during cyclisation and also that tri-n-butylstannane is both toxic and difficult to remove from product mixtures.
Curran52 has demonstrated that cyclisations involving tertiary iodides may be carried out using hexabutylditin, resulting in the re incorporation of the iodine atom back into the product ( Scheme 26 ). Further, Pattenden53*54 has developed a more flexible cyclisation- functionalisation process using cobalt mediated radical reactions ( Scheme 26 ). Cobalt reagents along with other reagents that will mediate radical reactions, such as tristrimethylsilylsilane55*56 provide alternatives to the use of undesirable tin compounds, though radical
31 chemistry still requires new reagents which are cheap and easy to prepare yet will mediate radical cyclisation reactions in yields comparable to tri-n-butylstannane.
Pattenden
Schem e 26
32 3} REARRANGEMENT BY GROUP TRANSFER
While radical rearrangements involving ring opening or ring closure processes have recently been exploited by the synthetic chemist; rearrangements involving group transfers have largely been neglected.
A notable exception to this is the Beckwith group who have studied and developed several group transfers which could prove synthetically useful. a) The neophvl rearrangement
The ring closure of o-alkenylaryl radicals generally proceeds via the exo mode; however, products arising apparently from cyclisation in the endo mode are more frequently observed than for those cyclisations not involving a-aryl radicals ( Scheme 27 ).
Detailed mechanistic studies by Beckwith57 have revealed that those products arising from endo cyclisation do so, not only from direct cyclisation in the endo mode, but also by a neophyl or 1,2 aryl migration of radicals formed initially by cyclisation in the exo mode ( Scheme27 ). The rate of neophyl rearrangement is slow when compared with the rate of hydrogen atom abstraction from tri-n- butylstannane by exo cyclised products and so substantial amounts of "endo" products are seldom observed as they can generally only arise via the disfavoured direct endo cyclisation process. If however, the experiment is carried out with low stannane concentrations and
33 CHOCHO CHO i •
Lexo
‘■ endo
CHO CHO
Bu3SnH
major product at high major product at low temperature and low temperature and high stannane concentration stannane concentration
Scheme 27
temperatures, then substantial amounts of neophyl rearrangement can take place and endo products can predominate.
34 Acvl and cvano group transfers
Other group transfers developed by Beckwith58*59 include cyano and acyl transfers, effected by consecutive homolytic addition and p-scission ( Scheme 28).
addition
1) P-Scission 2) Bu3SnH C 0 2Me 60%
CHO CHO O
38%
^ S c h e m ^ ^ ^
This process works in reasonable yield for aryl systems because the heat of formation of the aryl-carbon bond is considerably greater than for the alkyl-carbon bond.
35 c) 1.2 Migration of acvloxv groups
The 1,2 migration of acyloxy radicals is a well precedented6*7 radical rearrangement. Giese60 has observed such a rearrangement when the anomeric radical of the acetylated carbohydrate in Scheme 29 is produced. The subsequent 1,2 acyloxy migration provides an efficacious route to either 2-deoxy- or 2-alkyl substituted carbohydrate derivatives.
OAc OAc OAc
Scheme 29
The 1,2 acetoxy migration observed by Kocovsky61 in a radical debromination of an acetylated bromohydrin derivative of cholesterol
36 Mechanism for acyclic 1,2 acyloxy rearrangements
5-centred transition state
Mechanism for allylperoxyl rearrangement
5-centred transition state
Scheme 30
37 has been subsequently used in mechanistic studies by Beckwith62. He concluded that for cyclic systems, the migration proceeded via a three membered cyclic transition state or more probably via a tight anion- radical cation ion pair. This was in contrast to acyclic systems where the 1,2 acyloxy migration was found to proceed via a five membered transition state similar to the transition state found in the alkylperoxy radical rearrangement which has also been the subject of recent mechanistic scrutiny63'65 ( Scheme 30).
38 REFERENCES
1. D. H. R. Barton and S. W. McCombie, J. Chem. Soc. Perkin Trans. 1, 1975, 1574. 2. D. H. R. Barton, D. Crich and G. Kretzschmar, J. Chem. Soc. Perkin Trans. 1986, 1, 39. 3. S. D. Burke, W. F. Fobure and D. M. Armistead, J. Org. Chem., 1982, 46, 3348. 4. B. Gi5se and G. Kretzschmar, Chem. Ber., 1984, 3160. 5. B. Gisse "Radicals in Organic Synthesis : Formation of C-C Bonds", Pergammon Press, Oxford 1986. 6. J. W. Wilt in "Free Radicals", ed. J. K. Kochi, VoM, Chapter 8, Wiley, New York, 1973. 7. A. L. J. Beckwith and K. U. Ingold in "rearragements in Ground and Excited States", VoM, Essay 4, Academic Press, 1980. 8. J. D. Roberts and M. H. Mazur, J. Am. Chem. Soc. 1951, 73, 5931. 9. C. Roberts and J. C. Walton, J. Chem. Soc. Perkin Trans., 1983, 2 879. 10. M. Newcomb and W. G. Williams, Tetrahedron 1985, Lett., 26, 1179. 11. L. Matthem and J. Warkentin, J. Am. Chem. Soc., 1986, 108, 7983. 12. J. R. Bews, C. Glidewell and J. C. Walton, J. Chem. Soc. Perkin Trans.1982, 2 , 1447. 13. P. A. Krusic, J. Am. Chem. Soc., 1977, 99, 250. 14. C. P. Casey and M. C. Cesa, J. Am. Chem. Soc., 1979, 101, 4237. 15. A. J. Castellino and T. C. Bruice, J. Am. Chem. Soc., 1988, 110, 1313. 16. J. E. Baldwin, R. M. Adlington, B. P. Domayne-Hayman and G. Knight, J. Chem. Soc.f Chem. Common.., 1987, 1661.
39 17. M. Campredon, J. M. Kanobus-Kaminski and D. Griller, J. Org. Chem., 1988, 53, 5393. 18. P. Blum and A. G. Davies, J. Chem. Soc., Chem. Commun., 1976, 814. 19. A. G. Davies, M. Pereyre, M. Ratier and R. Sutcliffe, J. Chem. Soc. Perkin Trans. , 1984, 2 1907. 20. T. Morikawa, M. Uejima and Y. Kobayashi, Chem. Lett., 1988, 1407. 21. P. Mariano and E. Bay, J. Org. Chem. 1980, 45, 1763. 22. K. Miura, K. Fugami, K. Oshima and K. Utimoto, Tetrahedron Lett., 1989, 30, 1543. 23. K. Miura, K. Fugami, K. Oshima and K. Utimoto, Tetrahedron Lett., 1988, 29, 5135. 24. D. L. J. Clive and S. Daigneault, J. Chem. Soc., Chem. Commun., 1989, 332. 25. M. T. Crimmins and S. W. Mascurella, Tetrahedron 1987, Lett., 28, 5063. 26. D. P. Curran, M. H. Chen and D. Kim, J. Am. Chem. Soc., 1986, 108, 2489. 27. W. Treibs and G. Pellman Chem. Ber., 1954, 87, 1201. 28. H. C. Brown and M. M. Midland, J. Am. Chem. Soc., 1971, 93, 4078. 29. E. L. Stogryn and M. H. Gianni, Tetrahedron 1970, Lett., 11, 3025. 30. M. Cook, O. Hares, A. Johns and J. A. Murphy and C. W. Patterson, J. Chem. Soc., Chem. Commun., 1986, 1419. 31. A. Johns, J. A. Murphy,C. W. Patterson and N. F. Wooster, J. Chem. Soc., Chem. Commun., 1987, 1238. 32. J. A. Murphy,C. W. Patterson and N. F. Wooster Tetrahedron 29, Lett., 955. 33. D. H. R. Barton, R. S. Hay-Motherwell and W. B. Motherwell, J. 40 Chem. Soc. Perkin Trans. 1981, 2363. 1, 34. W. R. Bowman, B. A. Marples and N. A. Zaidi, Tetrahedron Lett., 1989, 30, 3343. 35. A. Johns and J. A. Murphy, Tetrahedron, 1988, Lett. 29, 837. 36. W. C. Danen, J. Am. Chem. Soc., 1972, 94, 4835. 37. G. Bentz, N. Besbes, A. Laurent and H. Stamm, Tetrahedron Lett., 1987, 28, 2511. 38. J. Werry, H. Stamm, P. Lin, R. Falkenstein, S. Gries and H. Irngarthinger Tetrahedron 1989, 16, , 5015. 39. A. L. J. Beckwith, R. Kazlauskus and M. R. Symer-Lyons, J. Org. Chem., 1983, 48, 4718. 40. T. L. Macdonald and D. E. O'Dell, J. Org. Chem., 1981, 46, 1501. 41. R. Freire, J.J. Marrero, M. S. Rodriguez and E. Suarez, Tetrahedron Lett., 1986, 27, 383. 42. H. Suginome and S. Yamada, Tetrahedron, 1987, 43, 3371. 43. C. G. Francisco, R. Freire, M. S. Rodriguez and E. Suarez, Tetrahedron 1987, Lett., 28, 3397. 44. R. Freire, R. Hernandez, M. S. Rodriguez and E. Suarez, Tetrahedron Lett., 1987, 28, 981. 45. S. L. Schreiber and W-F. Liew, J. Am. Chem. Soc., 1985, 107, 2980. 46. J. E. Baldwin, R. M. Adlington and J. Robertson, J. Chem. Soc., Chem. Commun. 1988, , 1404. 47. J. E. Baldwin, R. A. Adlington and J. Robertson, Tetrahedron, 1989, 45, 909. 48. M. Ramaiah Tetrahedron 1987, 43, 3541. 49. D. P. Curran, Synth., 1988, 489. 50. D. P. Curran and S. C. Kuo, J. Am. Chem. Soc., 1986, 108, 1106. 51. A. L. J. Beckwith, D. H. Roberts, S. C. Sheisser and A. Wallner, 41 Tetrahedron 1986, Lett., 26, 3349. 52. D. P. Curran, M. H. Chen and D. Kim, J. Am. Chem. Soc., 1986, 108, 2489. 53. G. B. Gill, G. Pattenden and S. J. Reynolds,Tetrahedron Lett., 1989,30, 3229. 54. M. J. Begley, H. Bhandal, J. H. Hutchinson and G. Pattenden, Tetrahedron, 1987, Lett. 28, 1317. 55. M. Lesage, C. Chatgilialoglu and D. Griller, Tetrahedron Lett. 1989, 30, 2733. 56. B. Giese and B. Kopper, Tetrahedron 1989, Lett., 30, 681. 57. A. N. Abeywickrema, A. L. J. Beckwith and S. Gerba, J. Org. Chem., 1987, 52, 4072. 58. A. L. J. Beckwith, D. M. O'Shea, S. Gerba and S. Westwood, J. Chem. Soc., Chem. Commun., 1987, 666. 59. A. L. J. Beckwith, D. M. O'Shea and S. Westwood, J. Am. Chem. Soc., 1988, 110, 2565. 60. B. Geise, K. S. Groninger, T. Witzel, H-G. Korth and R. Sustman, Angew. Chem. Int. Ed. Engl., 1987, 233. 26, 61. P. Kocovsky, I. Stary and F. Tureak, Tetrahedron 1986, Lett., 27, 1513. 62. A. L. J. Beckwith, P. J. Duggan, J. Chem. Soc., Chem. Commun., 1988, 1000. 63. A. L. J. Beckwith, A. G. Davies, A. Maccol, M. H. Mruzek, J. Chem. Soc., Chem.. Commun., 1988, 475. 64. N. A. Porter and J. S.Wujek, J. Org. Chem., 1987, 52, 5085. 65. A. L. J. Beckwith, D. M. O'Shea and D. H. Roberts, J. Am. Chem. Soc., 1986, 108, 6408.
42 RESULTS AND DISCUSSION
1. INTRODUCTION As we have demonstrated in the previous section, the development of preparatively useful radical rearrangement reactions is of relatively recent vintage. Indeed, although the cyclopropylcarbinyl radical rearrangement was a well known mechanistic probe1*2, its deliberate incorporation into a useful synthetic sequence had apparently been neglected. We envisaged that this rearrangement could be used in tandem with a 5-exo cyclisation to provide the key step in a novel sequence for regio- and stereoselective carbon-carbon bond formation.
Rates for 25° C , X = radical precursor
Schem e~T^
Scheme 1 shows the desired course of such a reaction. From a kinetic standpoint, the sequence seems entirely favourable. Thus,
43 rapid ring opening of the initially formed cyclopropyl carbinyl radical (1) (k= 1 .3 x1 0 8 M_1 s_1 )3 should be followed by a 5-exo ring closure process (k= 1 x 105 s_1 ) of the resultant radical (2 ) in a controlled cascade which is faster than either the rate of capture of primary or secondary alkyl radicals by tri-n-butyl stannane or the reverse reactions from radical ( 3 ) via ( 2 ) to ( 1 ). A more detailed examination however, appears to reveal a fundamental flaw in the overall process since success requires the regiospecific cleavage of bond (a) ( Scheme 1 ) with concomitant production of the higher energy primary radical in order to set up the 5-exo cyclisation, while the alternative homolysis of bond (b) to give a lower energy secondary radical would provide an unfavourable 4-exo type cyclisation.
In such a hypothetical monocyclic precursor as in Scheme 1 it has been demonstrated4 that products arising from cleavage of both bonds (a) and (b) are observed. Obviously, an additional control element was required in order to assure regioselectivity. Accordingly, our attention was focussed on the construction of bicyclo [ x. 1. 0. ] systems where there is ample precedent5*6 to suggest that regiospecific bond cleavage of the cyclopropane occurs as exemplified in Scheme 2 . Thus, the predominant products in both examples do not arise from thermodynamically controlled bond cleavage of the cyclopropane. In (i), cleavage has procede via a high energy primary radical instead of the much lower energy secondary benzylic radical and similarly in (ii), the cyclopropane has cleaved to to give a primary instead of a secondary radical.
44 ^Schem^^^
In order to explain these observations, Dauben6 has proposed an overlap control mechanism whereby the cyclopropane bond which is cleaved is the one which is most nearly aligned with the radical p. orbital and so provides the best overlap for the developing s-orbital. In current day parlance, these reactions would therefore be considered as subject to stereoelectronic control7. It is clear that in both (i) and (ii), the bond that would need to be cleaved to produce the thermodynamically lowest in energy radical is essentially orthogonal to the radical p-orbital whereas the bond that is actually cleaved is almost parallel to it.
Dauben8 has also observed this stereoelectronically controlled ring opening of cyclopropanes in liquid ammonia reductions of cyclopropylcarbinyl ketones ( Scheme 3 ). Although it is unclear whether ring opening occurs as a one or two electron process; bond
45 Scheme 3 cleavage is still subject to stereoelectronic control as the bond cleaved is the one which has better overlap with the s system of the carbonyl group ( or a ^-orbital containing a radical in the carbon atom of the carbonyl group)
Therefore our proposed reaction, with the kinetics of each step apparently firmly in our favour, appeared to stand a good chance of success, provided we could design a system that was subject to stereoelectronically controlled ring opening of the cyclopropane.
46 2.1 REGIOSELECTIVE SYNTHESIS OF SPIRQr 4. 5 .1DECANES
In the first instance, we elected to study the construction of the spiro-fused [ 4. 5. ] decane system found in the vetiver9 type sesquiterpenes and possessing the always challenging problem of quaternary centre generation10. On the basis of the foregoing analysis, precursors of type ( 4 ) ( Scheme 4 ) were accordingly required.
Scheme 4
A retrosynthetic analysis on ( 4 ) revealed that enone ( 5 ) would be an important intermediate; as the crucial cyclopropanation step might be directly effected via sulphur ylide chemistry or by reduction of the enone to an allylic alcohol followed by a Simmons-Smith reaction.
47 Scheme 5
There are many strategies in the literature for the synthesis of substituted enones11 but the addition of alkyl-metal reagents to the mono-enol ether derivatives of 1, 3-diketones and subsequent acid hydrolysis is a particularly flexible one12.
Thus dimedone, a readily available 1 ,3-diketone was refluxed with trimethylorthoformate in acidified methanol to give its corresponding mono-methyl enol ether13 ( 6 ), ( Scheme 6 ) in almost quantitative yield. For the side chain, which would eventually act as the radical trap, 3-butyn-1-ol was deprotonated with 2 equivalents of n-butyllithium at -30° C and then the dianion quenched with 2 equivalents of chlorotrimethylsilane. An acidic work up hydrolysed the silyl ether to afford 4-trimethylsilylbut-3-yn-1-ol in 82 % yield. The trimethylsilyl group replaced the acidic acetylenic C-H which would have caused problems in anion chemistry and further, there is evidence that the trimethylsilyl group can activate an acetylene towards the capture of nucleophilic alkyl radicals, presumably by back
48 donation of electron density from the acetylene rc-bonds into its vacant d-orbitals.
HC(OMe)3
cat. H2S04 96% MeOH ( 6 )
HO
1) t-BuLi
2) H3Oh 80%
( 9 )
Scheme 6
49 The alcohol was conveniently transformed to its corresponding iodide using either triphenylphosphine and iodine or triphenyl phosphite- methiodide, though the latter procedure generally gave marginally better and more consistent yields ( 80 %). This overall transformation of 3-butyn-1-ol to ( 8 ) (Scheme 6 ) was carried out in essentially an analogous fashion to that subsequently reported by Pattenden14.
The most obvious approach to coupling fragments (6 ) and (8 ) would be via a Grignard reaction12. In principle, magnesium should add electrons very readily to alkyl iodides; however, we were singularly unsuccessful in coaxing iodide (8 ) to react with magnesium despite employing many of the standard stratagems such as baking dry the magnesium turnings, stirring overnight under argon prior to use and the addition of initiators such as 1, 2-dibromoethane and iodine. Rather than attempt to prepare activated magnesium by the Reiche15 procedure for instance, we resorted to the principle of metal- halogen exchange using an alkyllithium. Thus, treatment of iodide (8 ) with 2 equivalents of 1-butyllithium followed by addition to ketone (6 ) and subsequent acidic work-up, afforded two products in roughly equal amount. One of these products was identified as the desired enone ( 9 ), whereas the other was identified as an adduct arising from the addition of i-butyllithium itself to ketone ( 6 ). Surprisingly therefore, 1-butyllithium proved to be a competitive nucleophile with the lithio-derivative of ( 8 ). The formation of this 1-butyl substituted enone was successfully suppressed by the use of a single equivalent of 1-butyllithium and under these conditions ( 9 ) was obtained in a good 79 % yield based on (5).
50 Having synthesized the enone intermediate (9), the next step was to introduce the crucial cyclopropane unit. Our initial attempts to effect this involved using the sulphur ylide chemistry developed by Corey16. When enone ( 9 ) was treated with the sulphur ylide generated from the deprotonation of trimethylsulphoxonium iodide by the dimsyl anion, a single product was obtained. The nmr of this compound revealed no characteristic resonances assignable to cyclopropyl protons and also the complete absence of the trimethylsilyl group.
Scheme 7 )
51 The ylide simply removed the trimethylsilyl group from ( 9 ) to give enone ( 10 ) in a reasonable 72 % yield. Presumably the potentially ambident ylide did not react as expected via carbon but via oxygen to form a strong silicon-oxygen bond with the trimethylsilyl group and thus displace it in an SN2 type process from the acetylene which would then simply pick up a proton on work up.
Having demonstrated that the sulphur ylide was incompatible with our substrate we were forced to turn to the Simmons-Smith cyclopropanation procedure17. Although the Simmons-Smith zinc- methylene carbenoids will insert into enones, the reaction procedes much faster and with superior yields for allylic alcohols as the hydroxyl provides an anchor for the zinc atom, facilitating ready delivery of the :CH2 .
O OH
OH OH
52 In turn, reduction of enones to allylic alcohols, is a well known process in organic synthesis. Diisobutylaluminium hydride18 is the reagent of choice as it does not give any conjugate reduction products. It will however, hydroaluminate acetylenes and so the best yields for the reduction of (9 ) were obtained with one equivalent of diisobutylaluminium hydride at -78° C. This afforded allylic alcohol (11) in 88 % yield ( Scheme 8 ).
For the Simmons-Smith cyclopropanation, a zinc / silver couple was used in preference to the originally reported zinc / copper coupie as Conia19 has demonstrated that the zinc / silver couple generally gives slightly higher and more reproducible yields. Thus, treatment of (11) with five equivalents of zinc / silver couple and diiodomethane resulted in the isolation of a compound with spectral and analytical data consistent with the desired product (12). Most notably, the nmr clearly showed the cyclopropyl resonances at 8 0.27 and 0.48. An important stereochemical consequence of this reaction which will become important later arises from the anchoring effect of the hydroxyl on the zinc carbenoid which results in a ds relationship between the hydroxyl group and the cyclopropane.
All that now remained in order to construct a precursor of the type ( 4 ) in Scheme 4 was to replace the hydroxyl group of compound ( 12 ) with a group from which a radical could be generated at the cyclopropylcarbinyl centre. Thiocarbonylimidazolide derivatives can be formed directly from alcohols under neutral conditions, often in very high yields. Barton and McCombie20 have shown that these thiocarbonylimidazolide derivatives react with tin 53 radicals and collapse as shown in Scheme 9 in an overall chain process; although these initial postulates have recently been brought into question21*22.
In our eagerness to obtain the requisite trigger, alcohol (12 ) was accordingly refluxed in dichloromethane with several equivalents of thiocarbonyldiimidazole reagent 'acquired' from our colleagues in the Whiffen laboratory whereupon work up afforded the corresponding thiocarbonylimidazolide ( 13 ) in essentially quantitative yield . As predicted however, from inspection23 of the thiocarbonyldiimidazole reagent, when (13) was treated with tri-n-butyl stannane in refluxing benzene with AIBN as initiator, only starting material was recovered. When however, we used a sample of thiocarbonyldiimidazole that we
54 had prepared ourselves24 from thiophosgene and imidazole to make ( 13 ), subsequent treatment with tri-n-butyl stannane and AIBN resulted in complete consumption of the starting material. Further investigation has revealed that the thiocarbonyldiimidazole supplied by the Aldrich Chemical Company appears to contain a potent radical inhibitor so that any thiocarbonylimidazolide derivatives prepared with this reagent will not reduce under radical conditions . The thiocarbonyldiimidazole supplied by the Fluka Chemical Company in keeping with the reagent we prepared ourselves, will produce derivatives which smoothly reduce under radical conditions.
The actual radical reaction was performed by dropwise addition of the tri-n-butyl stannane and AIBN to a solution of thiocarbonyl imidazolide ( 13 ) in refluxing benzene over 30 minutes. These reaction conditions provide a low tri-n-butyltin hydride concentration at any given time so that the initially formed radical (14) in Scheme 10 was given the maximum opportunity not only to open to radical (15) but also cyclise to ( 16 ) before hydrogen abstraction from the organostannane takes place. A t.I.c. analysis of the reaction mixture revealed the presence of only one non polar spot apart from the" Sn- S" residue formed from the breakdown of the thiocarbonylimidazolide. A molecular ion of 248 was correct for a reduced compound derived from (13) while the very simple i.r. spectrum showed no acetylenic stretches as would be expected if cyclisation had taken place. The 270 MHz proton nmr spectrum of the isolated product was consistent with the desired pathway of (13) to (17) having been followed as this would lead to the two vinyl silane geometric isomers which were found in a 1 :1 ratio. Also taking into account the presence of three alkenic protons and six allylic protons in the nmr then the structure of 55 SiMe3 ( 13 )
Bl^SnH
l r
SiMe3
SiMe3
( 16 )
Scheme 10
56 these two isomers was confidently assigned as ( 17 ), the desired spiro [ 4. 5. ] decane, formed in 71 % yield. No compound arising from hydrogen abstraction by intermediate (15) was observed. Further, no ring expansion compounds arising from cleavage of the "wrong " bond were observed. Thus the successful conversion of (13) to (17) in good yield substantiated our initial assertions that a stereo- electronically controlled cyclopropylcarbinyl radical rearrangement could be coupled with a 5-exo type cyclisation, to provide a highly regioselective intramolecular carbon-carbon bond forming process.
2.1 STEREOSELECTIVE SYNTHESIS OF SPIRO f 4. 5.1 DECANES
While the above example clearly revealed that the overall sequence was highly regioselective, we also wished to demonstrate that stereospecific construction was also possible, through the control exercised by the hydroxyl directed Simmons-Smith cyclopropanation. For this reason, the alkylated enone (18) in which the methyl group acts as a stereochemical marker was selected for study.
Methylation in the C-6 position was accomplished using lithium diisopropylamide as base and iodomethane as quench. The best yields for this reaction (77 %) were obtained when a reverse quench was employed. A stereoselective reduction of the carbonyl group of (18) was now required.
57 o o
SiMe3
(Scheme^^^
Examination of models of enone ( 18 ) revealed that its preferred conformation was likely to be one in which the C-6 methyl group adopted a pseudo equatorial position with the molecule as a whole having only one axial methyl group. In this conformer, both the C-6 methyl group and the pseudo axial methyl of the C-5 geminal methyl group appeared to shield one face of the cyclohexenone ring (Scheme 12). Thus, a bulky hydride might be expected to approach on a suitable vector from the opposite face giving rise to a "cis " relationship between the C-6 methyl group and the newly formed hydroxyl group. The choice of hydrides was obviously limited by the requirement to avoid conjugate reduction. Thus hindered aluminium and lithium aluminium hydrides were obvious candidates as suitable reducing agents. The results for several attempted stereoselective reductions are summarised in Table 1. The results for the diisobutylaluminium hydride and lithium tri-l-butoxyaluminium hydride were rather disappointing with a 3:1 ratio of the inseparable cis : trans
58 Reducing Temperature Ratio of Isolated Agent (°C) Diastereomers Yield
Diisobutyl- aluminium -78 3 : 1 92% Hydride
0 2:1 95% Lithium-tri- t-butoxy aluminium -30 3:1 92% hydride -40 No Reaction -
Cerium chloride sodium boro- -78 1 :1.25 98% Hydride
L-Selectride -78 >99:1 66 % of isolated alcohol
^ablejl
isomers ( measured by nmr and g.c. ) being the best selectivity observed. The cerium chloride / sodium borohydride reduction system developed by Luche also selectivity reduces enones to allylic alcohols. The chemo- and stereoselectivity of this reagent system has been postulated as arising from the rapid formation of methoxy boranes and not through complexation of the cerium to one face of the carbonyl as stereoselectivity increases on dilution. In the event, this reagent system exhibited almost no stereoselectivity in the reduction of (18).
59 SiMe3
Scheme 12
L-Selectride ( lithium tri-iso-butylborohydride ) is a very selective reducing agent for cyclohexanones but generally reduces enones in a conjugate fashion except where the terminal olefin carbon is sterically hindered in which cases carbonyl reduction can predominate. Enone (18) appeared to possess two structural features which might retard conjugate reduction in favour of carbonyl reduction by L-Selectride . Firstly, the homopropargyl side chain providing steric bulk at the the C-3 olefinic carbon and secondly the axial methyl of the C-5 gem methyl group which would provide unfavourable 1,3-diaxial interactions
60 with any incoming bulky hydride at C-3. Consequently, (18 ) was treated with one equivalent of L-Selectride at -78° C and then worked up by sequential addition of water and then ethanol. The i.r. and proton nmr spectra of the crude product revealed that very little conjugate reduction had taken place (c. 5 %) but that the ratio of ds : trans ally lie alcohols at 4 :1 was only marginally better that previous experiments. Of more interest however, was the observation that the organoborane complex of one of the isomers had not fully hydrolysed on work up. A closer examination of the work up procedure for this experiment revealed that only the major organoborane isomer was hydrolysed to its corresponding allylic alcohol at an appreciable rate by water alone while the other, minor organoborane isomer only hydrolysed on addition of the ethanol. Thus, this kinetic hydrolysis effect was fully capitalised upon to isolate exclusively the major diastereomer produced in this reaction in 66 % yield, simply by omitting ethanol from the work up procedure (Scheme 12).
The Simmons-Smith cyclopropanation of allylic alcohol ( 19 ) (Scheme 14 ) proceded in a good 73 % yield with the zinc again delivering the CH2 : exclusively cis to the hydroxyl thereby establishing the quaternary centre as a single diastereomer. An NOE experiment was carried out in order to prove that the cyclopropanated product did indeed have the anticipated relative stereochemistry of (20) ( Scheme 14). The results of this experiment shown schematically below ( see also appendix for actual results ) proves unequivocally that the compound is ( 20 ) with the C-4 methyl and the C-5 alcohol and cyclopropane all gjs to each other. This also means that the correct stereochemical assignment was made for compound (19).
61 OH OH Me Directed Me Simmons-Smith
Scheme 13
Having obtained compound ( 20 ) in which the methyl group and cyclopropane were cis. to each other, we also wished to obtain the " trans" compound in order to show that we could make either of the two diastereomeric spiro[ 5. 4. Jdecanes relative to the methyl group 62 marker. The obvious way to make the H trans" product (21 ) was to invert the stereochemistry of the hydroxyl group and hence use yet again the powerful tool of the directed Simmons-Smith cyclopropanation.
O
OH Ph
SiMe3
( 22 )
The procedure developed by Mitsunobu51 for the inversion of alcohols using diethylazodicarboxylate, triphenylphosphine and usually benzoic acid was chosen. This reaction allows a pure SN2 displacement to take place with the anion of benzoic acid as the incoming nucleophile and triphenylphosphine oxide as the leaving group. Base hydrolysis of the benzoic ester may then be used to liberate the inverted alcohol. When ( 19 ) was treated with this reagent system the pure inverted benzoic ester ( 21 ) was obtained in 60% yield. As a standard base
63 hydrolysis using potassium hydroxide in water / methanol would undoubtedly remove the trimethylsilyl group from the acetylene a titanium (IV) mediated transesterification52 was employed using titanium isopropoxide in isa-propanol. This gave alcohol ( 22 ) in 69 % yield. (22) had identical i.r and m.s data as (19) while the 270 MHz nmr showed the expected subtle differences most notable of which being the position of the methine proton adjacent to the alcohol group which moved from 8 4.29 to 8 3.72
Cyclopropanation of (22) via a Simmons-Smith reaction afforded ( 23) in an excellent 83% yield ( Scheme 15).
OH OH
SiMe3
Both compounds (20) and (23) reacted smoothly with 1,1' thiocarbonyldiimidazole to give their respective thiocarbonylimidazolide derivatives , ( 24 ) and (25 ) ( Scheme 16 ) in quantitative yield.
64 OH
The actual radical reactions were performed under analogous conditions to those for the conversion of (13) to ( 17 ) i.e. dropwise addition of a solution of tri-n-butylstannane and AIBN to the refluxing benzene solution of the thiocarbonylimidazolide derivative ( Scheme 17 ).
From these radical reactions, excellent yields ( 79 % and 81 % ), of cyclised products were obtained. As in the case of the spiro [5. 4. ] decene (17), both (26) and (27) were obtained as 50 :50 mixtures of geometrical vinyl silane isomers. While both ( 26 ) and ( 27 ) exhibited essentially identical i.r. and m.s characteristics, their proton 270 MHz nmr spectra although identical in the number and
65 distribution of hydrogens, showed considerable differences in 8 positions. Most importantly however, was the fact that the 1H nmr spectra of (26) and (27) were mutually exclusive of each other thus proving that this novel rearrangement-cyclisation strategy is not only highly regioselective but also stereoselective.
S
TMS
S
TMS
Scheme 17
66 3. ATTEMPTED SYNTHESIS OF SPRO-ETHERS
Having established both the regio- and stereospecificity of the rearrangement-cyclisation reaction it was of interest to examine a variety of examples involving different substitution and connectivity patterns, in order to probe the scope and general versatility of the reaction.
X
( 28 ) ( 29 )
C-X = homolisable bond
Scheme 18 )
Those rearrangement-cyclisation reactions already discussed, have involved all carbon systems and so an obvious extension to the reaction would be to attempt to synthesise heterocycles, particularly those with heteroatoms attached to the spiro-carbon. Thus Scheme 18 shows how the inclusion of an oxygen heteroatom might lead to the construction of a spiro-ether compound.
Simpkins25 has recently published a strategy for the synthesis of bicyclic spiro-ethers utilising as the key step a 5 or 6-exo radical ring
67 closure (Scheme 19). Although this methodology provides excellent regio-control there is no scope for its extension to stereocontrolled spiro-ether synthesis. In contrast, a cyclopropylcarbinyl rearrangement- cyclisation strategy should provide the necessary stereo-control element, via the directed Simmons-Smith cyclopropanation and would allow the stereospecific synthesis of spiro-ethers ( Scheme 20).
O O
Simpkins approach to spiro ethers
Scheme 19
Since the immediate precursor to cyclisation (28) is structurally similar to those radical precursors already prepared then a similar synthesis appeared to be possible. Thus, following the procedure of Simpkins25, dimedone was condensed with propargyl alcohol using acid catalysis and azeotropic removal of water to give enone ( 30) ( Scheme 20 ) in 81% yield. The reduction of this enone to provide allylic alcohol (31 ) for cyclopropanation was again accomplished with diisobutylaluminium hydride. Although a t.I.c. analysis suggested that the reaction had proceeded essentially "spot to spot" to a more polar compound, all that was isolated on aqueous work up was the less polar enone ( 33 ). To prevent this elimination taking place on work
68 up, the water was substituted by methanol which gave pure ally lie alcohol ( 31 ) in 98% yield.
O
zinc / copper couple CH2I2 DME
Scheme 20
An attempted cyclopropanation of ( 31 ) using a zinc / silver couple and methylene iodide yielded only enone (33). As the zinc / silver couple is formed in acetic acid solution and it is impossible to
69 remove the last traces of acid, then the failure of this reaction is not surprising considering the obvious acid lability of (31 ). The original zinc / copper couple used by Simmons and Smith can be generated simply by stirring zinc powder and copper( I ) chloride in refluxing ether26 and so this " acid free" couple was used in an attempt to prepare ( 32 ) from the allylic alcohol ( 31 ). It was also found necessary to add several equivalents of dimethoxyethane27 to the reaction in order to remove the Lewis acid, zinc iodide as a 1 : 1 complex. Under these conditions, a modest 39% of (32) was obtained
^ c h e m ^ 2 ^
along with a considerable amount of enone (33). Of some interest is the fact that all the starting material in this reaction was consumed
70 within 15 minutes yet still 39% of (32) was obtained illustrating just how reactive the electron rich double bond is towards the zinc carbenoid.
An attempted preparation of the thiocarbonylimidazolide derivative of (32) proved totally unsuccessful. In fact an almost quantitative yield of the diene ( 34 ) was obtained presumably via the precedented mechanism28 outlined in Scheme 21, which implies that imidazole hydrochloride was functioning as an effective acid catalyst.
S
( 32 )
Scheme 22^
Since alcohol (32 ) was found to be reasonably base stable, those radical precursors formed under basic conditions were considered. The procedure developed by Robbins29 involves reacting an alcohol with phenylchlorothionoformate in the presence of triethylamine as base, to give a thiocarbonate derivative. When the cyclopropylcarbinyl alcohol ( 32 ) was exposed to these conditions however ( Scheme 22), only starting material was recovered probably as a result of too much steric congestion. With this in mind, the preparation of the xanthate derivative was attempted as this only requires the relatively small carbon disulphide to react with the anion
71 of the alcohol, followed then by a methyl iodide quench. Thus, following the procedure of Barton and McCombie20, except that the reaction was carried out at room temperature instead of at reflux, resulted in an 80% yield ( by nmr) of the desired xanthate (35). This xanthate proved very unstable and could not be purified by chromatography on silica, florisil or basic alumina.
OH 1) NaH, Imidiazole 2) CS2 3) Mel _
An attempt was made to carry out the radical reaction on the crude xanthate using several equivalents of tri-n-butyl-stannane. The failure to isolate any identifiable product, let alone one resembling ( 29 ) was not surprising when considering the inherent instability of (35), the presence of sulphur " junk" that would react with tri-n-butyl- stannane and the refluxing toluene temperatures required for the breakdown of a secondary xanthate.
72 At this stage we decided to abandon this approach to spiro ethers; defeated by the acid lability of the radical precursors rather that by a failure of the radical reaction itself.
4. SYNTHESIS OF HYDRINDANE DERIVATIVES
We were also interested in an examination of different connectivity patterns of the radical acceptor chain relative to the cyclopropyl carbinyi trigger which would yield different carbon skeletons from the previously discussed spiro[ 5. 4. ]decanes.
X
Ijchem^^TJ
73 The hydrindane ( bicycle [ 4. 3. 0. Jnonane ) skeleton, which is common to a number of natural products including the Bakenolide9 and Oplapanone9 sesquiterpenes was of particular interest.
Thus, as shown in Scheme 24 , the precursor requires that the side chain is placed on the cyclopropyl carbon adjacent to the radical trigger. This cyclisation is made rather more interesting by the presence of the sp2 centre in the ring opened intermediate (37). It is well known in radical chemistry30 that for cyclisations of this general type ( Scheme 25 ) that if the radical bearing group and the radical trap are gig to each other as in (39 ) then the cyclised product (40 ) will be obtained in good yield.
( 39 ) ( 40 )
trans
( 41 ) ( 42 )
vs.
( 39 ) ( 37 )
Scheme 25 If however, the two groups are trans to each other as for (41 ) then the cyclisation process is kinetically so slow that hydrogen atom abstraction will be faster and only the uncyclised, reduced compound (42) will be obtained. Thus in our proposed transformation of (36 ) to (38 ), the sp2 centre provides a sort of "half way house" between the cis and trans cases. Another consequence of the sp2 centre is that the bond angle between the six membered ring and the radical acceptor chain is now 120° as opposed to 109.5° for a saturated six membered ring ( Scheme 25). Thus, the radical bearing carbon and the acceptor chain are forced further apart making the cyclisation even less favourable and so overall it was of some interest to determine whether the cyclisation was kinetically competitive with hydrogen atom abstraction.
1) KOH ------► 35% 2) Propargyl Bromide
Scheme 26
The synthesis of the required precursor was relatively straightforward. The different position of the side chain was obtained by direct C-alkylation of dimedone in the C-2 position. This reaction poses all the classical problems of C vs. O and mono- vs. dialkylation. The procedure followed31; namely the use of aqueous potassium
75 hydroxide as base with propargyl bromide ( Scheme 26 ), essentially failed to discriminate between mono- and dialkylation though O- alkylation was largely suppressed. The failure to obtain the mono C- alkylated product cleanly was due largely to the fact that it did not crystallise out as it was formed in the reaction mixture. Even so, the
1) DIBAL 2) H30 +
i r
OH o DEAL <------( 46 )
Zn/Ag ch2i2 1 f
76 reaction can be carried out on a large scale and the isolation of (43) is facilitated by its ability to be extracted into aqueous potassium hydroxide and then crystallised out upon acidification of the solution.
The transformation of ( 43 ) to the cyclopropylcarbinyl alcohol (47) follows essentially the same steps as for the transformation of (6 ) to (13) and is summarised in Scheme 27 .
^ c h e n r e ^ j^
The terminal acetylene of (47) would not provide the most efficient trap for the primary nucleophilic radical formed on ring opening of the cyclopropane, especially as the cyclisation is not as favoured as it might be due to the presence of the sp2 centre. We therefore decided
77 to introduce the srongly electron withdrawing -C02Me at the acetylenic terminus. The introduction of this group was initially attempted as a one pot procedure through a selective quench of the dianion ( 48 ), formed by treating ( 47 ) with two equivalents of n-butyllithium, with methylchloroformate. However, an almost equal mixture of the desired product (49 ) along with carbonates (50 ) and (51 ) as well as the starting material ( 47 ) was obtained. In order to circumvent this problem, the hydroxyl group of (47) was protected as its trimethylsilyl ether following the procedure of Rajanbabu32, to give (52 ) which was then deprotonated with one equivalent of n-butyllithium before quenching with methylchloroformate. Acidic work up led to hydrolysis of the silyl-ether to afford (49 ) in 83% yield from (47 ).
OH OTMS
1) n-BuLi 2) MeOC(0)Cl T
Scheme 29
As expected, the formation of the thiocarbonylimidazolide derivative of the neopentylic alcohol ( 49 ) was very slow requiring some thirty six hours heating at reflux to go to completion. When the nmr spectrum of the product was examined however, the product was
78 S OH
[Scheme 30 \
obviously a mixture of two very similar thiocarbonyiimidazolide derivatives ( 53 ) with all resonances doubled up and of similar 5 value, except for the methine proton adjacent to the -OC(S)lm group which resonates at the expected value of 8 4.37 in one compound and at 56.28 in the other.
It was difficult to envisage a mechanism by which epimerisation of this centre could occur and so at this stage we considered the possibility that two distinct rotomeric forms might exist for this obviously hindered thiocarbonyiimidazolide. If this was indeed the case then it should have no effect on the subsequent radical reaction. Thus, when (53 ) was treated with tri-n-butylstannane and AIBN added drop-wise in solution over one hour, two less polar compounds were isolated ( Scheme 31 ).
One of these compounds exhibited a methyl doublet signal at 5 0.86 in the nmr along with an acetylenic stretch at 2235 cm-1 in the i.r. and so was assigned as the ring opened but uncyclised product (55).
79 Addition time of Bu3SnH ( hrs) 1 3 S
70 100
+
f Scheme 31J
The other, major compound was a 50 :50 mixture of two geometrical isomers with a characteristic a-p-unsatu rated ester stretch at 1713 cm-1 in the i.r. as well as two sets of alkenic resonances in the nmr at 5 5.79 and 5 5.42 . This compound was assigned as the desired hydrindane derivative ( 54 ). The fact that the uncyclised compound ( 55 ) was isolated in a relatively substantial amount was a clear indication that hydrogen abstraction was taking place at a competitive rate. The experiment was therefore repeated with an even lower concentration of tri-n-butylstannane obtained by adding it over 3 hours instead of a single hour. Under these conditions, ( 54) was isolated exclusively with no (55) being detected by t.I.c analysis of the reaction mixture.
80 Despite the total suppression of compound ( 55 ) having been effected, the overall yield of ( 54 ) was still relatively low ( 30 % ). Rather unusually, this reaction had turned gradually black on addition of the organostannane thus suggesting that some decomposition was taking place. In this particular case, an alternative pathway could be envisaged involving the addition of the initially formed radical, generated by addition of Bu3Sn # to the sulphur atom of ( 53 ), undergoing a trans 6-exo type cyclisation directly with the activated acetylene instead of the normal fragmentation to the anticipated secondary radical ( Scheme 32 ). The product of such a cyclisation
SnBu3 S
Scheme 32
81 would be very unstable and almost certainly decompose.
However, in light of the low yield of ( 54 ), the uncertainty regarding the exact nature of the mixture of compounds ( 53 ) warranted further attention particularly since the nmr spectrum of this mixture is very sharp implying a rather larger energy barrier between the two supposed rotomers than would be expected.
If (5 3 ) was indeed a mixture of two rotomeric forms then hydrolysis of the thiocarbonylimidazole moeity should lead to recovery of alcohol (49 ), whereas if epimerisation had taken place then we should recover a mixture of two alcohols. We hoped that simple acid hydrolysis would be sufficient to hydrolyse a thiocarbonylimidazolide derivative back to its parent alcohol and in order to establish this we prepared the known thiocarbonylimidazolide derivative of p-cholestanol, dissolved it in THF, added dilute aqueous hydrochloric acid and then warmed it to 50° C for 30 minutes. p-Cholestanol was subsequently recovered in virtually quantitative yield.
In mechanistic terms, we would expect the acid to protonate the imidazole group and then when water attacks the thiocarbonyl group, the tetrahedral intermediate can collapse and kick out imidazole33. There is then a strong thermodynamic driving force to lose carbon oxysulphide to yield the original alcohol ( Scheme 33 ).
82 thiocarbonyldiimidazole Cholestanol
When the above procedure was repeated on ( 53 ), an apparently single compound was obtained in a clean reaction. The absence of an O-H stretch in the i.r. spectrum and the presence of a methine resonance at 5 3.35 precluded any form of alcohol. The mass spectrum helped identify the product as the thiol ( 56 ) ( Scheme 34 ), along with its corresponding disulphide as a single diastereomer, probably with the relative stereochemistry as shown although this has not yet been rigorously proven. The combined yield of thiol and disulphide amounted to 81% which means that both components of the mixture of ( 53 ) hydrolysed to give thiol ( 56 ) with some subsequent air oxidation accounting for the presence of its disulphide.
83 H30+/THF ( 53 )
Scheme 34
While considering the implications of this result we formed the thiocarbonylimidazolide derivative of alcohol ( 47 ) which is the same as (53 ) except that the ester group is not present. This reaction S
( 58 ) proceeded much faster than for ( 53 ) and the nmr showed the presence of essentially one compound (57) although a few percent of the "second thiocarbonylimidazolide" derivative could be detected.
84 When (57) was subjected to the acid hydrolysis it yielded exlcusively the thiol (58 ) in 77% yield. Thus we were left with a considerable problem in explaining why although the thiocarbonylimidazolide derivative of cholestanol hydrolyses to cholestanol as expected; both (57) and the mixture (53) hydrolyse exclusively to thiols (58) and ( 56). We therefore turned to the literature for a possible explanation for these experimental observations.
Rearrangements of xanthates are well precedented in the literature34. When an o-aromatic xanthate is heated to over 100° C, a Newman-Kwart type rearrangement takes place ( Scheme 36 ) with overall transposition between S and O via a postulated 4-centred transition state.
S 1 1
4-centred transition state
t 1) OH' O
The Newman-Kwart Rearrangement
Scheme 36
85 Base hydrolysis then yields a thiophenol. This thermal rearrangment generally only works for aromatic xanthates because if there are any removable p-hydrogens then a Chugaev elimination35 will take place via an E i mechanism.
H O + Ns^ < Chugaev Elimination - X SR
Scheme 37 i
The thione to thiol rearrangement of non-aromatic xanthates can be accomplished with a number of catalysts and at temperatures where Chugaev eliminations do not proceed. Hirsano36 has reported a number of such catalysts, the most effective of which being 4- dimethylaminopyridine-N-oxide. The mechanism he puts forward involves the totally unprecedented step of xanthate acting as a leaving group in an SN2 process with 4-dimethylaminopyridine-N-oxide as the incoming nucleophile. The leaving group then reacts with another molecule of the starting xanthate via the sulphur atom in a second SN2 process. Although this mechanism readily explains how the observed mixture of products might be obtained, the examples chosen R, R' = methyl, ethyl, benzyl ( Scheme 38 ) provide no evidence regarding the stereochemical outcome of the reaction.
86 An alternative mechanism however, more reminiscient of the Newman-Kwart type thermal rearrangements of xanthates and hence more in keeping with the known chemistry of xanthates ( Scheme 39) could be evoked to explain both the course of the reaction as well as the observed mixture of products.
87 ^Schem^^gJ
Thus, the 4-dimethylaminopyridine-N-oxide could attack the thiocarbonyl unit of the xanthate to set up a tetrahedral intermediate. An SNi process proceding via a 4 membered transition state then completes the transposition of S and O. If the 4-dimethylaminopyridine-N-oxide then attacks the carbonyl compound to form another tetrahedral intermediate then this could in principle collapse to release RS_.
88 Intermolecular processes would then equilibrate the products to the statistically observed ratios.
Thus, there is plenty of precedence for the rearrangement of xanthates to the thermodynamically more stable corresponding dithiocarbonate compound. It does not seem unreasonable therefore, to suggest that ( 53 ) is actually a mixture of thiocarbonyl and a rearranged carbonyl compound ( Scheme 40 ).
S O
(53) as a mixture of carbonyl and thiocarbonyl compounde
^Scheme~4^J
This would be consistent with the very different 5 values observered for the methine proton previously mentioned and also the low yield of reduced compounds recovered from the subsequent radical reaction. Acid hydrolysis of the rearranged carbonyl compound would be expected to yield the observed thiol ( 56) via loss of carbon dioxide instead of carbon oxysulphide.
However, this still does not solve the problem of explaining how the unrearranged portion of ( 53 ) and the thiocarbonylimidazolide derivative ( 57 ) which is essentially a single thiocarbonyl compound hydrolyse to the thiols ( 56 ) and ( 58 ). In comparison with the thiocarbonylimidazolide derivative of cholestanol , both ( 53 ) and ( 57)
89 possess a number of structural differences which might account for the difference in reaction pathways and at this stage it is impossible to say anything with any certainty except that the mechanism for the rearrangement does not involve any carbocation or radical intermediates as the cyclopropane remains in tact. This bizarre phenomenon certainly deserves further investigation. S OH
O
Loss of carbon dioxide ' r
A possible, but totally speculative expla nation could involve the acetylene playing a crucial part in the rearrangement to give thiols.
90 Thus, when water attacks the thiocarbonyl unit of the thiocarbonylimidazolide to set up a tetrahedral intermediate the acetylene and imidazole units are brought within close proximity to each other so that there is a rc-stacking interaction between them. This could stabilise the tetrahedral intermediate and retard its collapse surfficciently to allow the sulphur and oxygen atoms to transpose via a 4-centred transition state in an overall thermodynamically favourable process. When the tetrahedral intermediate eventually collapses imidazole is kicked out, with subsequent loss of carbon dioxide yielding the observed thiol ( Scheme 41 ).
As the use of the thiocarbonylimidazolide group as a radical trigger was unsatisfactory for this particular system, we therefore decided to examine some alternative radical triggers. Halides, and in particular iodides and bromides are probably the most extensively used groups from which to generate carbon centred radicals and so alcohol (49 ) was treated with triphenylphosphine dibromide in DMF37, in order to try and prepare the corresponding bromide. Unfortunately this reaction afforded an inseparable mixture of the desired bromide and ring expanded products even when proton scavengers were added to the reaction medium.
Since completion of our work on hydrindane skeletons, Clive38 has shown that cyclopropylcarbinyl alcohols may be transformed into their corresponding phenyl selenides with phenylseleno cyanide in connection with his work on cyclopropylcarbinyl radicals ( Scheme 42 ). This approach was not however, adopted in the present instance; instead we decided to take advantage of the opportunity of exploring an alternative radical trigger by formation of siloxyalkyl radicals.
91 OH SePh PhSeCN ■ C02Et ------► —••C02Et
Scheme 42
In 1973 Motherwell39 described a new reaction for the reductive deoxygenation of ketones to olefins using a zinc and chlorotrimethyl- silane system in an overall 2 electron process ( Scheme 43 ).
tm s tm sn+ tm s O o ' +le' (Zn) +le' (Zn) ► TMS-C1 TMS-C1 Loss of TMS-O-TMS Motherwell's reductive deoxygenation of ketones t
Carbenoid
Scheme 43
Corey40 realised that the initially formed radical species could be utilised as a source of siloxy substituted carbon centred radicals if it was provided with an alternative pathway from accepting a second electron and subsequent carbenoid formation. Thus in providing a
92 judiciously disposed radical trap he was able to effect the 5-exo cyclisation shown in Scheme 44.
O 1) Zn / TMS-CL / THF 78% 2) H30+ C02Me
Scheme 44
The rate of hydrogen atom abstraction from the THF solvent evidently proved slow enough not to compete with cyclisation.The application of this type of methodology to our system moreover, should result in the regiospecific creation of useful silyl enol ether functionality which might be further exploited.
OH
* Scheme 45
The required oxidation of alcohol ( 49 ) to the corrersponding ketone was carried out using Swern41 conditions in 94% yield.
It has been shown in connection with the deoxygenation of ketones that the replacement of zinc by zinc amalgam can facilitate difficult reactions and improve yields by aiding electron transfer and
93 cleaning the surface of the zinc42. Thus, treatment of cyclopropylcarbinyl ketone ( 59 ) with a large excess of chlorotrimethylsilane and zinc amalgam with collidine present as an acid scavenger, in refluxing THF, afforded after an aqueous work up, the ring opened and cyclised ketone (61) in 43% yield. Another mixture of compounds was also isolated, a major component of which we believe to be the ring opened but uncyclised compound (63).
Unlike Corey's cyclisations the initially formed secondary radical rearranges via the ring opening of the cyclopropane to a higher energy and more useful primary radical. Further, the ketone functionality is not lost in the reaction thus providing a convenient handle for further elaboration. In comparison, Corey's cyclisations yield a largely redundant tertiary alcohol.
As mentioned previously and outlined in Scheme 45, the actual product of the radical reaction prior to work up should be the regiospecific silyl-enol ether ( 62 ) which is even more useful for carbon-carbon bond formation than the isloated ketone ( 61 ). Consequently, an attempt was made to carry out a further Mukaiyama43 aldol condensation reaction on the crude product prior to the hydrolytic work up. On completion of the radical reaction, the solution was separated from the zinc using a filter stick and then the THF and residual chlorotrimethylsilane were blown off under a stream of argon. The residue was redissolved in DCM, cooled to -78° C and titanium(IV) chloride and benzaldehyde were added. The only identifiable product from this reaction was the previously isolated bicyclo[ 4. 3. 0. jnonanone ( 61 ). The failure of this reaction probably stems from the fact that these radical reactions were carried
94 TMS o O
.TMS
Scheme 46
95 out on a very small scale for a heterogeneous reaction and so a huge excess of both zinc amalgam and chlorotrimethylsilane was required. Consequently although some collidine was added; the reaction medium had probably become acidic by the end of the reaction and so the silyl-enol ether would already have been hydrolysed. In an attempt to circumvent this problem, chlorotrimethylsilane was substituted by chloro-i-butyldimethylsilane as its corresponding silyl-enol ethers are much les suceptible to hydrolysis. However, despite the fact that the two silicon electrophiles should behave similarly towards the ketone, only starting material was recovered when chloro-1-butyldimethylsilane was used. Perhaps the ketone of (59) is too sterically hindered for the approach of the larger silicon electrophile.
The work carried out so far with zinc amalgam and silicon electrophiles represents only a preliminary investigation and several aspects of the reaction need to be studied in greater depth. For example a) a variety of solvents or solvent mixtures need to be examined with the view to establishing a system with an optimum hydrogen atom abstraction rate so that the cyclisations can procede faster than abstraction b) a variety of silicon electrophiles need to be tested with the aim of trapping sufficiently stable silol-enol ethers for isolation. Scaling up the reactions may also help towards this goal as the ratio of potentially hydrolysable silyl chloride could be reduced with respect to the substrate.
96 4. ATTEMPTED SYNTHESIS OF BRIDGED BICYCLO COMPOUNDS
Although the yield of the hydrindane compound ( 54 ) from the free radical rearrangement-cyclisation reaction was only 30%, we were encouraged by the fact that under suitable, low stannane conditions, the ring opened but uncyclised compound ( 55 ) was completely suppressed in favour of the cyclised hydrindane compound (54).
( 64 ) X = radical precursor cyclisation via Y = activating group diaxial conformation
Scheme 47
This prompted us to consider attempting to apply this methodology to the synthesis of bridged bicyclo compounds as exemplified by Scheme 47. The cyclisation part of this process is disfavoured by the prerequisite that both the ring opened primary radical and the radical trap must adopt a high energy diaxial conformation (64) in order for the cyclisation to proceed.
The synthesis of the radical precursor ( 63 ) requires that the cyclopropane and radical trap be disposed in a relationship; posing
97 something of a synthetic problem. However, the known compound (-) cis-carveol 44 appeared to provide a ready made template from which we could readily access a suitable precursor which satisfied our stereochemical requirements.
MeI Me1 1) lithuim aluminium hydride r V A'"' 2) separation via benzoyl ester u A* A
(R)-Carvone (-) cis-carveol
Scheme 48
Thus ( R )-carvone was reduced with lithium aluminium hydride at -78° C to produce a 17 : 1 ratio of cis. to trans allylic alcohols. Recrystallisation of the 2, 4rdinitro-benzoyl esters and subsequent base hydrolysis yielded exclusively (-) cis-carveol. A Simmons-Smith reaction on (-) cis-carveol ( Scheme 49 ) provided, yet again, a means of
/\ /X
(-) cis-carveol ( 66 ) ^S chem ^^^
98 introducing the cyclopropyl ring with total stereocontrol to afford the cyclopropyl carbinyl alcohol ( 6 6 ) as a single enantiomer, in an excellent 83% yield.
( 66 ) Wadsworth-Emmons 11
61% Scheme 50 )
A C02Me
( 68 )
As with the hydrindane cyclisations, we felt that activation of the radical trap towards the nucleophilic alkyl radical was essential. Accordingly (6 6 ) was ozonised45 to afford the the hydroxyketone (67) ( Scheme 50 ) in 94% yield. This compound existed totally as the hydroxyketone with no observable lactolisation. Treatment of (67) with the sodium ylide of triethylphosphonoacetate, under standard Wadsworth-Emmons46 conditions provided compound 6 ( 8 ) in 61% yield as a 5 :1 mixture of E and Z isomers. Since it was anticipated that
99 this olefin functionality would be lost in the ensuing radical reaction; no attempt was made to separate the geometrical isomers.
Formation of the thiocarbonylimidazolide derivative ( 69 ) proceeded in essentially quantitative yield after an 18 hour reflux period in DCM ( Scheme 51). The radical reaction was carried out initially by adding 1.5 equivalents of tri-n-butylstannane over three hours via a syringe pump. A t.I.c analysis of the reaction mixture revealed two major compounds. The more polar of these compounds exhibited i.r. stretches at 1713 and 1639 cm-1 characteristic of the a, p-unsaturated ester radical trap built into the radical precursor, indicating that this was not the bridged bicyclic compound. The proton nmr of this compound possessed two alkenic protons at 8 5.02 and 5.64, and two allylic methyl groups at 8 2.16 and 2.10. A molecular ion of 222 confirmed that this compound was simply a reduction product of the thiocarbonylimidazolide ( 69 ). This compound was therefore assigned as the ring opened but uncyclised compound (70) (Scheme 51).
The other major, less polar compound, formed in a roughly equivalent amount, exhibited only a saturated ester stretch at 1729 crrr 1 in the i.r. and so we were initially encouraged to believe that this compound might be the desired bicyclo compound. However, the nmr revealed that we had a mixture of at least three compounds, probably diastereomers. The most striking feature of this spectrum was the complete lack of olefinic protons. The only other useful information that could be obtained from the nmr was the presence of three methyl groups in the molecule. One of these was associated with the ester group while the others were attached to tertiary and secondary carbons respectively. The mass spectrum gave a molecular ion of 222 100 C02Me C02Me
Bu3SnH / AIBN added via syringe pump
other + ( 70 ) + products
C02Me Addition time of Bu3SnH (hrs) 3 1
6 2
Scheme 51 and an accurate mass measurement provided a molecular formula of C14H22O2 . Thus, in the absence of any double bonds in the molecule ( substantiated by a 13C nmr spectrum ) we were forced to invoke a tricyclic compound in order to satisfy the molecular formula. A g.c analysis of the crude reaction mixture revealed that a substantial
101 number ( > 2 0 ) of other products had been formed although none in more than 3% and none of them were isolable. \ A re-examination of our proposed cyclisation revealed that if the desired cyclisation had taken place then there was the possibility of a further 5-exo cyclisation taking place between the radical adjacent to the ester group and the olefin formed on opening of the cyclopropane. This would lead to a tricyclic compound , probably diastereomeric at the centres where the methyl and ester groups are attached to the carbon skeleton. This noradamantane 47 structure appears to fit all the spectroscopic data we have so far ( Scheme 51 ).
It is important to note that for the second cyclisation to proceed, the first cyclisation must take place with a particular orientation ( Scheme 52) of the olefin radical trap so that the subsequent radical is suitably disposed to add to the second olefin. Since we have failed to isolate any bicyclic compounds then we must assume that the first cyclisation proceeds exclusively via this conformation though models fail to provide any obvious steric arguments which might be used to explain this phenomenon.
102 Major E isomer
C02Me
1) Ring opening 2) adoption of diaxial conformation H Me Me
5-exo cyclisation
Me
noradamantane skeleton
Scheme 52
103 When the radical reaction was repeated with a slower addition of stannane (6 hours ), the ratio of tricyclic to monocyclic compounds (71 ) and (70) altered to 2:1. Efforts to separate the diasteromers of nordamantanes by preparative capillary g.c are currently under way so that a more extensive characterisation may be carried out. This separation however, is proving extremely difficult.
104 5. A REAPPRAISAL OF THE REGIOCHEMISTRY OF THE CYCLOPROPYL CARBINYL RADICAL REARRANGEMENT IN B1CYCLIC SYSTEMS
During the development of our tandem radical cyclopropyl carbinyl rearrangement and 5-exo cyclisation strategy we naturally searched the literature for any work involving cyclopropyl carbinyl radicals, especially those involving bicyclic systems.
Freeman 48 has examined the addition of thiyl radicals to bicyclo [3. 1.0. ]hexylidene systems. A number of products were obtained ( Scheme 53) but the only ring opened ones arose from
predominantly trans product
^SchemeT^^ stereoelectronically controlled reactions i.e. via the higher energy primary methyl radicals and not through the lower energy ring
105 expanded secondary radicals. These results are in complete accord with our own observations.
By way of contrast however, Davies 49 included the addition of carbon tetrachloride to sabinene under free radical conditions in a paper on radical rearrangements of strained systems. As can be seen in Scheme 54, only the ring expanded product was isolated through
Scheme 54 )
106 the cleavage of bond ( a ), a result which appears to be in direct contradiction of our stereoelectronic arguments and observations as this bond is essentially orthogonal to the ^-orbital of the radical. Our initial sceptism at this result disappeared when we repeated the experiment, improved the yield, and found that the structural assignment by Davies was indeed correct.
Realising that abstraction of chlorine atoms from carbon tetrachloride to quench the sabinene radical would be very slow even when carbon tetrachloride was used as a solvent we examined some alternative free radical additions to sabinene where the radical would be quenched at a significantly faster rate. Thus, when sabinene was heated with thiophenol and AIBN in benzene, two products were obtained. One of these was the addition compound ( 78 ) where the cyclopropane is still intact, illustrating just how fast hydrogen atom abstraction from thiophenol is; while the other product was the five membered ring compound (77 ) which is the expected product of a stereoelectronically assisted process. Similarly, reaction of sabinene with benzoyl peroxide and bromotrichloromethane which features the same chain carrying trichloromethyl radical as in the case of carbon tetrachloride but differs only in the faster rate of abstraction of the weaker carbon-bromine bond, yielded the five membered ring compound ( 76 ) as the exclusive product.
Thus it would appear that providing atom capture from a reagent is sufficiently slow then the radical from stereoelectronically controlled ring opening can reclose and open again and eventually give the thermodynamic product. We believe however, that the observed products should not simply be dismissed on a straightforward
107 kinetic basis as the alignment of the radical ja-orbital and bond ( a ) of the cyclopropane ( see Scheme 55) which needs to cleave to produce the thermodynamic product are still essentially orthogonal to each other.
7t-orbital
Stereoelectronically controlled ring opening taking Walsh orbitals into account
Scheme 55
A more refined explanation is that the ring opening process to afford the thermodynamic six-membered ring takes place via a higher energy
108 puckered conformation of the molecule ( Scheme 55) where the radical p-orbital is now aligned with bond ( a ) so that stereoelectronics are actually still controlling the opening process. The amount of pucker in the molecule required to provide sufficient overlap between the p- orbital and bond ( a ) is reduced when the Walsh type orbitals of the cyclopropane are taken into account ( Scheme 55).
Scheme 56 )
Paquette 50 has added carbon tetrachloride to the bicyclo[ 3.1. 0 ] hexane shown in Scheme 56. Here, stereoelectronically controlled opening of a non-puckered conformer initially takes place but because the fl£m-methyl group makes this radical a very low energy tertiary radical it is the five membered ring compound that is quenched by a chlorine atom and not the thermodynamically higher in energy and non-kinetic six membered ring compound.
109 It is amusing in retrospect to reconsider the situation which pertains in our initial study of the tandem ring opening-cyclisation process wherein the development of the ^.-orbital containing the unpaired electron from the 2 C-0 bond of the thiocarbonylimidazolide will initially result in a conformation which is apparently more suitable for stereoelectronically assisted cleavage of the inner bond ( a ).
S
Radical in familiar p-orbital favours cleavage of external internal cylopropane cylopropane bond bond
Scheme 57
Evidently in these reactions, the adoption of a conformation which favours the observed ring opening mode is extremely rapid. In the final analysis, there is no substitute for experimentation. 110 EXPERIMENTAL SECTION
Melting points were determined on a Kofler-hot stage and are uncorrected. Infrared spectra were recorded on a Perkin Elmer 983G grating infrared spectrometer as thin films or chloroform solutions. NMR spectra were obtained for solutions in d-chloroform with residual protic solvent or tetramethylsilane as internal standard, and were recorded on JEOL FX90Q ( 90 MHz ), JEOL GSX-270 ( 270 MHz ), Bruker WM-250 ( 250 MHz ) or Bruker AM-500 ( 500 MHz ) instruments. Mass spectra were obtained with VG micromass 7070B instrument. Microanalyses were performed by the Imperial College Chemistry Department microanalytical laboratory. Optical rotations were measured on an Optical Activity AA-1000 polarimeter.
Analytical thin layer chromatography ( t.I.c.) was performed on pre-coated glass backed plates ( Merck Kieselgel 60 F 2 5 4 ). Preparative column chromatography was performed at low positive pressure on Merck Kieselgel 60 (230-400 mesh).
"Petrol" refers to redistilled light petroleum ether with b.p. 40-60° C unless otherwise indicated. "Ether" refers to diethyl ether. Ether, tetrahydrofuran, toluene and benzene were distilled from sodium- benzophenone ketyl under argon immediately prior to use. Dimethyl- formamide and dimethylsulphoxide were distilled from calcium hydride at reduced pressure, and stored over 4A molecular sieves under an argon atmosphere. Pyridine was distilled from potassium hydroxide and stored over 4A molecular sieves under an argon atmosphere. Dichloro- methane was distilled from phosphorous pentoxide, under an argon 111 atmosphere, immediately before use. All other solvents and reagents were purified by standard methods. Solvents were removed with a rotary evaporator at water pump pressure, followed by static evaporation with an oil pump.
112 PREPARATION OF 5. 5- DIMETHYL-3- METHOXYCYCLOHEX-2-ENONE
Dimedone (20g, .14 mol ) was dissolved in methanol ( 500 ml ). Trimethylorthoformate ( 93 m l) and cone, sulphuric acid ( 8 m l) were added and the mixture refluxed for 5 hours. After allowing to cool, most of the methanol was removed in vacuo . The residue was then neutralised to pH 7 with aqueous sodium bicarbonate before extraction with chloroform (3 x 200 m l). The combined organic phases were dried over anhydrous sodium sulphate followed by evaporation of solvent in vacuo . The crude product was purified by vacuum distillation ( 74° C at 1mm of Hg ) to afford 5, 5-DIMETHYL-3- METHOXYCYCLOHEX-2- ENONE ( 20.91 g , 95% ) as a colourless oil (lit13 ). vmax (film) 2959, 1654, 1609, 1374, 1225, 1155, 823 enr 1 ; 8 h (90 MHz , CDCI3 ) 1.07 ( 6 H, s, C-5 gem Me ) , 2.21 & 2.72 ( 4H, s+s, C-4 & C-6 -CH2 ) , 3.69 ( 3H, s, MeO- ), 5.39 ( 1 H, s, C-2 alkenic =CH).
113 PREPARATION OF 4-TR1METHYLSILYLBUT-3-YN-1 -OL
(7 )
To anhydrous THF (450 m l) cooled to -30° C under an atmosphere of argon, was added n-BuLi (140 ml of a 2.5M solution in hexanes, .35 m ol). 3-Butyn-1-ol (10g , .143 mol) was then added dropwise over 15 minutes with stirring. After a further 20 minutes freshly distilled chlorotrimethylsilane ( 40ml , .3 m ol) was added in one portion. The cooling bath was removed and the reaction allowed to warm up to room temperature. 2M HCI (250 ml ) was added and the mixture stirred vigorously for 15 minutes. The organic and aq. phases were separated and the aq. phase extracted with ether ( 3 x 100 m l). The combined organic phases were washed with aq. sodium bicarbonate and dried over MgSC>4 . Solvents were removed in vacuo and the yellow residue distilled (Kugelrohr oven, 120°, 2mm Hg ) to afford 4- TRIMETHYLSILYL BUT-3-YN-1-OL (16.4g , 82% ) as a colourless oil ( lit14 ). Vmax (film) 3435, 3312, 2959, 2874, 2175,1250,1070, 894,843 cm-1; 5H (90 MHz , CDCI3 ) .12 ( 9H, s, SiMe3 ) ,1.83 (1H, br. s, O-H) , 2.47 (2H, t, J = 10 Hz, C-2-CH2) , 3.68 (2H, t, J = 1 0 Hz, C-1 -CH2).
114 PREPARATION OF 4- IQDO-1 -TRIMETHYLSILYLBUT-1 -YNE
Triphenylphosphite ( 34 ml , 130 mmol) and iodomethane (11.5 ml, 184 mmol ) were refluxed together with a gradually increasing oil bath temperature until the internal temperature reached 120°C. The heating bath was removed and 4 -trimethylsilyl-3-butyn-1-ol (16.6g , 117 mmol ) was added while the mixture was still warm. After stirring the mixture at room temperature for 6 hours, all volatile components were distilled off under vacuum (2mm of Hg , oil bath at 100°C). The resultant clear oil was passed through a short silica column eluting with light petroleum ether. Removal of the solvents at reduced pressure afforded 4-IODO-1-TRIMETHYLSILYLBUT-1-YNE (23.7g , 80% ) as a colourless oil ( lit 14 ). Vmax (film) 2960, 2176, 1250, 1172, 844 cm-1 ; 8h (90 MHz , CDCI3 ) .12 ( 9H, s, SiMe3 ) , 2.47 ( 2H, t, J = 1 0 Hz, C-2 -CH2 ) , 3.24 ( 2 H, t, J = 10 Hz, C-1 -CH2 ).
115 PREPARATION OF 5.5 DlMETHYL-3-( 4-TRIMETHYLSILYLBUT-3-YNYU CYCLOHEX-2- ENONE
O o
OMe SiMe3 + I Me3Si ( 9 )
4-lodo-1-trimethylsilylbut-1-yne (10g, 40 mmol) was dissolved in anhydrous pentane and cooled to -78° C under argon, t- Butyllithium (26 ml, 44 mmol) was added dropwise over 10 minutes with stirring. After 1 hour the solution of the lithio species was transferred via a lagged cannular to a vessel containing 3-methoxy-5,5 dimethyl- cyclohex-2-enone (8 g , 52 mmol) in THF (250 ml) at -78° C under argon. After a further 1 hour of stirring the mixture was allowed to warm up to room temperature. Distilled water (10 ml ) was added followed by removal of solvents in vacuo. To the residue was added 2M HCI 200 ml) and THF (200 ml ) and the homogeneous mixture was then stirred for 2 hours at room temperature. The mixture was then neutralised with aq. sodium bicarbonate and extracted with diethyl ether (3 x 150 m l). The combined organic phase were dried over magnesium sulphate before removal of solvents in vacuo. The resultant yellow oil was chromatographed (silica, 5% ether /petrol) to afford 5, 5-DIMETHYL-3-(4-TRIMEYHYLSILYLBUT-3-YNYL )
116 CYCL0HEX-2-EN0NE (7.85g, 80% ) as a colourless oil. v max 2958, 2176, 1670, 1628, 1250, 844 cnr 1 ; SH (270 MH z ; CDCI3 ) .109 (9H, s, SiMe3 ), 1.02 ( 6 H, s, gem Me), 2.20 ( 4H, m, C-2' -CH2 and C-6 -CH2 ), 2.40 ( 4H, m, C-6 -CH2 and C-V -CH2 ), 5.89 ( 1 H, m, =CH); mi_z 247(m-H+), 192, 177, 146, 73. ( Found C 72.40 , H 9.80 ; C 15H 24 Si O requires C 72.52 , H 9.74 %)
PREPARATION OF 5.5 DIMETHYL-3-f 4-TRIMETHYLSILYLBUT-3-YNYL) CYCLOHEX-2-EN-1 -OL
(9) (11)
Enone (9 ) (6 g , 24 mmol) was dissolved in anhydrous toluene (100 m l) and cooled in an ice / salt bath to -20° C under argon. DIBAL ( 1.5 M in toluene , 16 ml , 24 mmol ) was added dropwise over 15 minutes. After stirring for a further 3 hours at -20° C, water (3 m l) was added dropwise and the reaction allowed to warm up to room temperature. Sodium sulphate ( 40g) was then added and the mixture stirred overnight. The solid alum was filtered off and solvents were removed in vacuo. The crude product was chromatographed (silica , 30% ether / petrol ) to afford 5,5 DIMETHYL-3-( 4-TRIMETHYL
117 SILYLBUT-3-YNYL )CYCL0HEX-2-EN-1-0L ( 5.32g, 88% ) as a colourless oil. vmax 3324, 2954, 2176, 1249, 843 cm-1 ; 8h (250 MHz ; CDCI3 ) 0.13 ( 9H, s, SiMe3 ) , .8 8 (3H, s, C-5 Me) , .99 (3H, s, C-5 Me) , 1.24 (1H, b, O-H ) , 1.70-1.93 (4H, m, C-4 allylic -CH 2 and C-6 CH2 ) , 2.17 ( 2 H, m, C-1 ' allylic -CH2 ) , 2.32 ( 2H, t, J = 7HzC-2' propargylic -CH2 ) , 4.25 (1H, bm, HO-C-H ) , 5.47 ( 1H, m, C-2 alkenicCH). (Found: C 71.80, H 10.61; Ci 5 H26OSi requires : C 71.93, H 10.46%)
Preparation of Zinc / Silver Couple
AgOAc (30 mg , .18 mmol ) was dissolved in hot acetic acid (15ml). Zinc (1.9 3g, 30 mmol ) was added in one portion with vigorous stirring. After 1 minute the acetic acid was decanted off. The zinc / silver couple was then washed repeatedly with dry ether ( c. 6 times ) until no trace of acetic acid could be detected.
118 PREPATION OF MRS. 5SR. 6 RS 1-3.3-DIMETHYL-1 -(4-TRIMETHYL SILYLBUT-3-YNYL1-BICYCLO T4 .1 .0 1 HEPTAN-5- OL
OH OH
(11 ) ( 1 2)
Alcohol (11 )(2g, 8 mmol) and diiodomethane ( 2.58ml, 32 mmol) were dissolved in anhydrous ether ( 50 ml ) and added to freshly prepared zinc / siver couple (2g , c. 32 mmol ). The mixture was refluxed under argon with vigorous stirring for 3 hours. After cooling to 0° C, saturated aq. ammonium chloride ( 30ml ) was added dropwise. The aqueous and organic phases were separated and the aqueous phase extracted with ether ( 2 x 25ml ). The combined organic phases were dried over MgS 0 4 before removal of solvents in vacuo . The residue was chromatographed ( silica , 40% ether/ petrol ) to afford! 1RS, 5SR, 6 RS )-3,3-DIMETHYL-1-(4-TRI METHYLSILYLBUT-3-YNYL)-BICYCLO[4. 1. 0 ] HEPTAN-5- OL ( 1.39g , 6 6 %) as a colourless oil. Vmax (film) 334, 2953, 2175, 1249, 1032, 841 cm-1 ; SH (270 MHz, CDCI3 ) .14 ( 9H, s, SiMe3 ) , .27 (1H, t, J = 4.8 Hz C-7 C-H, ) , .48 ( 1H, dd, J = 4.8, 9 Hz, C-7-CH) .71 ( 1 H, m, C-2 H-C cis to cyclopropane) , .84 (3H, s, C-3 Me-) , .87 ( 3H, s, C-3 Me-) , 1.16-1.27 (3H, m, C-1'-CH 2 and C-2 C-H trans to
119 cyclopropane) , 1.39-1.62 (3H, m, C-4 -CH 2 and O-H) , 2.28 ( 2H, t, J = 7.8 Hz , C-2' CH2) , 4.27 ( 1H. m, C-5 -CH ); rn/z 264 (M+), 249, 231 75, 73 . (Observed M+ 249.1674 ; C*i 6 H28SiO requires M+ 249.1679 .)
PREPARATION OF o-IMRS. 5SR. 6RS 13.3-DIMETHYL-1-(4-TRI - METHYLSILYLBUT-3-YNYL VBlCYCLOf4. 1. O 1HEPT-5-YL11- IMIDAZOLE THIOCARBOXYLATE
(12) (13)
Alcohol ( 12 ) ( 500mg , 1.90 mmol) was dissolved in dry dcm ( 25 m l). N , N'-thiocarbonyldiimidazole (2g , 10.99 mmol ) was added and the mixture refluxed under argon for 5 hours. On cooling, the mixture was diluted with more dcm ( 25 ml ) and then washed sequentially with water, dilute HCI, sat. NaHC 0 3 . water and brine before drying over Na 2 S0 4 . Removal of solvents in vacuo afforded o-[ ( 1RS, 5SR, 6RS)-3,3-DIMETHYL-1-(4-TRIMETHYLSILYLBUT-3-YNYL)
120 -BICYCL0[4.1.0]HEPT-5-YL] 1-IMIDAZOLETHIOCARBOXYLATE ( 71 Omg , quantitative yield) as a yellow oil. Vmax (film) 2955, 2175, 1688, 1215, 8 8 8 , 843 cm-1 ; 8h ( 270 MHz , CDCI 3 ) .14 ( 9H, s, SiMe3 ) , .27 (1H, t, J = 4.8 Hz , C-7 C-H ) , .52 ( 1H, dd, J = 4.8, 9 Hz , C-7 -CH ) .71 ( 1H, m, C-2 H-C cis to cyclopropane) , .87 (3H, s, C-3 Me-) , .1.02 (3H, s, C-3 Me-) , 1.16-1.27 ( 3H, m, C-1’ -CH 2 and C-2 C-H trans to cyclopropane) , 1.39-1.62 (2H, m, C-4 -CH 2 ) , 2.28 (2H, t, J = 7.8 Hz C-2’ CH2) , 4.39 (1H, m, C-5 -CH ) , 7.06 ( 1 H, m, imidazole CH) , 7.44 (1H, m, imidazole CH) , 8.17 ( 1 H, m, imidazole CH ).
PREPARATION OF (2 EZ ) . 9.9-DIMETHYL-2-r (TRIMETHYLSILYU METHYLENE1 SPIRO T4.51 DEC-6 -ENE
S
Thiocarbonylimidazolide derivative (13 ) ( 700 mg , 1.86 mmol ) was dissolved in dry degassed benzene (30 m l). The solution was brought to reflux under argon and a solution of tri-n-butyltinhydride (1 ml , 3.7 mmol ) and AIBN (50 mg) in benzene ( 4 ml ) was added dropwise over 30 minutes. The mixture was then allowed to reflux overnight. 121 On cooling CCI4 (5 m l) was added with stirring for 10 minutes. A dilute solution of iodine in ether was titrated in until a faint yellow colour persisted. After dilution with more ether (50 m l), the mixture was washed three times with 5% aqueous potassium fluoride and then dried over MgS0 4 before removal of solvents in vacuo. The residue was chromatographed ( silica , pentane ) to afford ( 2 EZ ) , 9,9-DIMETHYL-2-[ (TRIMETHYLS1LYL) METHYLENE ]-SPIRO [4 .5 ] DEC-6-ENE (327 mg , 71 % ) as a colourless oil. Vmax (film) 2951, 1622, 1246, 872, 837 cm-1 ; 8h (270 MHz , CDCI 3 ) .07 ( 4.5H, s , SiMe 3 of one isomer) , .09 (4.5H, s, SiMe3 of one isomer) , .93 (1.5H, s, C-9 Mg- of one isomer ) , .95 (3H, s, C-9 Me-) , .945 (1.5H, s, C-9 Me- of one isomer ) , 1.40 (1H, s, C-10 -CH2 of one isomer ) , 1.41 ( 1 H, s, C-1 0 -CH2 of one isomer ) , 1.52-1.75 ( 2 H, m, C-4 -CH2 ) , 1.78 ( 2 H, m, C-8 allylic -CH2 ) , 2.24-2.46 (4H, m , C -1 & C-3 allylic -CH 2 's) , 5.36 ( 1 H, m, exocyclic C=CH ) , 5.53 ( 2H, m, C-6 & C-7 alkenic =CH 's ) ; m /z 248 (M +), 233, 192, 174, 73. (Observed M+ : 248.1960 ; C i6 H2sSi requires 248.1960 )
122 PREPARATION OF 5.5.6-TRlMETHYL-3-(4-TRIMETHYLSILYLBUT-3- YNYD-CYCLOHEX-2-ENONE
O o
SiMe3 SiMe3 (9 ) (18)
Diisopropylamine (1.58 ml, 11.24 mmol ) was dissolved in dry THF (40 m l) and cooled to 0° C under argon. n-BuLi (4.50 ml of a 2.5M solution in hexanes , 11.24 mmol ) was added dropwise and the mixture left to stir for 30 minutes. On cooling the solution of LDA to -78° C, enone ( 9 ) ( 2.5 g , 10 mmol ) in THF ( 3 ml ) was added dropwise. After 1 hour at -78° C, the orange-red solution of the anion was transferred via cannula to a flask containing Mel (700 ml, 11.24 mmol ) in THF ( 50 m l) at -78° C under argon. The mixture was stirred at -78° C for 30 minutes and then allowed to warm up to room temperature before addition of water ( 50 m l), followed by extraction with ether ( 3 x 30 m l). The combined organic extracts were washed sequentially with dilute aq. HCI, saturated aq. Na 2 S0 4 aq. sodium thiosulphate, water, brine and then dried over MgS 0 4 . Removal of solvents in vacuo yeilded a yellow oil which was chromatographed ( silica, 10% ether / petrol ) to afford 5,5,6-TRI METHYL-3-(4- TRIMETHYLSILYLBUT-3-YNYL )-CYCLOHEX-2-ENONE ( 2.03 g, 77 % ) as a colourless oil. vmax (film) 2961, 2872, 2176, 1672, 1636, 1250, 843 crrr 1 ;
123 8 h (270 MHz , CDCI3 ) .11 ( 9H, s, SiMe3 ) , .87 ( 3H, s, C-5 Me-) , 1.04 ( 3H , s, C-5 Me-) , 1.04 (3H, d, J = 6 .8 Hz,C -6 Me-) , 2.12- 2 .2 1 ( 3H, m, C-4 allylic -CH 2 & C-6 -CH) , 2.38-2.44 (4H , m, C-1' & allylic -CH 2 , propargylic -CH 2 ) ,5.84(1H, m, C-2 alkenic =CH ); m /z 262 (M+), 234, 177, 73. ( Observed M+ : 262.1745 ; C i 6 H26 SiO requires : 262.1753 )
PREPARATION OF MRS. 6 RS 1-5.5.6-TRIMETHYL-3T4-TRIMETHYL SILYLBUT-3-YNYL 1-CYCLOHEX-2-EN-1 -OL
O OH
(18) (19 )
Lithium tri-/so- butylborohydride ( L- Slectride ) (3.8 ml of a 1 M solution in THF, 3.74 mmol ) was dissolved in THF ( 3 ml ) and cooled to -78° C under argon . Enone (18 ) (1 g , 3.75 mmol) dissolved in THF (1 m l) was added dropwise over 10 minutes. After stirring at -78° C for 2 hours, the cooling bath was removed and water ( 2 0 m l) was added with stirring for 20 minutes. The mixture was extracted with ether ( 4 x 5 m l) and the combined organic phases were dried over
124 MgS0 4 before removal of solvents in vacuo . The residue was then chromatographed ( silica , 30 % ether / petrol) to afford (1RS, 6RS )- 5,5,6-TRIMETHYL-3-(4-TRIMETHYLSILYLBUT-3-YNYL)- CYCLOHEX-2-EN-1-OL (695 mg , 69 % ) as a white, low melting point waxy solid. Vmax 3341 , 2960 , 2878 , 2175 , 1671 , 1249 , 843 cm-1 ; 8h (270 MHz , CDCI3 ) .18 ( 9H, s, SiMe3 ) ,.92 (3H, d, J = 8 Hz, C-6 Me- ) , .975 ( 3H, s, C-5 Me-) , .98 (3H, s, C-5 Me-) , 1.28 (1H, d, 7.8 Hz, O-H ) , 1.59 (1H, m, C-6 -CH) , 1.65 (1H, d, J = 16Hz, C- 4 -CH ) , 1.88 (1H, d, J = 16 Hz, C-4 -CH ) , 2.18 ( 2H, t, J = 7 Hz, C-1' allylic -CH 2 ) ,2.37 ( 2H, t, J = 7Hz, C-2' propargyl -CH 2 ) , 4.29 ( 1H, b, C-1 -CH } , 5.42 (1H, m, C-2 alkenic -CH); m /z 264 (M +), 249, 152, 73. (Found : C 72.56 , H 10.95 ; C i 6 H2 8 SiO requires : C 72.66 , H 10.67 %)
PREPARATION OF ( 1RS. 4SR. 5SR. 6RS 1-3.3.4-TRIMETHYL-1-(4- TRIMETHYLSILYLBUT-3-YNYL 1-BICYCLO T4.1.01 HEPTAN-5-OL
OH c m
(19) (20 )
Freshly prepared zinc/silver couple ( 660 mg , c. 10.9 mmol ) was added to a solution of alcohol (19) (480 mg, 1.81 mmol) in dry ether (15 ml). The mixture was brought to reflux under argon with vigorous
125 stirring. Diiodomethane ( 265 pi, 10.86 mmol) was added dropwise over 10 minutes and the mixture was then refluxed for 3.5 hours. After cooling to 0° C, saturated aq. NH 4CI ( 20 ml ) was added dropwise. The phases were separated and the aqueous phase extracted with ether (2 x 1 0 m l). The combined organic phases were dried over MgS0 4 before removal of solvents in vacuo. The residue was chromatographed (silica, 30% ether / petrol ) to afford ( 1RS, 4SR, 5SR, 6RS)-3,3,4-TRIMETHYL-1-(4-TRIMETHYLSILYLBUT- 3-YNYL )-BICYCLO [4.1.0] HEPTAN-5-OL ( 374 mg, 74%) as a colourless oil . Vmax (film) 3374, 2960, 2912, 2866, 2174, 124, 1021, 842 cm-1 ; 5h (250 MHz , CDCI3 ) .14 ( 9H, s, SiMe3 ) , .38 ( 1 H, dd, J = 9,4.8 Hz, C-7 cyclopropyl -CH into cyclohexane ring ) , .49 (1H, t, J = 4.8 Hz, C-7 cyclopropyl -CH out of cyclohexane ring ) , .77 ( 3H, , J = 7.6 Hz, C-4 Me-) , .85 ( 3H, s, C-3 Me-) , .96 (3H, s, C-3 Me-) , 1.09 (1H, dt, J = 9, 4 Hz, C-6 ring fusion-CH ) , 1.20-1.33 (3H, m, C-2 -CH2 & O-H ) , 1.48- 1.80 ( 3H, m, C-1 ' -CH2 & C-V-CH) , 2.30 ( 2H, m, C-2' propargyl -CH 2 ) , 4.44 (1H, t, J = 6.5 Hz, C-5 C-1’ -CH); m /z 278 (M+) , 263, 245, 219, 75, 73. ( Found :C 73.57, H 11.07; Ci 7 H30SiO requires : C 73.31, H 10.86%)
126 PREPARATION OF CM MRS. 4SR. 5SR. 6 RS.) 3.3.4-TRIMETHYL-1-(4- TRIMETHYLSILYLBUT-3-YNYn-BICYCLOr4.1 . 0.1HEPT-5-YL 11- IMIDAZOLE THIOCARBOXYLATE
S
(20 ) (24 )
Alcohol (20 ) ( 250 mg, .9 mmol) and 1,1' thiocarbonyldiimidazole (1 g , 5.6 mmol ) were dissolved in dry DCM ( 20 ml ) and refluxed under argon for 8 hours. After cooling, the mixture was diluted with more DCM (20 m l) and washed sequentially with water, dilute aq. HCI, aq. NaHC 0 3 , water and brine before drying over MgS0 4 . Removal of solvents in vacuo afforded 0-[ ( 1RS, 4SR, 5SR, 6 RS, ) 3,3,4- TRIMETHYL-1-(4-TRIMETHYLSILYLBUT-3-YNYL) BICYCLO[4. 1. 0.]HEPT-5-YL ] 1-IMIDAZOLETHIOCARBOXY-LATE as a yellow oil. Vmax (film) 2956, 2173, 1688, 1215, 841 cm-1 ; 5H (270 MHz , CDCI3 ) .13 ( 9H, s, SiMe3) , .48 ( 1 H, t, J = 4.8 Hz, C-7 cyclopropyl -CH ) , .55 ( 1 H, dd, J = 9,4.8 Hz, C-7 cyclopropyl - CH , ) , .89 ( 3H, s, C-3 Me- ) , .92 ( 3H, d, J = 7 Hz, C-4 Me- ) , 1.13 ( 3H, s, C-3 Me ) , 1.22- 1.42 ( 2 H, m, C-2 -CH2 ) , 1.57-1.85 ( 3H, m, C-1' -CH2 & C-4 -CH ) ,2.31 ( 2H, m, C-2' propargyl -CH2 ) , 4.74 ( 1 H, t, J = 6 .6 Hz, C-5 -CH ) , 7.09 ( 1H, m, imidazole ring
127 =CH ) , 7.49 (1H, m, imidazole ring =CH ) , 8.22 ( 1H, m, imidazole ring =CH).
PREPARATION OF (2EZ. 5RS. 81 8.9.9-TRIMETHYL-2-r ( TRIMETHYL- SILYL 1 METHYLENE ISPIROf 4. 5 1DEC-6-ENE
S
3 (24) (26)
The thiocarbonylimidazolide derivative ( 24 ) ( 245 mg , .63 mmol )u/as dissolved in dry benzene (15 ml) and brought to reflux under argon. Tri-n-butyltinhydride (.5 ml, 1.89 mmol) and AIBN ( 20 mg , .095 mmol) in benzene (2 m l) were added dropwise to the refluxing solution over 1 hour. The mixture was then allowed to reflux overnight. On cooling, CCU (5ml) was added with stirring for 10 minutes. A dilute solution of iodine in ether was then titrated in until a faint yellow colour persisted. After dilution with more ether ( 50 ml ), the mixture was washed three times with 5% aq. potassium fluoride and then dried over MgS0 4 before removal of solvents in vacuo. The residue was chromatographed ( silica, pentane ) to afford (2EZ, 5RS, 8SR ) 8,9,9- TRI METHYL-2-[ ( TRIMETHYLSILYL JMETHYLENE] SPIRO [ 4. 5 ]DEC-6-ENE ( 130 mg , 79% ) as a colourless oil.
128 Vmax (film) 2954, 1623, 1246, 872, 836 cm-1; §H (270 MHz , CDCI3 ) .07 ( 4.5H, s, SiMe 3 ) , .08 (4.5H, s, SiMe3 ) , 0.78 ( 1.5H, s, C-9 Me) , .80 (1.5H, s, C-9 Me) , .87 (1.5H, d, J = 7.2 Hz, C -8 Me) , .90 ( 1.5H, d, J = 7.2Hz, C-8 Me) , .93 (3H, s, C- 9 Me) , 1.26- 1.65 ( 4H , m, C-4 & C-10 -CH 2 ) , 1.89-1.93 ( 1 H, m, C-8 allylic -CH ) , 2.14-2.23 ( 4H, m, C -1 & C-3 allylic-CH 2 ) , 5.28- 5.47 (3H, m, alkenic =CH at C -6 C-7 & C-1' ); m /z 262 (M+), 192, 163, 75, 73 . ( Observed M+ 262.2121 ; Ci7 H3oSi requires 262.2117 )
PREPARATION OF BENZOIC ACID f M RS. 6SR ) 5.5.6-TRIMETHYL-3- (TRIMETHYLSILYLBUT-3-YNYL) CYCLOHEX-2-ENYL 1 ESTER
OH
(21)
Alcohol ( 19 ) ( 500 mg, 1.87 mmol) and triphenylphosphine (487 mg, 1.87 mmol) dissolved in dry ether (3 m l) were added dropwise to a solution of diethylazodicarboxylate ( 300 ml , 1.87 mmol) and benzoic acid ( 227 mg , 1.87 mmol) in ether ( 7 m l). After stirring at room temperature overnight, the solid precipitate was filtered off and the filtrate was concentrated in vacuo . The residue was chromatographed ( silica , 5 % ether / petrol to afford BENZOIC ACID [ ( 1 RS, 6 SR ) 5,5,6-TRIMETHYL-3-(TRIMETHYLSILYL BUT-3-
129 YNYL ) CYCL0HEX-2-ENYL ] ESTER ( 478 mg , 65 % ) as a white crystalline solid ( m.p. 64- 65 0 C ). Vmax ( CHCI 3 ) 2961, 2175, 1713, 1268, 1111, 842 cm-1 ; 6 h (270 MH z , CDCI3 ) .1 2 ( 9H, s, SiMe3 ) , .89 ( 3H, s, C-5 Me-) , 95 (3H , d, J = 7 Hz, C-6 Me- ) , 1.02 (3H, s, C-5 Me-) , 1.77 (2H, m, C-4 allylic CH2 ) , 2.03 ( 1 H, m, C-6 -CH) , 2.16 ( 2 H, m, C-1 * allylic CH2) , 2.34 (2H , t, J = 7.1 Hz, C-2' propargyl -CH 2 ) , 5.28 ( 1H, m, C-1 -CH) , 5.49 ( 1 H, b, C-2 alkenic =CH) , 7.42 (2H, m, m- aromatics) , 7.55 (1H, m, aromatic) , 8.05 (2H, m, £- aromatics) ; m /z 368 (M + ), 353, 279, 263, 105. ( Found : C 74.69 , H 8.63 ; C 23 H32Si0 2 requires : C 74.95 , H 8.74 % )
PREPARATION OF (1RS. 6 SR )-5.5.6-TRIMETHYL-3-(4-TRIMETHYL- SILYLBUT-3-YNYL 1 CYCLOHEX-2-EN-1 -OL
Titanium IV iso - propoxide ( 760ml, 2.6 mmol) was added to ester (21 ) ( 950 mg , .2.6 mmol ) dissolved in iso -propanol (50 ml ) and the solution was refluxed under argon of 8 hours. After cooling to room temperature, water ( 50 ml ) was added followed by extraction with ether ( 3 x 25 m l). The combined organic extracts were dried over MgS0 4 before removal of solvents in vacuo. The residue was
130 chromatographed (silica, 25% ether / petrol) to afford (1RS, 6SR )- 5,5,6-TRIM ETH YL-3-(4-TRIMETH YLSILYLBUT-3-YN YL)-CYCLO- HEX-2-EN-1-OL ( 470 mg , 69 % ) as a low melting point waxy solid. Vmax (film) 3315, 2960, 2891, 2177, 1250, 1017, 842 cm-1 ; 8h (270 MHZ , CDCI3 ) .14 (9H, s, SiMe3 ) , .77 (3H, s, C-5 Me-) , .94 (3H , s, C-5 Me-) , 1.01 (3H, d, J = 6 .8 Hz, C-6 Me-) , 1.20- l. 29 ( 1 H, , O-H ) , 1.63- 1.70 (2H, m, C-4 allylic -CH2 ) , 1.94 ( 1 H, m, C-6 -CH) , 2.16 (2H, m , C-1* allylic -CH2 ) , 2.33 ( 2H, t, J = 7 Hz, C-2' allylic -CH 2 ) , 3.72 ( 1 H, b, C-1 -CH) , 5.44 (1H, br. s, alkennic =CH ); m /z 264 (M + ), 249, 20 , 165, 152, 75, 73. (Observed M+ : 264.1909; Ci6 H2sSiO requires 264.1909 )
PREPARATION OF ( 1RS. 4RS. 5SR. 6 RS 1-3.3.4-TRIMETHYL-1-(4- TRIMETHYL SILYLBUT-3-YNYL VBICYCLO T4.1.01 HEPTAN-5-OL
OH OH
(22) (23)
Freshly prepared Zinc / Silver couple (410 mg , c. 6.79 mmol ) was added to a solution of alcohol (22)(300 mg , 1 .1 0 mmol) in dry ether ( 1 0 ml). The mixture was brought to reflux under argon with vigorous stirring. Diiodomethane (160 ml , 6.79 mmol ) was added dropwise over 10 minutes and the mixture was then refluxed for 2 hours. After cooling to 0° C, saturated aq. NH 4 CI ( 20 m l) was added dropwise. The phases were separated and the aqueous phase extracted with
131 ether ( 2 x 10 ml ). The combined organic phases were dried over MgS0 4 before removal of solvents in vacuo . The residue was chromatographed (silica, 35% ether / petrol) to afford ( 1RS, 4 RS, 5SR, 6 RS )-3,3,4-TRIMETHYL-1-(4-TRIMETHYLSILYLBUT-3- YNYL )-BICYCLO [4.1.0] HEPTAN-5-OL (264 mg , 84%) as a colourless oil . Vmax (film ) 3370, 2960, 2912, 2866, 2175, 1249, 1021, 843 cm-1 ; 8h (270 MHz , CDCI3 ) .13 (9H, s, SiMe3 ) , .28 ( 1 H, t, J = 4.7Hz, C-7 cyclopropyl -CH ) , .48 (1H, dd, J= 4.8, 9 Hz , C-7 cyclopropyl -CH ) , .77 ( 3H, s, C-3 Me- ) , .79 ( 3H, s, C-3 Me-) , .93 ( 3H, d, J = 7.2 Hz, C-4 Me-) , 1.12 (1H, d, J = 8 Hz, O-H ) , 1.20-1.65 ( 6 H, m, C-1' & C-5 -CH2 ; C-6 & C-3 -CH) , 2.29 (2H, t, J = 7Hz, C-2' propargyl -CH 2 ) , 3.76 ( 1 H, m, C-5 HO-CH) ; m /z 278 (M+1. 263, 245, 219, 179, 75, 73 . ( Observed M+ : 278.2069 ; C i 7 H2oSiO requires 278.2066 )
132 PREPARATION OF O-rMRS. 4RS. 5SR. 6RS .) 3.3.4-TRIMETHYL-1- (4-TRIMETHYLSILYLBUT-3-YNYLVBICYCLOr 4. 1.0.1HEPT-5-YL11 - IMIDAZOLE THIOCARBOXYLATE
S
IV
SiMe3 SiMe3 (23) ( 2 5 )
Alcohol (23 ) ( 230 mg , .83 mmol ) and 1,1' thiocarbonyldiimidazole (1 g , 5.5 mmol) were dissolved in dry DCM ( 15 ml ) and refluxed under argon for 5 hours. After cooling, the mixture was diluted with more DCM ( 15 ml ) and washed sequentially with water, dilute aq. HCI, aq. NaHC 0 3 , water and brine before drying over MgSO/j. Removal of solvents in acuo afforded 0-[(1RS, 4RS, 5SR, 6RS, ) 3,3,4-TRIM ETHYL-1-(4-TRIMETHYLSIL YLBUT-3-YN YL)-BICYCL [ 4. 1. 0. JHEPT-5-YL ] 1-IMIDAZOLE THIOCARBOXYLATE as a yellow o il. Vmax (film ) 2960, 2174, 1688, 1466, 1363, 1292, 1269, 1214, 886, 843 cm-1 ; 5h (270 MHz , CDCI 3 ) .12 ( 9H, s, SiMe3 ) , .24 ( 1 H, t, J = 4.5 Hz, C-7 cyclopropyl -CH ) , .65 ( 1H, dd, J = 9 , 4.5 Hz , C-7 cyclopropyl -CH ) , .84 ( 3H, s, C-3 Me- ) , .91 ( 3H, s, C-3 Me ) , .99 ( 3H, d, J = 7 Hz , C-4 -CH 3 ) , 1.08 (1H, m, C-6 -CH ) , 1.23- 1.70 ( 5H , m, C-2 C-1' -CH2 & C-4 -CH), 2.28 ( 2 H, m, C-2 ' propargyl -CH 2 ) ,
133 4.08 (1 H, dd, C-5 -CH ) , 7.11 (1H, m, imidazole ring =CH ) , 7.53 (1H, m , imidazole ring =CH) , 8.17 (1H , m, imidazole ring =CH) .
PREPARATION OF (2EZ. 5RS. 8 ) 8.9.9-TRIMETHYL-2- T ( TRIMETHYLSILYL 1 METHYLENE 1 SPIROf 4. 5 1DEC-6-ENE
S
( 25 ) (27 )
The thiocarbonylimidazolide derivative ( 25 ) ( 300 mg, .77 mmol ) dissolved in dry benzene (15 ml) and brought to reflux under argon. Tri-n-butyltinhydride (.61 ml, 2.30 mmol) and AIBN (40 mg, .184 mmol) in benzene ( 2 m l) was added dropwise to the refluxing solution over 1 hour. The mixture was then allowed to reflux overrnight. On cooling CCU (5 m l) was added with stirring for 10 minutes. A dilute solution of iodine in ether was titrated in until a faint yellow colour persisted. After dilution with more ether (50 m l) the mixture was washed three times with 5% aq. potassium fluoride and then dried over MgS 0 4 before removal of solvents in vacuo . The residue was chromatographed ( silica , pentane ) to afford (2EZ, 5RS, 8 RS ) 8,9,9-TRIMETHYL-2-[ ( TRIMETHYLSILYL ) METHYLENE ] SPIRO[ 4. 5 ]DEC-6 -ENE ( 130 mg , 79 % ) as a colourless oil.
134 vmax (film) 2955, 2900, 1646, 1462, 1247, 844 cm-1 ; 5h (270 MHz , CDCI3 ) .02 ( 4.5H, s, SiMe3 ) , .03 (4.5H, s, SiMe3 ) , 0.78 ( 1.5H, s, C-9 Me) , .795 (1.5H, s, C-9 Me) , .8 6 (1.5H, d, J = 7 Hz, C -8 Me) , .8 8 ( 1.5H, d, J = 7Hz, C-8 Me) , .91 ( 3 H, s, C-9 Me) , 1.40- 1.61 ( 4H , m, C-4 & C-10 -CH 2 ) , 1 .8 6 ( 1 H, m, C-8 allylic -CH) , 2.07- 2.23 ( 4H, m, C-1 & C-3 allylic -CH 2 ) , 4.84 ( .5H, m, C-1' =CH) , 5.01 ( ,5H, m, C-1’ =CH) , 5.22 ( ,5H, t, J = 3 Hz, C-6 =CH ) 5.26 ( .5H, t, J = 3 Hz, C -6 =CH ) 5.39 ( ,5H, d, J = 3 Hz , C-7 =CH ) 5.60 ( .5H, d, J = 3 Hz, C-7 =CH ); m /z 262 (M +), 247, 192, 163, 75, 73. ( Observed M+ 262.2120 ; Ci7 H3oSi requires 262.2117 )
135 PREPARATION OF 3-(OXYPROP-2-YNYL1-5.5-DlMETHYLCYCLOHEX-2- ENONE
O O
o +
( 30) Dimedone ( 15 g, .105 mol ), propargyl alcohol ( 12 ml, .21 mol ) and jl-toluenesulphonic acid (1.5g, 7.8 mmol) were dissolved in benzene ( 300 m l) and the mixture refluxed in a vessel fitted with a Dean- Stark separator for 8 hours. On cooling, the organic phase was washed with aq. Na 2C0 3 (75 m l) and then dried over Na2 S0 4 . After removal of solvents in vacuo , the resultant yellow solid was recrystallised from 90 % petrol / ether to afford 3-(OXYPROP-2-YNYL)- 5,5-DIMETHYLCYCLOHEX-2-ENONE { 15.23 g, 81% ) as a white crystalline solid ( m.p. 53 0 C ). Vmax (solvent : chloroform ) 3305, 2957, 2129, 1638, 1607, 1381, 1356, 1158, 1114, 1012 cm -1 ; 6 h ( 270 MHz , CDCI3 ) 1.05 ( 6 H, s, C-5 gem Me-) , 2.20 (2H, s, C-6 -CH2 ) , 2.28 ( 2 H, s, C-4 allylic -CH2) , 2.56 ( 1 H, t, J = 2.4Hz , C-3' acetylenic CH) , 4.52 (2H, d, J = 2.4Hz, C-1' propargyl -CH 2 ) , 5.41 ( 1 H, br.s, C-2 alkenic=CH) ; m /z 178 (M +), 121, 94, 69. (Found : C 73.94 , H 7.91; CnH 140 2 requires C 74.13 , H 7.92% )
136 PREPARATION OF 3-(OXYPROP-2-YNYLV5.5-DIMETHYLCYCLOHEX-2- EN-1-OL
Enone (30) (3 g, 18.80 mmol) was dissolved in toluene (30 ml ) and cooled to -78° C under argon. DIBAL ( 16.8 ml of a 1.5M solution in toluene, 25.2 mmol ) was added dropwise over 10 minutes. After stirring at -78° C for 3 hours, methanol (30 m l) was added dropwise and the mixture allowed to warm to room temperature. Celite ( 5 g ) was added with stirring and the mixture was then filtered through a celite pad . Solvents were then removed in vacuo to afford 3-(OXY PROP-2-YNYL)-5,5-DIMETHYLCYCLOHEX-2-EN-1-OL ( 2.98g , 98 %) as a colourless oil. Vmax (film) 3292, 2953, 2122, 1663, 1139, 1024 cm-1 ; 6 h (270 MHz , CDCI3) .93 (3H, s, C-5 gem Me-) , 1.03 (3H, s, C-5 gem Me-) , 1.26 (1H, dd, J = 13.8, 8.2 Hz, C-6 -CH) , 1.40 (1H, d, J= 8.0 Hz, O-H) , 1.77 (2H, m, C-4 allylic -CH & C-6 -CH ) , 2.05 ( 1H, m, C-4 allylic-CH) , 2.49 (1H, t, J = 2.7Hz , C-3' acetylenic CH) , 3.48 (1H, br. d, J = 5.4Hz, C-1 -CH) , 4.40 (2H, d, J = 2.7 Hz, C-1' propargyl -CH 2 ) , 4.78 (1H, m, C-2 alkenic =CH ); m /z 180 (M+) , 162, 95, 85, 71, 68. ( Observed M+ 180.1153 ; C11H16 O2 requires 180.1150 )
137 PREPARATION OF MRS. 5RS. 6RS ) U OXYPROP-2-YNYLV3.3- DIMETHYLBICYCLOr 4. 1.0.1HEPTAN-5-QL
Zinc dust ( 1.8 g, 27.6 mmol) and copper ( I ) chloride ( 274 mg, 2.76 mmol ) were added to ether ( 25 ml) and the mixture was refluxed with stirring under argon for 40 minutes. Diiodomethane ( .6 8 mmol, 27.6 mmol) and DME (3.1 m, 30 mmol ) were added and the reflux continued for 40 minutes. Allylic alcohol (31 ) (1 g, 5.52 mmol) in ether ( 3 ml ) was then added dropwise over 10 minutes. After a further 30 minutes, the reaction mixture was cooled to 0° C and ammonium chloride ( 20 ml ) was added dropwise. The aqueous and organic phases were separated and the aqueous phase extracted with ether ( 2 x 10 ml). The combined organic phases were dried over MgS0 4 before removal of solvents in vacuo . The residue was chromatographed ( basic alumina, 60% ether / petrol ) to afford ( 1RS, 5RS, 6 RS ) 1-( OXYPROP-2-YNYL)-3,3-DIMETHYL BICYCLO[ 4. 1.0. ]HEPTAN-5-OL ( 419 mg, 39 % ) as a colourless oil. Vmax (film) 3300, 2925, 2862, 2119, 1451, 1363, 1223, 1041 cm-1 ; 8h (270 MH z , CDCI3 ) .38 (1H, pt, J = 6 Hz, C-7 cyclopropyl -CH) , .64 (1H, t, J = 11.5 Hz , C-7 cyclopropyl -CH) , .90 ( 3H, s, C-3 gem Me-) , .98 (3H, s, C-3 gem Me- , 1.03 (1H, m, C-6 -CH) , 1.25 (1H , br.d, J = 5.4Hz, O-H) , 1.49 (1H, ddd , J = 13, 6.2, 1.8 Hz, C-4 -CH) , 1.59 ( 1H, br.d, J = 14.5 Hz, C-2 -CH) , 1.73 (1H,
138 dt, J = 6.2,10.5 Hz, C-4 -CH ) , 1.85 (1H, br. d , J = 14.5 Hz , C-2- CH) , 2.40 ( 1H, t, J = 2.5 Hz , C-3' acetylenic CH) , 4.08 (1H, d, J = 2.5 Hz, C-1* propargyl -CH ) , 4.09 ( 1H, d, J = 2.5 Hz, C-1* propargyl -CH), 4.31 ( 1H, m, C-5 -CH) ; m iZ 194 (M +), 193, 176, 161, 155, 85, 55. ( Observed M+ 194.1307 ; C12H18O2 requires 194.1302 )
PREPARATION OF 3-(OXYPROP-2-YNYn-5.5-DIMETHYLCYCLOHEPT- 1.3-DIENE
S
(34) Alcohol ( 32 ) (350 mg, 1.80 mmol ) was dissolved in dry DCM (25 ml). 1, I'-thiocabonyldiimidazole (1 g, 5.5 mmol ) was added and the mixture refluxed under argon for 8 hours. On cooling, the mixture was diluted with more DCM ( 25 ml ) and then washed sequentially with water, dilute HCI, sat. NaHC 0 3 , water and brine before drying over Na 2 S0 4 . Removal of solvents in vacuo afforded a yellow oil which contained no imidazole or cyclopropyl protons in the nmr. This 139 product has been assigned as 3-(OXYPROP-2-YNYL)-5,5-DI- METHYLCYCLOHEPT-1,3-DIENE ( 292 mg, 9 2 % ). Vmax (film) 3294, 2953, 1647, 1620, 1592, 1200, 1137 cm-1 ; §H (270 MHz , CDCI3 ) .99 ( 6H, s, C-6 gem Me-), 2.03 ( 2H, m, C-5
acetylenic CH ) , 4.40 (2H, d, J = 2.5 Hz, C-1' propargylic -CH 2 ) , 5.02 (1H, d, J = 6Hz, C-2 alkenic =CH) , 5.60 (1H, m, C-4 alkenic =CH) , 5.79 ( 1H, m, C-3 alkenic =CH) ; m lz 177 (MH+) , 176 , 155, 85, 55, 41 . (Observed M+ 176.1201; C 12H16 O requires 176.1201 )
PREPARATION OF o-rMRS. 5RS. 6RS ) U OXYPROP-2-YNYLV3.3- DIMETHYLBICYCLOr 4. 1. 0 . 1HEPTAN-5-YL1 S-METHYL DITHIO CARBONATE
( 32) ( 33)
Alcohol (32 ) ( 300 mg, 1.54 mmol), sodium hydride dispersion ( 60 %, 123 mg , 3.08 mmol) and imidiazole (5 mg ) were stirred in THF (5 m l) at room temperature under argon for 3 hours. Carbon disulphide (2 m l) was added, and after a further 1 hour stirring, methyl iodide (2 m l) was added, and stirring continued for a further 1 hour. After addition of acetic acid (2 m l) the reaction was worked up, dilution
140 with DCM ( 30 ml ) and then sequential washing with sodium bicarbonate solution, water and brine, and drying over Na 2 S0 4 . Removal of solvents in vacuo afforded o-[(1RS, 5RS, 6RS ) 1-( OXYPROP-2-YNYL)-3,3-DIMETHYLBICYCLO[ 4. 1. 0. ] HEPTAN-5-YL] S-METHYL DITHIOCARBONATE ( 330 mg , 80% ) as a yellow oil. Vmax (film) 3291, 2925, 2118, 1619, 1461, 1219,1052, 974cm-1 8h (270 MHz , CDCI 3 ) .57 (1H, t, J = 6 Hz, C-7 cyclopropyl -CH ) , .83 (1H, pt, J = 11.5 Hz, C-7 cyclopropyl -CH) , .93 ( 3H, s, C-3 gem Me-) , .99 ( 3H, s, C-3 gem Me-), 1.18 (1H, m, C-6 -CH) , 1.60 ( 2 H, m, C-2 -CH2) , 2.02 ( 2H, m, C-4 -CH ) , 2.40 (1 H, t, J = 2.5 Hz,C-3' acetylenic CH ) , 4.08 (2H, d, J = 2.5Hz , C-1' propargyl -CH ) , 6.18 ( 1 H, m, C-5 -CH).
ATTEMPTED RADICAL REACTION WITH XANTHATE (331
(35)
Xanthate derivative (13 ) ( 570 mg , 2 mmol ) was dissolved in dry degassed toluene (30 ml ) . The solution was brought to reflux under argon and a solution of tri-n-butylstannane (1.7ml, 10 mmol) and AIBN 141 (50 mg) in toluene ( 4 m l) was added dropwise over 45 minutes. The mixture was then allowed to reflux for three hours. On cooling, CCI 4 (5 m l) was added with stirring for 10 minutes. A dilute solution of iodine in ether was titrated in until a faint yellow colour persisted. After dilution with more ether (50 m l), the mixture was washed three times with 5% aq. potassium fluoride and then dried over MgS 0 4 before removal of solvents i n vacuo. Nmr revealed a complex mixture of hydrostannylated materials.
142 PREPARATION OF 5.5-DIMETHYL-2-(PROP-2-YNYD-CYCLOHEXAN-1.3- DIONE
(43 )
To a solution of potassium hydroxide (10 g , .18 mol) in water (120 ml) was added dimedone ( 25 g , .18 mol). After heating to 50° C, propargyl bromide ( 14 ml , .2 mol) was added with vigorous stirring for 2 hours. On cooling, potassium hydroxide (10 g , .18 mol) was added with stirring and the aqueous and organic phases were then separated . The aqueous phase was acidified with 1 M aq. HCI to pH 5 t and the off white precipiate of crude product was filtered off. Chromatography ( silica, 70 % ether / petrol ) afforded 5,5-DI METHYL-2-(PROP-2-YNYL)-CYCLOHEXAN-1,3-DIONE (11.0 g , 35% ) as a white chrystalline solid ( m.p. 111-113° C ) Vmax (solvent: chloroform) 3391, 3303, 2956, 1735, 1704, 1623, 1379, 1041 cm-1 ; 8h (90 MHz , CDCI3 ) .82 (1.2H, s, Me-of diketone tautomer) , 1.07 ( 3.6H, s , gem Me- of enol tautomer) , 1.20 (1.2H, s, gem Me- of enol tautomer) , 1.89 ( .4H, t, J = 2.5Hz , acetylene CH of diketone tautomer) , 2.20 ( . 6 H, t, J = 2.5 Hz, acetylene CH of enol tautomer ) , 2.26-2.76 (5.4H, m, O-H of enol tautomer, propargyl -CH 2 of diketone tautomer, C-4&C-6 -CH 2 of both tautomers ) , 2.29 ( 1 .2 H,
143 d, J = 2.5 Hz ) , 3.55 ( .4H, t, J = 5 Hz, C-2 CH of diketone tautomer) ; m /z 178, 163, 149, 145, 135, 122, 94, 83. ( Found: C 73.85, H 7.89 ; CnH 140 2 requires : C 74.13 , H 7.89%)
PREPARATION OF 3-METHOXY-5.5-DlMETHYL-2-(PROP-2-YNYL1 CYCLOHEX-2-ENONE
(44)
To the dione ( 43 ) ( 5.75 g , 32 mmol) dissolved in methanol ( 250 m l) was added trimethylorthoformate (30 m l) and cone. H2SO4 ( 2 m l). The solution was then stirred at room temperature for 6 hours. Most of the methanol was then removed in vacuo and the residue neutralised to pH 7 with aq. NaHC 0 3 . The mixture was then extracted with chloroform (3x150 ml) and the combined organic phases dried over Na 2 S0 4 . Removal of solvents in vacuo gave a yellow solid. Recrystallisation ( 20 % ether / petrol ) afforded 3-METHOXY-5,5- DIMETHYL-2-(PROP-2-YNYL)-CYCLOHEX-2-ENONE ( 5.27g, 85%) as colourless needle shaped crystals ( m.p. 87-88°C).
144 Vmax (solvent :CHCI3 ) 3306, 2555, 2117, 1369,1076,1617 cnr1 ; 6 h (90 Mz , CDCI3 ), 1.06 (6H, s, C-5 gem Me- ) , 1.79 (1H, t, J = 2.8 Hz, C-3' acetylenic CH ) , 2.22 ( 2H, s, C-6 -CH2 ) , 2.42 ( 2H, s, C-4 allylic -CH2 ) , 3.14 ( 2H, d, J = 2.8Hz, C-1' allylic / propargylic -CH2) , 3.83 ( 3H, s, -OMe); m /z 192, 177, 149, 145, 135 121. (Found: C 74.92 , H 8.42 ; C 12H16 O2 requires: C 74.97, H 8.38%)
PREPARATION OF 5.5-DIMETHYL-2-(PROP-2-YNYU-CYCLOHEX-2- ENONE
(44) (45)
Ketone (4 4 )(3 .6 4 g , 18.96 mmol) was dissolved in toluene (50 ml ) and cooled to O0 C under argon. DIBAL ( 19.04 ml of a 1.5M solution in toluene, 28.44 mmol) was added dropwise over 10 minutes. After stirring at O0 C for 2 hours, water ( 15 ml ) was added dropwise followed by 2M HCI (10 ml) and the mixture was stirred vigorously for 30 minutes. The phases were separated and the aqueous phase extracted with ether (2 x1 5 ml). The combined organic extracts were washed with aq. NaHCC >3 and then dried over MgS0 4 before removal of solvents in vacuo.. Chromatography ( silica, 40 % ether / petrol )
145 afforded 5,5-DIMETHYL-2-(PROP-2-YNYL)-CYCLOHEX-2-ENONE (3.40 g, 96 %) as a colourless oil . Vmax (film) 3290, 2957, 2121, 1673, 1378 cm-1 ; 8h (270 MH z , CDCI3 ) , 1.03 ( 6 H, s, C-5 gem Me-) , 2.17 (1H, s, J = 2.7 Hz, C-3’ acetylenic CH ), 2.31 ( 4H, br. s, C-4 ally lie & C-6 - CH2 ) , 3.17 ( 2 H, m, C-1 ' allylic / propargylic -CH 2 ) ,7.01 ( 1 H, m, C- 3 alkenic =CH). ( Found: C 81.37 , H 8.89 ; C 11H14O requires :C 81.44 , H 8.70%)
PREPARATION OF 5.5-DIMETHYL-2-(PROP-2-YNYn-CYCLOHEX-2-EN- 1-OL
(45) (46)
The enone ( 45 ) ( 2.80 g, 17.28 mmol) was dissolved in toluene (50 ml) and cooled to -78° C under argon. Dibal ( 11.56ml of a 1.5M solution in toluene , 17.28 mmol ) was added dropwise over 10 minutes . After stirring at -78° C for 3 hours, water ( 4 ml ) was carefully added and the mixture allowed to warm up to room temperature. Ethyl acetate ( 150 m l) was added followed by a large excess of Na 2 S0 4 ( 40 g ) and the mixture stirred at room temperature overnight. The solid alum was filtered off before removal of solvents
146 in vacuo. The crude material was chromatographed ( silica, 30 % ether / petrol ) to afford 5,5-DIMETHYL-2-(PROP-2-YNYL)-CYCLO HEX- 2-EN-1-OL (2.61g, 92% ) as a colourless oil. Vmax (film) 3306, 2950, 2117, 1465, 1043 cm-1 ; §H (270 MHz , CDCI3 ) .91 ( 3H, s, C-5 Me-) , .99 ( 3H, s, C-5 Me-) , 1.41 ( 1H, dd, J =8 .8 , 12.7 Hz C -6 -CH ) , 1.58 (1H , br.d, J = 7 Hz, -O-H) , 1.74- 1.97 ( 3H, m, C-4 allylic -CH 2 & C-6 -CH ) , 2 .1 2 (1H, t , J = 2.7 Hz, C-3' acetylenic CH ) , 3.08 (2H, m, C-V allylic/ propargylic -CH 2 ) , 4.24 ( 1 H, b, C-1 HO-CH) , 5.75 ( 1 H, m, C-3 alkenic =CH ); m lz 164 ( M+ ), 125, 109 , 79. ( Found :C 80.32 , H 10.08 ; C11H16 O requires C 80.44 , H 9.82 %)
PREPARATION OF MRS. 2SR. 6SR1 4.4-DlMETHYL-1-(PRQP-2-YNYU1- BICYCLOf4.1.01HEPTAN-2-QL
(46) (47)
To Zinc / Silver couple (4.32g , c. 72 mmol) was added allylic alcohol (46) (2 g, 12 mmol) and diiodomethane (1.76 ml, 72 mmol) in dry ether ( 40 m l). The mixture was refluxed under argon with vigorous stirring for 2 hours. On cooling to O0 C, sat. ammonium chloride solution ( 40 ml ) was added dropwise. The two phases were separated and the aqueous phase extracted with ether ( 2 x 20 m l).
147 The combined organic phases were dried over MgS0 4 before removal of solvents in vacuo. The residue was chromatographed ( silica, 40 % ether / petrol ) to afford (1RS, 2SR, 6SR) 4,4-DIMETHYL -1-(PROP-2-YNYL)]-BICYCLO [4.1.0] HEPTAN-2-OL( 1.68 g, 77% ) as a colourless oil . Vmax (film) 3306, 2950, 2117, 1465, 1043 cnr 1 ; §H (270 MHz , CDCI3 ) .31 ( 1H, t, J = 5 Hz, C-7 cyclopropyl-CH ) , .62 (1H, dd, J = 5, 9 Hz, C-7 cyclopropyl-CH ) , .82 (1H, m, C -6 -CH ) , .84 (3H , s, C-4 Me-) , .90 (3H, s, C-4 Me-) , 1.05- 1.20 ( 2 H, m, C-5 -CH2 ) , 1.50-1.72 ( 3H, m, C-3 -CH 2 & O-H ) , 2.02 (1H, t, J = 2.5 Hz, C-3' acetylenic CH) , 2.14 ( 1H, dd, J = 2.5, 17 Hz, C-1' -CH) , 2.57 ( 1H, dd, J = 2.5, 17 Hz, C-1* -CH ) 4.26(1H, m, C-2 HO-CH) ; m/z 178 (M+),163 ,112, 82. (Found : C 80.77 ,H 10.47 ; C i 2 H180 reqiures C 80.85, H 10.18 %)
PREPARATION OF (1RS. 2SR. 6SRM.4-DIMETHYL-1-f 3-f METHYL ACETATE )PROP-2-YNYL 1 BICYCLO T4.1.01HEPTAN-2-QL
(4 7 ) (4 9 )
Alcohol ( 47 ) (1.38 g , 7.7 mmol) and diisopropylamine ( 1.40 m l, 8.4 mmol) were dissolved in pyridine ( 20 m l) and DCM ( 20 m l), and
148 the mixture was cooled to 0° C under argon. Freshly distilled chloro- trimethylsilane ( 2.07 ml , 16.3 mmol ) was added and the mixture stirred at room temperature overnight. After dilution with petrol ( 250 m l) the mixture was washed sequentially with sat. K 2HPO4 solution ( 50 m l) and water (50 m l). The combined aq. washings were back washed with petrol (2 x 25 m l) and then the combined organic phases were washed with aq. NaHC 0 3 ( 50 m l) before being dried over Na2 S0 4 . After removal of solvents in vacuo the silyl ether (1.86g, 7.44 mmol) was dissolved in dry THF and cooled to -78° C under argon. n-BuLi ( 9.38 ml of a 1.6M solution in hexanes , 15 mmol ) was added dropwise and the reaction was left to stir at -78° C for 30 minutes. Methyl chloroformate ( 2.4 ml, 20 mmol ) was then added dropwise accompanied by the removal of the cooling bath and the reaction left to warm up to room temperature over 1 hour. 2M HCI (15 ml) was added with stirring for 2 0 minutes before neutralisation with aq. NaHC 0 3 . The mixture was extracted with ether (3 x 20 m l) and the combined organic phases were dried over MgS0 4 before removal of solvents in vacuo . The residue was chromatographed ( silica, 40 % ether / petrol ) to afford (1RS, 2SR, 6SR)-4,4-DIMETHYL-1-[ 3- (METHYLACETATE )PROP-2-YNYL ] BICYCLO [4.1.0] HEPTAN- 2-OL (1.51 g , 8 6 % ) as a white waxy soid ( m.p. 69° C ). Vmax (solvent :chloroform) 3376, 2951, 223, 1713, 1255 cm -1 ; 8 h (270 MH z , CDCI3 ) , .33 ( 1H, t, J = 5 Hz , C-7 cyclopropyl-CH ) , .63 ( 1H, dd, J = 5, 9 Hz , C-7 cyclopropyl -CH ) .76 ( 1H, m, C-6 -CH ) , .84 ( 3H, s , C-4 Me-) , .91 ( 3H, s, C-4 Me-) , 1.10-1.21 ( 2H, m, C5 -CH2 ) , 1.30 ( 1 H , br. d , J = 9 Hz, O-H ) , 1.55 (1 H, m, C-3 -CH ) 1.68 (1 H, m, C-3 -CH ) , 2.19 ( 1 H, d, J = 17.5 Hz, C-1’ propargyl - CH) , 2.88 ( 1H, d, J = 17.5 Hz, C-1' propargyl -CH ) , 3.75 ( 3H, s, C3' Me- ester) , 4.21 (1H, m, C-2 HO-CH).
149 (Found : C 71.40 , H 8.77 ; C 14H20O3 reqiures C 71.16, H 8.53% )
PREPARATION OF 0-fMRS. 2SR. 6SRM.4-DIMETHYL-1-f 3- ( METHYL ACETATE ) PROP-2-YNYL 1 BICYCLO T4.1.01HEPT-2-YL 1 1-IMIDAZOLE THIOCARBOXYLATE
S
As a mixture of two compounds ( 49 ) ( 53 )
Alcohol ( 49 ) ( 200 mg , .85 mmol) and N, N thiocarbonyldiimidazole (1 g , 5.5 mmol) were dissolved in dry DCM ( 25 m l) and refluxed under argon for 36 hours. After cooling , the mixture was diluted with more DCM ( 25 ml ) and washed sequentially with water, dilute aq. HCI, aq. NaHC 0 3 , water and brine before drying over MgS0 4 . Removal of solvents in vacuo afforded 0- [(1RS, 2SR, 6SR)-4,4-DI METHYL-1-[ 3- ( METHYLACETATE )PROP-2-YNYL ]BICYCLO [4.1.0]HEPT-2-YL ] 1-IMIDAZOLE THIOCARBOXYLATE ( 294 mg , quantiative yield) as a yellow oil. vmax («lm) 2999, 2237, 171, 1385, 1285, 1257, 970 cm ’ 1 ;
150 8h (270 MHz , CDCI 3 ) .38 ( .5H, t, J = 5 Hz, C-7 cyclopropyl-CH ), .69 .5H, t, J = 5 Hz, C-7 -CH) , .80-1.00 (2H, m, C -6 & C-7 cyclopropyl -CH ) , .85 (1.5H, s, C-4 gem Me-) , .89 (1.5H, s, C-4 gem Me-) , 1.03 (1.5H, s, C-4 gem Me-) , 1.06 (1.5H, s, C-4 gem Me-) , 1.25 ( 2 H, m, C-5 -CH2 ) , 1.83 ( 1.5H, m, C-3-CH2) , 1.97 ( .5H, m, C-3-CH2) , 2.2 1 ( .5H , d, J = 17Hz, C-1 ' propargyl -CH) , 2.38 ( .5H , d, J = 17 Hz, C-1 ' propargyl -CH ) , 2.59 ( .5H , d, J = 17 Hz, C-1'propargyl -CH ) , 2.94 ( .5H , d, J = 17Hz, C-1'propargyl -CH) , 3.67 ( 1.5H, s, C-4’ -OMe ) , 3.74 (1.5H, s, C-4' -OMe) , 4.37 ( .5H , m , C-7 O-CH) , 6.28 ( .5H, m, C-7 O-CH) , 7.02 ( .5H, dd, J = 1.5 , 0.9 Hz, imidazole N=CH ) , 7.09 ( .5H, dd, J = 1.5,0.9 Hz, imidazole N=CH ) , 7.44 ( .5H, t, J = 1.5Hz, imidazole N=CH ) , 7.64 ( .5H, t, J = 1.5 Hz, imidazole N=CH ) , 8.19 ( .5H, t , J = 0.9 Hz , imidazole N=CH ) , 8.34 ( .5H, t, J = 0.9 Hz , imidazole N=CH ) .
PREPARATION OF (E.ZVMETHYL-(2. 3. 3a. 4. 5. 6-HEXAHYDRO-5.5- DIMETHYL-1H-INDEN-2-YLIDENE 1 ACETATE
S
(53) (54)
The thiocarbonylimidazolide derivative ( 53 ) ( 260 mg, .75 mmol ) dissolved in dry benzene (15 m l) and brought to reflux under argon. Tri-n-butyltinhydride (.225 ml, .90 mmol) and AIBN (15 mg, .08 mmol)
151 in benzene ( 5 m l) were added dropwise to the refluxing solution over 2 hours. The mixture was then allowed to reflux for 5 hours. On cooling, CCI4 (5 m l) was added with stirring for 1 0 minutes before removal of solvents in vacuo. The residue was chromatographed ( silica, 5% ether / petrol ) to afford (E,Z)-METHYL-(2, 3, 3a, 4, 5, 6-HEXAHYDRO 5,5-DI METHYL-1 H.-INDEN-2-YLIDENE ) ACETATE ( 49 mg , 30 %) as a colourless oil. Vmax (film) 2949, 2239, 1713, 1656, 1208, 1123 cm“l ; 8h (500 MHz , CDCI3 ) .91 ( 3H, s, C-4 gem Me-) , .97 ( 3H, s, C-4 gem Me-) , 1 .6 8 ( 1 H, m, C-5 homoallylic -CH 2 ) , 1.76 ( 1 H, m, C-5 homoallylic -CH 2 ) , 1.89 ( 1 H, m, C-5 allylic -CH 2 ) , 2.15 ( 1 H, m, C-5 allylic -CH 2 ) , 2.44 (1.5H, m, C-6 allylic & C-7 allylic-CH) , 2.72 ( .5H, dd, J = 7.6, 16 Hz, C-7 allylic-CH 2 ) , 3.07 ( .5H, m, C-7 allylic -CH 2 ) , 3.21 ( .5H, m, C-7 allylic -CH 2 ) , 3.45 ( 2 H, br. s, C-9 allylic) , 3.68 (1.5H, s, C-2' -OMe) ,3.69 1.5H, s, C-2’ -OMe) , 5.40 ( .5H, m, C-2 alkenic -CH 2 ) , 5.44 ( .5H, m, C-2 alkenic -CH 2 ) , 5.77 ( .5H, m, C- 1 ' alkenic -CH 2 ) , 5.81 ( .5H, m, C-1 ' alkenic -CH 2 ); m /z 220 (M+) , 205 , 189 , 164 , 105 . (Found : C 76.17 , H 9.15 ; C 14H2o02 requires C 76.33 , H 9.15%)
152 HYDROLYSIS OF THE THIOCARBONYLIMIDAZOLIDE DERIVATIVE OF 3B-CHOLESTANOL
The thiocarbonylimidazolide derivative of 3p-cholestanol ( 1 g, 2 mmol ), was dissolved in THF ( 10ml) and this solution was added to 2M aq. HCI ( 1 0 ml) . The resultant mixture was heated with stirring at 50° C for 1 hour . On cooling , the mixture was extracted with ether (3 x 5 m l). The combined organic extracts were dried over MgS0 4 before removal of solvents in vacuo. The resultant yellow solid was recrystallised ( ether / methanol) afforded p-CHOLESTANOL ( 723 mg , 93%) as a white crystalline solid ( m.p. 138-139° C ) ( [a ]20 + 22.6° ) ( lit20). Vmax ( solvent: chloroform ) 3601, 3441, 2930, 1583, 1381, 1074 cm"1; SH (270 MHz, CDCI3 ) .63 (3H, s, C13Me-), .78 (3H , s , C19 Me-) , .80-1.98 (CH envelope), 3.58 (1 H , m, H-C-OH).
153 PREPARATION OF 0-fHRS. 2SR. 6SRV4.4-DIMETHYL-1-( PROP-2- YNYL1 BICYCLO T4.1.01HEPT-2-YL) 1-IMIDAZOLE THIOCARBOXYLATE
S OH
(47) (57 )
Alcohol ( 47 ) ( 50 mg, .28 mmol) and 1, 1' thiocarbonyldiimidazole (300 mg, 1.65 mmol) were dissolved in dry DCM (8 ml) and refluxed under argon for 18 hours. After cooling, the mixture was diluted with more DCM ( 15 m l) and washed sequentially with water, dilute aq. HCI, aq. NaHC 0 3 , water and brine before drying over MgS0 4 . Removal of solvents in vacuo afforded 0- [(1RS, 2SR, 6SR)-4,4-DI METHYL-1-( PROP-2-YNYL) BICYCLO [4.1.0]HEPT-2-YL ) 1- IMIDAZOLE THIOCARBOXYLATE (81 mg , quantitative yield ) as a yellow oil. Vmax (film) 3122 , 2951 , 2118 , 1688 , 1216 , 887 cnr1; 5H (270 MHz , CDCI 3 ) , .30 (1H, t, J = 5 Hz, C-7 cyclopropyl) , .72 (1H, dd, J = , 9 Hz , C-7 cyclopropyl) , .84 (3H, s, C-4 gem Me-) 1.03 ( 3H, s, C-4 gem Me- ) , 1.12-1.26 ( 3H, m C-5 -CH 2 & C-6 -C) , 1.72-1.84 ( 2 H, m, C-3 -CH2 ) , 1.93 ( 1 H, dd, J = 17, 2.7 Hz , C-1 propargyl-CH ), 2.02 ( 1 H, t, J = 2.7Hz, C-3' acetylenic CH ) , 2.82 1 H, dd, J = 17, 2.7Hz , C-1' propargyl-CH ), 4.42 ( 1 H, dd, J = 12, 6 Hz , C-10-CH) , 7.04 (1H, dd, J = 1.5,0.9 Hz, imidazole N=CH ) , 7.42 (1H, t , J = 1.5Hz , imidazole N=CH) , 8.15 ( 1H , t , J = 0.9
154 Hz , imidazole N=CH ).
ATTEMPTED HYDROLYSIS OF THIOCARBONYL1MIDAZOLIDE ( 53 )
S
R = H, disulphide dimer
(53 ) (56)
Thiocarbonylimidazolide derivative ( 53 ) ( 53 mg, .15 mmol ), was dissolved in THF ( 2 m l) and this solution was added to 2M aq. HCI (2 m l). The resultant mixture was heated with stirring at 50° C for 3 hours. On cooling, the mixture was extracted with DCM ( 3 x 1 m l). The combined organic extracts were dried over MgS0 4 before removal of solvents in vacuo to afford a yellow oil characterised as a mixture of (1RS, 2SR, 6SR)-4,4-DIMETHYL-1-[ 3-(METHYL ACETATE )- PROP-2-YNYL ] BICYCLO [4.1.0] HEPTAN-2-THIOL and its corresponding disulphide. vmax(film) 2951, 2925, 2236, 1713, 1255, 1034 cm " 1 ; 8h ( 270 MHz , CDCI 3 ) .22 (1, t , J = 5.3 Hz, C-7 cyclopropyl -CH ) .62 (1H, dd, J= 9,5.3 Hz, C-7 cyclopropyl-CH ) , .82 ( 3H, s, C-4 Me-) , .93 ( 3H, s, C-4 Me-) , 1.20-1.31 (2H, m, C-5 & C-6 -CH) , 1.37 ( 1 H, d, J = 8 .8 Hz, C-5 -CH ) , 1.70 ( 2H, m, C-3 -CH 2 ) , 2.04 ( 1 H, d, J = 17.5 Hz , C-1’ propargyl-CH) , 3.13 (1H, d, J = 17.5 Hz , C-1* propargyl-CH) , 3.35 ( 1 H, m, C-2 S-C-H ) , 3.75 (3H, s, C-4'
155 -OMe) ; m lz 502 ( M+ of disulphide) , 470 , 443 , 431 , 300 , 284 , 268 , 252 ( M+ of thiol) . ( Observed M+: 502.2200; C28H38O4 S2 requires 502.2211 )
PREPARATION OF MRS. 2SR. 6 SFP 4.4-DlMETHYL-1-(PROP-2-YNYLY|- BICYCLO [4.1,01 HEPTAN-2-THIOL
S
Thiocarbonylimidazolide derivative ( 57 ) (81 mg, .28 mmol ) was dissolved in THF ( 4 m l) and this solution was added to 2M aq. HCI (4 m l). The resultant mixture was heated at 50° C with stirring for 30 minutes . On cooling, the mixture was extracted with DCM (3 x 2.5 m l). The combined organic extracts were dried over MgS0 4 before removal of solvents in vacuo to afford (1RS, 2SR, 6 SR) 4,4-DI METHYL-1-(PROP-2-YNYL)]-BICYCLO [4.1.0] HEPTAN-2-THIOL as a yellow oil. Vmax (film) 3302, 2950, 2922, 2117, 1460, 637 cnr 1 ; 8h (270 MH z , CDCI3 ) .17 (1H, t, J = 5.3 Hz , C-7 cyclopropyl-CH ) .60 ( 1H, dd, J= 9,5.3 Hz, C-7 cyclopropyl-CH ) , .82 ( 3 H, s, C-Me-) , .93 ( 3H, s, C-4 Me-) , 1.10-1.25 ( 3H, , C-5 & C-6 -CH & S-H) , 1.37 ( 1 H, d, J = 8.5 Hz , C-5 -CH) , 1.74 ( 2 H, m, C-3 -CH2 ) ,
156 1.86 (1 H, dd, J = 17.5,2.7 Hz, C-1' propargyl -CH ) ,1.96 (1H, t, J= 2.7 Hz, C-3' acetylenic CH) , 2.96 ( 1 H, dd, J = 17.5,2.7 Hz , C-1* propargyl-CH) , 3.42 (1H, m, C-2 S-C-H ) ; m /z 194, 181, 161, 105, 98. ( Observed M+: 194.1130; C12H18S requires 194.1129 )
PREPARATION OF MRS. 6SR ) 4.4-DIMETHYL-1-r 3-(METHYL ACETATE^PROP-2-YNYL 1 BICYCLOtt. 1.0.1HEPTAN-2-ONE
(49) (5 9 )
Oxalyl chloride ( 167 pi, 1.9 mmol) was added to dry DCM ( 10 m l) under argon and cooled to -78° C. DMSO ( 274 pi, 3.8 mmol) was added dropwise and after stirring at -78° C for 10 minutes, alcohol ( 49 ) ( 380 mg, 1.6 mmol) in dry DCM (2ml) was added in one portion. After a further 20 minutes, triethylamine ( 800 ml, 5.78 mmol ) was added to the cloudy solution and the cooling bath removed. When the reaction had warmed up to room temperature, water (15 ml ) was added and the two layers separated. The aqueous phase was extracted with DCM ( 2 x 8 ml ) and the combined organic phases were dried over MgS0 4 before removal of solvents in vacuo. The residue was chromatographed ( silica, 40% ether / petrol ) to afford (1RS, 6 SR ) 4,4-DIMETHYL-1 -[ 3- (METHYLACETATE )-PROP-
157 2-YNYL ] BICYCL0[4. 1. 0. ]HEPTAN-2-ONE ( 357 mg , 94%) as a colourless oil. Vmax (film) 2955, 2237, 1710, 1689, 1257 cm " 1 ; 8h (270 MHz , CDCI 3 ) .89 ( 3H, s, C-4 gem Me-) , .94 ( 3H, s, C-4 gem Me-) , 1.03 (1H, t, J = 5.2Hz , C-7 cyclopropyl-CH) , 1.37 ( 1H, dd, J = 8.8 Hz , J = 5.2 Hz, C-7 cyclopropyl -CH ) , 1.56 ( 1H, d, J = 14 Hz , C-3 -CH) , 1.77 (1H , m , C-6 -CH ) , 1.85-1.94 (2H, m, C-5 -CH2 ) , 2.10 ( 1 H, d , J = 14Hz, C-3 -CH) , 2.64 ( 1 H, d, J = 18 Hz, C-1' -CH) , 2.94 ( 1H, d, J = 18 Hz , C-1' -CH ) , 3.71 ( 3H, s, C-4' -OMe); nv[z 234 (M+), 219, 204, 91, 41 . (Found : C 71.52 , H 7.73 ; C i 4 H180 3 requires: C 71.77, H 7.74%)
PREPARATION OF ( 2EZ. 3aRS. 7aSR ) 2. 3. 3a. 4. 5. 6 . 7. 7a. OCTAHYDRO-5.5-DlMETHYL-2-( METHYLACETATE 1 METHYLENE! -1H- INDEN-7-ONE
O O
(59) ( 62 )
Ketone ( 59 ) ( 1 0 0 mg , .42 mmol) , collidine ( 94 pi , .75 mmol) and 'proton free' chlorotrimethylsilane (3.18 ml , 25 mmol) were dissolved in dry THF (10 ml ). This solution was added to flame dried Zinc amalgam ( 2g , c. 30 mmol) and the mixture was refluxed under argon with vigorous stirring for 24 hours . On cooling sat. NH 4 CI solution
158 ( 20 ml) was added dropwise. The mixture was then extracted with DCM (3x15ml) and the combined organic phases were dried over MgS0 4 before removal of solvents in vacuo. The residue was chromatographed to afford ( 2EZ, 3aRS, 7aSR ) 2, 3, 3a, 4, 5, 6 , 7, 7a, OCTAHYDRO-5,5-DIMETHYL-2-( METHYLACETATE ) METHYLENE] -1H-INDEN-7-ONE as a white low melting point waxy solid . Vmax (film) 2951, 1710, 1657, 1207 cm*1 ; 8h (270MH z , CDCI3 ) .99 ( 1 .5 H, s, C-4 gem Me-) , 1.00 (1.5H, s, C-4 gem Me-) , 1.09 (1.5H, s, C-4 gem Me-) , 1.10 ( 1 .5 H, s, C-4 gem Me-) , 1.25-1.81 ( 2 H, m, C-5 -CH2 ) , 2.07 ( 1 H, m, C-6 -CH) , 2.44 (2H , m , C-3 -CH2 ) , 2.62-3.05 ( 4H, m, C-9 & C-7 allylic -CH 2 ) , 3.22 ( 1 H , m , C-1 -CH) , 3.77 (3H, s, -OMe) , 5.96 (1H, m, C-1' alkenic =CH) ; m /z 236 (M+) , 221, 204, 151, 119. (Found : C 71.01 , H 8.70 ; C 14H20O3 requires : C 7.16 , H 8.70%)
159 PREPARATION OF ( O CIS CARVEOL
Lithium Aluminium Hydride ( 6g, 158 mmol) was added to ether (500 m l) and the mixture cooled to -78° C under argon. Carvone ( 50 ml, 319 mmol) was then added dropwise over 45 minutes with stirring. After stirring at -78° C for a further 2 hours the reaction was allowed to warm up to room temperature. Water ( 200 ml ) was added, dropwise at first, and the phases were separated. The aqueous phase was extracted with ether (2 x 50 m l) and the combined organic phases were dried over MgS0 4 before removal of solvents in vacuo to afford (-) carveol (43.2g ,90%) as a mixture of isomers (c . 17:1 of cis : trans) . (-) Carveol (43.2 g , 284 mmol) was dissolved in triethylamine (100 m l) and DCM ( 300 ml). On cooling to O0 C , 4- dimethylaminopyridine ( 200 mg, 1.6 mmol) was added followed by 3, 5, dinitrobenzoylchloride (45g, 196 mmol). The mixture was then allowed to warm up to room temperature and stirred for 2 hours before washing sequentially with 2M aq. HCI and aq. NaHC 0 3 . The organic phase was dried over MgS0 4 before removal of solvents in vacuo . The residue was passed through a silica pad eluting with DCM . On removal of solvents in vacuo the resultant crystalline solid was recrystallised from ethanol to afford the 3, 5, dinitrobenzoylester of (-) cis carveol (56.88 g , 79 % ).
160 Potassium hydroxide (20.19 g, 360 mmol) was dissolved in water (50 ml ) and added to methanol ( 200 ml ) . The 3, 5, dinitrobenzoylester ( 50g, 144 mmol) was added and refluxed for 2 hours. On cooling the mixture was diluted with water (200 ml ) and then extracted with DCM (3x150 m l). The combined organic extracts were washed sequentially with 2M aq. HCI and aq. NaHC 0 3 before drying over MgS0 4 . After removal of solvents in vacuo the residue was distilled (85° C at 1 mm of Hg ) to afford (-) cis carveol as a colourless oil ( 18.89 g, 80 %) (lit53). Vmax (film) 3333, 2966, 2916, 1642, 1449, 1038, 889 cm-1 ; 5H ( 270 MHz , CDCI3 ) 1.51 (1H, m, C-4 -CH) , 1.61(1H, br., O-H) , 1.73 ( 3H, m, allylic Me- ) , 1.75 (3H, m, allylic Me-) , 1.93-2.27 ( 4H, m, C-4 -CH & C-5 -CH2 & C-6 -CH) , 4.18 ( 1 H, b, C-1 HO- CH) , 4.72 ( 2H, m, C-2' alkenic =CH 2 ) , 5.49 ( 1 H, m, C-3 alkenic =CH) .
PREPARATION OF (1S. 2R. 4R. 6 R 1 1 -METHYL-4-ISOPROPENYL BICYCLOr4.1.01HEPTAN-2-QL
T ( 66 )
To Zinc / Silver couple (19.68g, c. 324 mmol ) was added diiodomethane (14 ml, 162 mmol) in dry ether (100 ml). The mixture was refluxed under argon with vigorous stirring for 40 minutes. (-) cis
161 carveol (10 g, 6 6 mmol ) was added dropwise and refluxing continued for a further 2.5 hours. On cooling to 0° C, sat. ammonium chloride solution ( 100 ml ) was added dropwise. The two phases were separated and the aqueous phase extracted with ether ( 2 x 50 m l). The combined organic phases were dried over MgS0 4 before removal of solvents in vacuo. The residue was chromatographed ( silica, 30 % ether / petrol ) to afford (1S, 2R, 4R, 6R ) 1-METHYL-4-ISO PROPENYL BICYCLO[4.1.0]HEPTAN-2-OL ( 9.06 g , 83% ) as a white waxy solid ; [ a ] D20 -40.2° ( c = 1.0 , CHCI3 ) Vmax ( Solvent : chloroform ) 3591,3451, 2935,1640, 1033, 894 cm-1; 8 h ( 270 MHz , CDCI 3 ) .33 (1H , t , J = 5 Hz , C-7 cyclopropyl -CH ) , .44 1H , dd , J = 5, 9 Hz, C-7 cyclopropyl -CH ) , .84 ( 1 H, m, C-5 - CH) , .96 ( 1H, m, C -6 -CH ) , 1.18 ( 3H, s, C-1 Me-) , 1.43 ( 2H,m, O-H, C-5 -CH) , 1.67 ( 3H, m, C-1' allylic Me-) , 1.69- 1.95 ( 2H,m, C-3 -CH2 ) , 2.05 ( 1 H, m, C-4 allylic =CH ) , 3.91 ( 1 H, m, C-2 -CH) , 4.63 ( 2H, m, C-2' alkenic =CH2 ) ; m /z 166 (M +), 148, 107, 98, 69 . (Found : C 79.41 , H 11.13 ; CnH180 requires: C 79.47 , H 11.13%)
162 PREPARATION OF MS. 2R. 4R. 6R ^ 4-ACETYL-1-METHYL BICYCLO
14.1.01HEPTAN-2-QL
7 y ° H OH
4:
(66) (67)
Alcohol ( 6 6 ) ( 5 g, 30 mmol) was dissolved in methanol ( 200 m l), cooled to -78° C and a stream of O 2 / O3 was passed through the solution for 8 hours. After the solution had been purged of O 3 with a stream of O 2 only, triphenylphosphine ( 11.36 g, 43 mmol) was added and the mixture stirred at room temperature overnight. Solvents were removed in vacuo and the residue chromatographed ( silica, 50 % ether / petrol ) to afford ( 1S, 2R, 4R, 6 R ) 4-ACETYL-1-METHYL ] BICYCLO[4.1.0]HEPTAN-2-OL (5.95 g, 90%) as a colourless oil ; [a]D20 = - 6 0 .2 0 ( c = 1.0 , CHCI3 ). vmax (film) 3399, 2945, 1702, 1039 crrM ; 8h ( 270 MHz , CDCI 3 ) .36 ( 1 H, t , J = 5 Hz , C-7 cyclopropyl -CH ) , .49 (1H , dd , J = 5, 8 .8 Hz , C-7 cyclopropyl -CH ) , .83- 1.05 ( 2H, m, C-5 & C-6 -CH) , 1.98 ( 1H, ddt , J = 2.2, 5.5, 5.7 Hz, C-4 -CH ) , 2.11 ( 3H, , C-2' Me-) , 2.17 (1H, m, C-3 -CH ) , 2.31 ( 1 H, m, C-3 -CH ) , 4.93 ( 1H, m, C-2 -CH ) ; mZz 168 (M +), 107, 71, 43. ( Found : C 71.24 , H 9.66 ; C 10H16 O2 requires C 71.39 , H 9.59%)
163 PREPARATION OF MS. 2R. 4R( 2EZL6R n-METHYL-4-(ISO- PROPENYL- 2-ETHYL ESTER 1- BICYCLOr4.1.Q1HEPTAN-2-OL
( 67 ) ( 68 )
DME ( 20 ml ) was added to sodium hydride ( 260 mg of a 60 % dispersion in oil, 6.53 mmol ) under argon. Triethylphosphonoacetate (1.3ml, 6.53 mmol) was added dropwise with stirring over 20 minutes and after a further 40 minutes alcohol ( 67 ) (1 g, 5.97 mmol) was added. The mixture was then stirred at room temperature for 1 hour and at reflux for 1.5 hours. On cooling, water (150ml) was added and the mixture was extracted with ether ( 3 x 50 m l). The combined organic extracts were dried over MgS0 4 before removal of solvents in vacuo . The residue was chromatographed ( silica, 40 % ether / petrol ) to afford ( 1S, 2R, 4R( 2EZ), 6 R ) 1-METHYL-4-( ISOPROP ENYL-2- ETHYL ESTER )- BICYCLO[4.1.0]HEPTAN-2-OL ( 863 mg , 61 %) as a low melting point white waxy solid . vmax (film) 3411, 2934, 1710, 1639, 1212, 1150, 1038 cm-1 ; 5h (270 MHZ , CDCI 3 ) .25 ( 1/6 H, dd , J = 5, 9.1 Hz, C-7 cyclopropyl -CH of Z isomer) , .34 ( 5/eH, t, J = 5Hz, C-7 cyclopropyl -CH of E isomer) , .42 ( 5/6H, dd, J = 5, 9.1 Hz, C-7 cyclopropyl -CH of E isomer) , .52 ( 1/eH , t , J = 5Hz , C-7 cyclopropyl -CH of Z isomer) , 80-1.10 (3H, m, C-5 -CH2 &C-6-CH) ,1.15 (3H, s, C-1 Me-) , 1.21 ( 3H, t, J = 7 Hz , C-5' Me-) , 1.72 (1H, m, C-4 -CH) , 1.82
164 ( 1 H, b, 0-H) , 1.84-2.05 ( 2 H, m, C-3 -CH2) , 2.04 ( 5/6 H, d, J = 1Hz , C-1' allylic Me- of E isomer) , 2.08 ( 1/6 H , d, J = 1 Hz , C-1' allylic Me- of Z isomer) , 3.87 ( 5/6 H, m, C-2 -CH of E isomer) , 3.97 ( 1/6 H, m C-2 -CH of Z isomer ) , 4.09 ( 2H, q, J = 7 Hz, C-4' - CH2 ) , 5.56 ( 5/6 H, m , C-2' =CH of E isomer) , 5.62 ( 1/6 H, m, C-2' =CH of Z isomer) ; ( Found : C 70.27 , H 9.29 : C-i 4 ^ 3 0 3 requires : C 70.56 , H 9.30 %)
PREPARATION OF CM ( 1S. 2R. 4R( 2EZL 6 R ) 1-METHYL-4-f ISOPROP- ENYL-2-ETHYL ESTER VBICYCLO[4.1.01HEPT-2-YL1 1-IMIDAZOLE THIOCARBOXYLATE
(68) (69)
Alcohol ( 6 8 ) ( 500 mg, 2.1 mmol ) and N, N thiocarbonyldiimidazole (1.5 g , 8.2 mmol) were dissolved in dry DCM (25 ml) and refluxed under argon for 18 hours. After cooling , the mixture was diluted with more DCM (25 m l) and washed sequentially with water, dilute aq. HCI , aq. NaHC 0 3 , water and brine before drying over MgS0 4 . Removal of solvents in vacuo afforded 0-[ ( 1S, 2R, 4R( 2EZ), 6R )1-METHYL- 4- ( ISOPROP ENYL-2-ETHYL ESTER )- BICYCLO[4.1.0]HEPT-2- YL ] 1-IMIDAZOLE THIOCARBOXYLATE ( 660 mg, 91 % ).
165 Vmax ( film ) 3122, 2940, 2876, 1688, 1638, 1465, 1363, 1270, 1215, 1150, 886 cm-1 ; 6 h ( 270 MHz , CDCI3) .41 ( 5/6H, t, J = 5Hz , C-7 cyclopropyl -CH of E isomer) , .50 ( m, C-7 cyclopropyl -CH 2 of Z isomer) .64 ( 5/6H ,dd, J = 5, 9.1 Hz , C-7 cyclopropyl -CH of E isomer) , .80-1.15 ( 3H , m, C-5 -CH2 & C-6 -CH) , 1.23 (3H, s, C-1 Me-) , 1.25 (3H, t, J = 7.1 Hz, C-5' Me-) , 1.80-2.05 (3H, m, C-3 -CH2& C-4 -CH) , 2.10 ( 5/6H, d, J = 1Hz, C-1' allylic Me- of E isomer) , 2.16 ( 1/6H, d, J = 1Hz , C-1' allylic Me- of Z isomer) , 4.11 ( 1H, m, C-2 -CH of E isomer) 4.12 (2H,q, J = 7.1 Hz, C-4’ -CH2 ) , 5.59 ( 5/6 h , m, C-2' =CH of E isomer) , 5.60 ( 1/6H, m, C-2' =CH of Z isomer) ; 7.09 ( 1H, dd , J = 1.5, 1 Hz, imidazole N=CH ) , 7.46 (5/6H, pt ,J = 1.5 , imidazole N=CH of E isomer) , 7.48 ( 1/6H , pt , J = 1.5 , imidazole N=CH of Z isomer) 8.19 ( 5/6H, pt, J = 1 Hz, imidazole N=CH of E isomer ) , 8.21 ( 1/6H, pt, J = 1 Hz, imidazole N=CH of Z isomer ).
RADICAL REACTION ON THICARBONYLIMIDAZOLIDE (69)
( 69 ) ( 70 ) ( 71 )
The thiocarbonylimidazolide derivative ( 69 ) ( 650 mg , 1.86 mmol ) dissolved in dry benzene ( 20 m l) and brought to reflux under argon. Tri-n-butyltinhydride (540 ml , 2 mmol) and AIBN ( 42 mg , .25 mmol)
166 in benzene ( 1 0 ml) was added dropwise via a syringe pump to the refluxing solution over 6 hours. The mixture was then allowed to reflux for 6 hours. On cooling, CCI 4 (5 m l) was added with stirring for 1 0 minutes before removal of solvents in vacuo . The residue was chromatographed ( silica , 5% ether/petrol) to afford to major products. (4R ( 2E ), 6R ) 1,6 DIMETHYL-4-(ISOPROPENYL-2-ETHYL ESTER ) CYCLOHEX-1-ENE ( 95 mg , 23 % ). Vmax 2926, 1713, 1639, 1443, 1367, 1148 cm-1; SH (270 MHz, CDCI 3 ) , .80-1.10 (2H, m, -CH2 ) , 1.15 ( 2 H, d, J = 6 Hz, Me-) , 1.21 ( 3H, t, J = 7Hz, M£-CH2-C(0)0) , 1.45-2.05 (4H, m, 3 allylic protons &Me-C-H) , 2.11 (3H, br. s, allylic Me- ) , 2.17(3H, br.s, allylic Me- ), 4.09 (2H,q, J = 7Hz, Me-CHZ-C(Q10 ) , 5.02 (1H, m, alkenic =CH) , 5.64 ( 1H, m, alkenic =CH of a, p unsat. ester); m /z 222, 176, 149, 134, 94, 79. ( Observed M+ 222.1625; C 14H22O2 requires 222.1620) 7, 9-DIMETHYL-6-ETHYLESTER TRICYCLO[3 .3.1. 03’7] NONANES ( 198 mg , 48% ). vmax 2934, 2868, 1729, 1455, 1377, 1174 cm-1; 5H (270 MH z, CDCI3 ) .8 8 , .99, 1.02 ( 3H, d, J = 5 Hz, C-9 Me-), 1.25 ( 3H , t, J = 7 Hz, Me-CH 2-C(0)0 ), 1.26 (3H, s, Me-), 1.4-2.7 (11H, m ), 4.08 (2H , m, Me-CH2-C(0)0); m/z 2 2 2 , 207, 193, 176, 149, 93. ( Observed M+ 222.1623; C 14H22O2 requires 222.1621 )
167 PREPARATION OF 4-CHLORO-4-ISOPROPYL-1-( 2.2.2-TRICHLORO
ETHYL 1 CYCLOHEX-1-ENE
Sabinene ( 62 mg, .49 mmol ) was dissolved in carbon tetrachloride (1.5m l) containing benzoylperoxide (7 mg, .03 mmol) and one drop of d6-benzene . The solution was degassed with a stream of argon and then sealed in an nmr tube. The tube was heated at 80° C for 8 days, monitoring the progress of the reaction by 90 MHz nmr . On completion of the reaction, the tube was cooled, snapped open and the carbon tetrachloride removed in vacuo. Chromatography ( silica, light petroleum ) afforded 4-CHLORO-4-ISOPROPYL-1-( 2,2,2-TRI CHLOROETHYL ) CYCLOHEX-1-ENE ( 116 mg, 8 8 % ) as a colouless liquid and as the sole product . 8h (270 MH z , CDCI3 ) 1.04 ( 3H, d, J = 3.5Hz, isopropyl Me-) ,1.10 ( 3H, d, J = 3.5 Hz , isopropyl Me-) , 1.70 ( 1H, m, C -6 -CH ) , 1.89 ( 1H, m, C-6 -CH) , 2.12 ( 1 H, m, Me2CH_) , 2.23-2.72 ( 4H, m , C- 2 & C-5 CH2 ) , 3.35 (1 H, br. s, C-1 ’ -CH ) , 3.38 ( 1 H, br. s, C-1 ' - CH) , 5.73 (1H, m, C-3 alkenic =CH) .
168 PREPARATION OF 3-METHYL-1-(PHENYLTHlOMETHYU-3-ISOPROPYL CYCLOPENT-1 -ENE
major minor 6 1 (77) (78)
Sabinene (70 mg, .51 mmol), AIBN ( 2mg, .01 mmol) and thiophenol ( 47 pi, .46 mmol) were dissolved in bezene (1 m l) and the solution was degassed with a stream of argon and then sealed in an nmr tube. The tube was heated at 60° C for 2 hours . On cooling , the tube was snapped open and the benzene removed in vacuo. Chromatography ( silica, light petroleum ) afforded pedominantly 3- METHYL-1-(PHENYLTHIOMETHYL)-3-ISOPROPYLCYCLO PENT-1-ENE as a colouless liquid along with an amount (c. 15% ) of inseperable unopened cycloproane-thiophenol adduct ( 118 mg of both thiophenol adducts, 93 % ). Vmax (film) 2955, 2868, 1583, 1479, 1437, 737, 690 cirr 1 5H (270 MH z , CDCI3 ) .67 ( 3H, d, J = 3.5 Hz , Me- of isopropyl) .70 (3H, d, J = 3.5 Hz , Me- of isopropyl) , .80 ( 3H, s, C-1 Me-) , 1.41 ( 2H, m, C-5 CH2 ) , 1.72 ( 1 H, m, Me2CH ) , 2.28 ( 2H, m, C-4 allylic -CH2 ) , 3.51 ( 2H, s, C-1' -CH2-S ) , 5.23 ( 1 H, br. s , C-2 alkenic =CH) , 7.26 ( 5H, m, SPh ) ; m /s 246, 231, 218, 203, 93. (Observered M+: 246.1449; Ci6 H22S requires 246.1442)
169 PREPARATION OF 3-BROMQMETHYL-1-( 2.2.2-TRlCHLORQETHYL \
CYCLOPENT-1 -ENE
Sabinene (65 mg, .51 mmol) was dissolved in bromotrichloromethane ( 1 m l) containing benzoylperoxide (7 mg, .03 mmol) . The solution was degassed with a stream of argon and then sealed in an nmr tube before heating at 100° C overnight. On cooling, the tube was snapped open and the bromotrichloromethane removed in vacuo . Chromatography ( silica, light petroleum ) afforded 3-BROMO METHYL-1-( 2,2,2-TRICHLOROETHYL )CYCLOPENT-1-ENE (118 mg , 74 % ) as a colouless liquid . Vmax (film) 2959, 2871, 1724, 1463, 796, 707 cm-1 ; 8h ( 270 MHz , CDCI 3 ) .87 ( 3H, d, J = 3.4 Hz , Me- of isopropyl) .89 (3H, d, J = 3.4Hz , Me-of isopropyl) , 1.75 (1H, m, C-5CH) , 1.94 ( 1H, m, Me2CLL) , 2.04 ( 1H, m, C-5 -CH ) , 2.57 ( 2H, C-4 allylic CH2) , 3.47 ( 2H , d , J = .7 Hz ,C-1’ -CH2Br) , 3.50 (1H , d , J = 1 Hz, C-1' -CH2 ) , 3.53 (1H, d, J = 1 Hz , C-1' -CH2) , 5.60 ( 1H, m , C-2 alkenic =CH ) ; m / z ( with isotope patterns ) 334 ( m-H+ for 35ci, 80Br) , 333 ( m+ for^SQI, 80Br ) , 291 , 239 , 195 , 133 .
170 APPENDIX
171 7 7 s o
strsri
,is£s:.\
i2 d d 3IMZ03 till REFERENCES
1. P. A. Krusic, J. Am. Chem. Soc., 1977, 99, 250. 2. C. P. Casey and M. C. Cesa J. Am. Chem. Soc., 101, 4236. 3. B. Mailard, D. Forest and K. U. Ingold, J. Am. Chem. Soc., 98, 7024. 4. P. A. Blum, A. G. Davies, M. Pereyre and M. Patier J Chem. Res. (S), 1980. 5. E. C. Friedrich and R. L. Holmstead, J. Org. Chem., 1971, 36, 970. 6. W. G. Dauben, C. Shutte, R. E. Wolff and E. J. Denny, J. Org. Chem., 1969, 34, 2512. 7. P. Deslongchamps," Stereoelectronic Effects in Organic Chemistry", Pergammon Press, Oxford, 1983. 8. W. G. Dauben and R. E. Wolf, J. Org. Chem., 1970, 35, 2361. 9. The Total Synthesis of Natural Products, Ed. J. Apsimon, p265, John Wiley and Sons 1983. 10. A. P. Krapco Synth., 1974, 383. 11. G. H. Posner, Org. React., 1972, 19, 1972. 12. M. C. Pirrung, J. Am. Chem. Soc., 1981, 103, 82. 13. M. L. Quesada, R. H. Slessinger and W. H. Parsons, J. Org. Chem., 1978, 43, 3968. 14. G. Pattenden and S. J. Teague, J. Chem. Soc. Perkin Trans. 1988, 1, 1077. 15. R. D. Rieke and S. E. Bales Org. Synth. Coil. VoI 844. VI, 16. E. J. Corey and M. Chaykovsky, J. Am. Chem. Soc., 1965, 87, 1965. 17. H. E. Simmons and R. D. Smith, J. Am. Chem. Soc., 1959, 81, 4256. 18. K. E. Wilso, R. T. Seidner and S. Masamune, J. Chem. Soc., Chem. Commun., 1970, 213.
176 19. C. Girard and J. M. Conia, Tetrahedron 1974, Lett., 16, 3327. 20. D. H. R. Barton and S. W. McCombie, J. Chem. Soc. Perkin Trans. 1, 1975, 1574. 21. S. J. Zard and J. E. Forbes Tetrahedron 1989, Lett., 30, 4367. 22. D. Crich, Tetrahedron 1988, Lett., 29, 5805. 23. W. B. Motherwell, personal communication. 24. J. J. Fox, N. Miller and I. Wempen, J Med. Chem, 1966, 101. 25. N.. Simpkins Tetrahedron 1988, Lett., 29, 1315. 26. R. J. Rawson and I. T. Harrison, J. Org. Chem., 1970, 35, 2057. 27. H. E. Simmons, E. P. Blanchard and R. D. Smith, J. Am. Chem. Soc., 1964, 86, 1347. 28. J. M. Conia, Synth. Comm., 1973, 3, 261. 29. M. J. Robbins and J. S. Wilson, J. Am. Chem. Soc., 1981, 103, 932. 30. A. L. J. Beckwith, Tetrahedron, 1983, 10, 3056. 31. H. O. House, Modern Synthetic Reactions, 2nd Ed. 32. T. V. Rajanbabu and G. S. Reddy, J. Org. Chem., 1986, 51, 5458. 33. M. K. Kaloustian, M. I. A-L de Guterrez and R. B. Nade, J. Org. Chem., 1979, 44, 666. 34. The Chemistry of the Thiol Group, Parti, Ed. S. Patai, J. Wiley & Sons, 1974. 35. H. R. Nace, Org. React., 1962, 12, 57. 36. K. Harano, H. Kiyonaga, S. Sagimoro and T. Hirsano, Heterocycles, 1988, 27, 2327. 37. G. A. Wiley, R. L. Hershkowitch, B. M. rein and B. C. Chug, J. Am. Chem. Soc., 1964, 86, 964. 38. D. L. J. Clive and S. Daigneault, J. Chem. Soc., Chem. Commun., 1989, 333. 39. W. B. Motherwell, J. Chem. Soc., Chem. Commun., 1973, 935. 40. E. J. Corey and S. G. Pyne, Tetrahedron 1983, Lett., 24, 2821.
177 41. A. J. Manuso and D. Swern, Synth. 1981, 165. 42. G. Day, Ph. D. Thesis, London, 1989. 43. T. Mukaiyama, Angew. Chem. Int. Ed. Engl., 1977, 16, 817. 44. L. Garver, P. van Eikeren and J. E. Byrd J. Org. Chem., 1976, 41, 2773. 45. L. I. Smith, F. L. Greenwood and O. Hudrlik, Org. Synth. Coll. Vol, 3 1955, 673. 46. W. J. Wadsworth jr. Org. React., 1977, 25,73. 47. H. Stetter, Angew. Chem. Int. Ed Engl., 1962, 1, 286. 48. P. K. Freeman, M. F. Grostic and F. A. Raymond, J. Org. Chem., 1971, 36, 905. 49. J. A. Claisse, D. I. Davies and L. T. Parfitt, J. Chem. Soc., 1970,(C), 258. 50. R. A. Roberts, V. Schuli and L. A. Paquette, J. Org. Chem., 1983, 48, 2076. 51. O. Mitsunobu, Synth., 1981,1. 52. S. Bittner, Z. Barneis and S. Felix, Tetrahedron 1975, Lett., 44, 3871.
178 ADDENDUM p127 as a yellow oil. = ( 351 mg, quantitative ), as a yellow oil. p133 as a yellow oil. = ( 323mg, quantitative ), as a yellow oil. p155 and its corresponding disulphide. = and its corresponding disulphide (31 mg, 81%). p156 as a yellow oil. = ( 41 mg, 77% ), as a yellow oil. p159 as a white low melting point solid. = ( 42mg, 43%), as a white low melting point solid. p167 to afford to major products. = to afford two major products, both as colourless oils. p178 Reference 53 K. Narajan, P. K. Talkwalker, R. K. Shah, S. R. Mehta, G. V. Nayak, IndianJ.. Chem Sect(B), 1985, 24B(1), 98. Generation of Spirocyclic Quaternary Centresvia a Tandem Free Radical Cyclopropylcarbinyl Rearrangement-Cyclisation Strategy! J. D. Harling and W. B. Motherwell* Department of Chemistry, Imperial College of Science and Technology, South Kensington, London S W 7 2AY, U.K. Regio- and stereo-specific construction of spirocyclic quaternary centres may be achieved by hydroxy-directed Simmons-Smith cyclopropanation of an allylic alcohol followed by a tandem free radical cyclopropylcarbinyl rearrangement-cyclisation reaction; generation of the spiro-fused systems is subject to stereoelectronic and kinetic control.
Reprinted from the Journal of The Chemical Society Chemical Communications 1988 1380 J. CHEM. SOC., CHEM. COMMUN., 1988
Generation of Spirocyclic Quaternary Centresvia a Tandem Free Radical Cyclopropylcarbinyl Rearrangement-Cyclisation Strategyt J. D. Harling and W. B. MotherweH* Department of Chemistry, Imperial College of Science and Technology, South Kensington, London S W 7 2AY, U.K. Regio- and stereo-specific construction of spirocyclic quaternary centres may be achieved by hydroxy-directed Simmons-Smith cyclopropanation of an allylic alcohol followed by a tandem free radical cyclopropylcarbinyl rearrangement-cyclisation reaction; generation of the spiro-fused systems is subject to stereoelectronic and kinetic control. Within the last decade, new methods for the controlled centred radicals (1) or (2) (Scheme 1). The rapidity of production of carbon-centred free radicals1 have been coupled cyclopropane ring opening by an adjacent carbon-centred with intramolecular cyclisation reactions of predictable regio- radical has proved to be an extremely useful mechanistic and stereo-selectivity2 to provide a powerful technique forprobe.7 carbon-carbon bond formation in organic synthesis.3 A second feature of the design, as illustrated for the Although such substitution and addition reactions have been particular case of the spiro-fused exomethylene cyclopentane, extensively studied, preparative sequences incorporating anis that incorporation of the rearrangement in a rigid bicyclo- intermediate free radical rearrangement as part of the chain [x.1.0] system leads to a stereoelectronically controlled are much less common,4 particularly in all-carbon systems. cleavage of bond (a) to produce the higher energy primary Consideration of the relative rates of ring opening and radical (2), as opposed to the thermodynamically favoured reclosure of cyclopropylcarbinyl radicals and of hex-5-ynyl species (3) [bond (b)j. We now describe our preliminary radicals5 suggested to us that the elaboration of a tandem6 results on the construction of spiro[4.5]decanes which support rearrangement-cyclisation strategy should prove possible the foregoing analysis. under conditions of kinetic control without competing A suitable bicyclic precursor was readily assembled as hydrogen abstraction from stannane by intermediate carbon shown in Scheme 2. 1,2-Addition of 4-lithio-l- trimethylsilylbut-l-yne to the methyl enol ether (4) of dimedone gave, after aqueous acidic work-up, enone (5). t First presented at the Royal Society of Chemistry AutumnReduction Meeting. of the enone with di-isobutylaluminium hydride Nottingham, 24th September, 1987. followed* by hydroxy-directed Simmons-Smith J. CHEM. SOC., CHEM. COMMUN., 1988 1381
Scheme 2. Reagents: i, Me3Si-E-CH2CH2Li; ii, H30 + (Steps i and ii Scheme 3. Reagents: i, L-Selectride, aqueous work-up (65%); ii, 80% yield); iii, di-isobutylaluminium hydride (88%); iv, diethylazodicarboxylate, CH2I2) triphenylphosphine, benzoic acid (65%); iii, Zn/Ag couple (70%); v, thiocarbonyldi-imidazole, (100%);titanium vi, tetraisopropoxide, propan-2-ol (60%); iv, Zn/Ag couple, tri-n-butyltin hydride, azobisisobutyronitrile, benzene, (71%).CH2I2 [from (10) 74% yield from(12) 84% yield]; v, thiocarbonyldiimidazole, (100%); vi, tri-n-butyltin hydride, cyclopropanation furnished the bicyclic cyclopropyl carbinol azobisisobutyronitrile, benzene(13) [from79% yield, from(14) 81% (6). Finally, quantitative conversion to the corresponding yield], thiocarbonyl imidazolide derivative (7) yielded a suitable precursor9 for carbon-centred radical generation. necessary stereochemical marker (Scheme 3). Efficient Dropwise addition of tri-n-butylstannane to a refluxing chemo- and stereo-selective reduction was achieved through solution of (7) with addition of azobisisobutyronitrile (AIBN) use of L-Selectride, a feature of interest being that the major as initiator led smoothly to the desired spirocyclic system (8)allylic alcohol (10) was readily separated in pure form from (71%), thus confirming the regiospecific nature and the epimer (12) by simple aqueous work-up involving kinetically kinetic and stereoelectronic features of the proposed controlled hydrolysis of the intermediate borate esters. sequence. Inversion of configuration at the hydroxy centre(10) of was It was also of interest to demonstrate that this method can accomplished via Mitsunobu reaction10 followed by be used in a stereospecific manner. Accordingly, alkylation of titanium(iv) isopropoxide11 mediated solvolysis of the the kinetic lithium enolate of enone (5) with methyl iodide intermediate benzoate ester (11). The power and effectiveness gave the trimethyl analogue (9) (77%) containing theof the stereospecific hydroxy directed Simmons-Smith J. CHEM. SOC., CHEM. COMMUN., 1988 cyclopropanation sequence was demonstrated by the 2 A. L. J. Beckwith, Tetrahedron, 1981, 37, 3073. consistently high yields of cyclopropyl carbinols(13) [74% 3 For carbon-carbon bond forming radical chain reactionsinter see alia: D. J. Hart, Science (Washington C.), D. 1984, 223, 883: G. yield from (10)] and (14) [84% yield from (11)]. Stork, ‘Current Trends in Organic Synthesis,’ Pergamon Press, Transformation via reductive deoxygenation of their derived Oxford, 1983, 359; B. Giese, ‘Radicals in Organic Synthesis: thiocarbonyl imidazolides(13) and (14) with tri-n-butyl tin Formation of Carbon-Carbon Bonds,’ Pergamon Press, Oxford, hydride as described above, generated spirocycles(15) (79%) 1986. and (16) (81%) respectively which differ only in their relative 4 For radical-induced epoxide cleavage see: D. H. R. Barton, R. S. orientation of the vinyl silane moiety with respect to the Hay-Motherwell, and W. B. Motherwell,J. Chem. Soc., Perkin methyl group marker. Trans. 1,1981,2363; M. Cook, O. Hares, A. Johns, J. A. Murphy, Although a variety of connectivity patterns may be enviand C. W. Patterson,J. Chem. Soc., Chem. Commun., 1986, saged for the tandem rearrangement-cyclisation strategy, the 1419. 5 D. Griller and K. U. Ingold,Acc. Chem. Res., 1980, 13, 317. above examples in the spiro-mode are of particular interest 6 D. P. Curran and D. M. Rakiewicz,/.Am. Chem. Soc., 1985,107, inasmuch as they present a solution to the challenging problem 1448 and references therein. of stereospecific elaboration of a quaternary centre.12 7 J. W. Wilt, in ‘Free Radicals,’ ed. J. K. Kochi, Wiley-Interscience, We are grateful to the S.E.R.C., and to Quest International New York, 1973, vol. 1, ch. 8, p. 399; E. C. Friedrich and R. L. for the provision of a CASE studentship. Holmstead,J. Org. Chem., 1971, 36, 971. 8 H. E. Simmons, T. L. Cairns, S. A. Vladuchick, and C. M. Received, 13th June 1988; Com. 8I02376J Hoiness,Org. React. (N. Y.) 1973, 20, 1; R. E. Ireland and S. C. Welch, J. Am. Chem. Soc., 1970, 92, 7232. 9 D. H. R. Barton and S. W. McCombie,J. Chem. Soc., Perkin References Trans 1, 1975, 1574. 1 For recent reviews on the development of new free radical10 O. chainMitsunobu, Synthesis, 1981, 1. reactions seeinter alia: D. H. R. Barton and W. B. Motherwell,11 D. Seebach, E. Hungerbuhler, R. Naef, P. Schnurrenberger, B. Pure Appl. Chem., 1981, 53, 15; Heterocycles, 1984, 21, 1; T. Weidmann, and M. Zuger,Synthesis, 1982, 138. Ramaiah, Tetrahedron, 1987, 43, 3541. 12 S. F. Martin, Tetrahedron, 1980, 36, 419.