Cascade Radical Mediated Cyclisations Leading to Polycycle Constructions

Gerald Pattenden

Organic Chemistry Section, Department of Chemistry, Nottingham University, Nottingham NG7 2RD, England

Preamble. - The synthesis of ring systems has been both a preoccupation and a fascination for synthetic chemists since the earliest beginnings of modern organic chemistry. The fascination comes from the diverse and alluring molecular architecture of the ring systems, often associated with natural products, both carbo- and heterocyclic, with profound biological activities. In the past two decades a variety of cyclisation protocols, but particularly electrophilic polyene cyclisations,^[1] tandem transition metal-catalysed reactions,^[2] and pericyclic cycloadditions,^[3] have proved to be especially useful in the elaboration of polycyclic compounds. Another group of reactions that has played a dominant and determining role in the development of polycyclic ring synthesis during the past few years are free radical reactions, particularly so when they are carried out in an intramolecular and sequential ring forming fashion. Examples of the scope of these powerful radical cyclisation methods abound in the contemporary literature,^[4] including our own studies with allyl radical macrocyclisations,^[5] organocobalt reagents,^[6] tandem macrocyclisation - transannulation processes,^[7] and polycycle constructions involving oxy-radical fragmentations.^[8] There is no question that radical-mediated processes offer unique opportunities for the rapid, stereocontrolled synthesis of a wide range of highly functionalised polycyclic natural products, e.g. terpenes, steroids and alkaloids, under mild conditions.

Nature elaborates polycyclic terpenes and steroids by way of a controlled series of enzymic reactions triggered by carbonium ion intermediate formation, and many of these transformations can be mimicked in the laboratory, e.g. Johnson's synthesis of progesterone 2 from the polyene 1 ^[9] and the biomimetic synthesis of pentalenene 4 from the 5,8-ring fused diene 3.^[10]

Among many questions we have put to ourselves in exploring the scope for radical reactions in synthesis are the following:

can we use radical intermediates and mimic the terpene and steroid ring constructions found in Nature?
can we take a polyene-based radical system, pre-organise the chain, and promote polycycle construction either via a macrocyclisation - transannulation manifold (folding A) or by sequential endo-cyclisations (folding B)?

In this article we summarise our earliest results investigating and probing the scope for the two strategies towards polycycle construction, highlighted in Scheme 1.

Cascade Macrocyclisation - Transannulation Reactions. - In some of our first investigations, we examined the radical macrocyclisation - transannulation sequence involving the iodotrienone 5, with a view to a 'one-pot' synthesis of 1-decalone.^[11] Earlier work,^[12] based on the precedent set by the studies of Porter et al,^[13] had demonstrated the need for an electron deficient alkene electrophore, e.g. a conjugated enone, to promote macrocyclisations with nucleophilic radical centres. Thus, when a solution of 5 in benzene was heated in the presence of 1.1 equivs of Bu3SnH and a catalytic amount of AIBN for 0.5 h, workup and chromatography gave a 2:3 mixture of the cis-isomer 6 and trans-isomer 7 of 1-decalone, in a combined yield of 72%; treatment of this mixture with DBU (25°C, 24 hr) allowed the isolation of trans-1-decalone in essentially quantitative yield (Scheme 2).

By contrast, treatment of the positional isomer 8 of 5 with Bu₃SnH-AIBN led to a 1:1 mixture of trans-decalone 6 and cis-octahydroazuler-1-one 10 in a combined 68% yield, resulting from competitive 6-exo/5-exo transannulation from the intermediate cyclodecenone radical 9 (Scheme 3).

In further investigations of the scope for sequential radical macrocylisation - transannulations in bicycle constructions we showed that whereas the iodo-dienone 11 underwent tandem 9-endo-5-exo cyclisation producing the cis-tetralone 12 in reasonable yields (~50%), the corresponding E-octadienone 13 led only to the Z-cyclooctenone 14, and the iododienone 16 led to the 4-cyclopentyl substituted cyclohexanone (17; 95%), ie not to the anticipated 7,6-bicyclic ketone 15, on treatment with Bu₃SnH-AIBN.

15 to 17

In a novel, and certainly very different, strategy towards the tricyclo[9.3.1]pentadecane ring system in the taxanes, we have also achieved the radical macrocyclisation - transannulation sequence 18 -> 19 -> 20, in an unoptimised 25% yield.^[14] The use of an ynone electrophore in the initial macrocyclisation step in this sequence raises the overall yield of the corresponding tricycle to >60%, which bodes well for the future development of this strategy.^[15]

We next turned to an investigation of the elaboration of the 6,6,5-tricycle by way of a 13-endo trig macrocyclisation, ie 22 -> 23, followed by two successive 5-exo, 6-endo-trig transannulations, starting from the iodotrienone 21.^[16] When 21 was treated with Bu₃SnH-AIBN, under the usual conditions, work up and chromatography led to a single tricyclic ketone in 55% yield. However, spectroscopic data did not distinguish between the 6,6,5- and the 5,7,5-ring fused tricycles, 24 and 27 respectively. Accordingly we prepared the crystalline 2,4-DNP derivative of the product and recorded its X-ray crystal structure. This determination established unambiguously that the tricyclic produced from the cascade radical cyclisation of 21 was the cis, anti,trans 5,7,5-ring fused tricyclic ketone 27. ^[17] The tricycle 27 is produced from 21 via a sequential 13-endo-trig macrocyclisation, followed by two successive 5-exo-trig transannulation processes involving the radical intermediates 23 and 25/26. 21 to 24; 23 to 27

In other attempts to form tricycles from the cascade radical cyclisations of iodotrienone precursors we also examined the compounds 28 and 31 with an eye to the synthesis of the corresponding 5,6,5- and 5,5,5-ring tricycles, 30 and 33 respectively. To our surprise when 28 was treated with Bu₃SnH-AIBN instead of leading to 30 it gave a 3:1 mixture of diastereoisomers of the cyclopropane-cyclopentane (29; 53%), and the only product isolated from treatment of the iodotrienone 31 with Bu₃SnH-AIBN was the macrocyclisation product 32 in a meagre 16% yield.

In an attempt to rationalise the preliminary experimental results we have observed in the above sequential radical macrocyclisation - transannulation reactions, we have carried out some systematic MMC calculations.^[18] Indeed, a reasonably satisfactory rationale of the experimental results was forthcoming from these calculations, but unfortunately space does not allow a discussion of these calculations in this article.

Sequential 6-Endo Trigonal Cyclisations. - It is now thirty years since Breslow et al,^[19] and later Julia et al^[20] first examined the possibility of a free-radical mechanism for the oxidative cyclisation of squalene, and more recently this hypothesis has been re-visited by Snider et al^[21] and by Zoretic et al^[22] amongst others.^[23] The construction of fused polycycles by way of sequential radical mediated cyclisation reactions from alkyl centred radicals is well documented. Furthermore, with few exceptions 5-exo trig cyclisations are generally preferred over 6-endo trig closures from hex-5-en-1-yl radical intermediates,^[24] and attempts to use consecutive 6-endo trig cyclisations from (5-, 9-, 13-) polyolefinic alkyl radical precursors in the formation of linear and angular 6-ring fused constructions have met with failure.^[25] The unusual tendency of hex-5-en-oyl, i.e. acyl, radicals to cyclise via the 6-endo trig mode,^[26] leading to six-ring carbocycles, has prompted us to evaluate the consecutive cyclisations of a range of (5-, 9-, 13-, 17-) polyolefinic acyl radical intermediates with a view to the synthesis of linear and angular fused 6-ring systems include steroid constructions.^[27] We chose phenylselenyl esters as the most practical and convenient source of acyl radical intermediates.

We first examined the cyclisations of the acyl radical intermediates produced from the Z- and E-isomers of the diene selenyl ester 34.^[28] When solutions of the pure Z- and E-isomers of 34 were treated separately with Bu₃SnH-AIBN (reflux 8 h), each was found to undergo two consecutive 6-endo trig cyclisations leading to the trans-decalone 37 in 70-80% yield. The formation of a single diastereoisomer of a single regioisomer of 37 from either the Z- or E-isomer of 34 is significant, and is best rationalised on the basis of: i, rapid inversion of the stereochemistry of the beta-keto radical intermediate (35 to 36) prior to the second ring forming reaction, and ii, preference for formation of the most stable radical product in the second 6-endo cyclisation leading to 37 from 36. The importance of substitution on the 5- and the 9- double bonds in 34 in determining the regiochemical outcome of the bi-cyclisation leading to 37, was demonstrated by cyclisations of the related phenyl selenyl esters 38, 39 and 40, to 41, 42 and 43 respectively.

In studies similar to those carried out with 39 and 40 we also examined the cyclisations of the Z-dienoate 44 and the Z- and E-isomers of the phenyl ester 46. To our satisfaction cyclisation of 44 under the standard conditions led exclusively to the trans, anti, trans tricycle 45 in 72% yield, and reductive cyclisation of Z- or E- 46 gave a 1:1 mixture of diastereoisomers of the D-homosteroid 47 in 78% yield. 44 to 45; 46 to 47

Additional model studies, probing the influences of the stereochemistry and the alkyl group substitutions on the various olefinic bonds in the (5-, 9-, 13, 17) polyene selenyl ester cyclisations, finally led us to an investigation of the cascade radical cyclisations of the all-E polyene phenyl selenyl esters 48, 50 and 52. These three polyene esters were each found to undergo consecutive 6-endo trig cyclisations, in a remarkably regio- and stereo-selective fashion, leading to the ring-fused tricycle 49 and tetracycles 51 and 53, in very good yields (50-70%).

Each of the products 49, 51 and 53 was produced as a mixture of ring C or ring D methyl epimers; the full structure and stereochemistry of 49 were determined by X-ray measurements on the corresponding 2,4-dinitrophenylhydrazone derivative. The stereoselectives observed in these latter reactions are quite remarkable, and they offer enormous scope for the rapid construction of a range of linear and angular-fused ring systems with the opportunity of extensions into aza-steroids and interesting hetero-atom ring junction-substituted analogues. It is tempting to rationalise the observed stereospecifities in the cyclisations of the aforementioned polyene acyl radicals on the basis of fully concerted mechanisms, but our prejudice at this time would favour step-wise pathways for the sequential 6-endo-trig cyclisations.

Acknowledgements

It is a pleasure to have this opportunity to acknowledge my young colleagues, named in the references, who have contributed greatly to the ideas, and carried out all the experimental work, associated with the studies summarised here. Additional and sincere thanks are also extended to Dr. Chris Boden for converting this document into its current hypertext format. Finally, it is also a pleasure to thank Fisons Pharmaceuticals, Glaxo Group Research, Rhone-Poulenc Rorer, SmithKline Beecham, Pfizer Central Research and Zeneca, together with the SERC/EPSRC, for their financial support of our research.

References

[1] For recent reviews see: J.K. Sutherland "Polyene Cyclisations", in Comprehensive Organic Synthesis, Vol 3, 341, Ed. B.W. Trost, Pergamon Press, 1991; S.K. Taylor, Org. Prep. Proc. Int., 1992, 24, 247.

[2] e.g. N.E. Carpenter, D.J. Kucera and L.E. Overman, J. Org. Chem., 1989, 54, 5846; A. de Meijere, F.E. Meyer and P.J. Parsons, J. Org. Chem., 1991, 56, 6487; M.J. Dorrity, R. Grigg, J.F. Malone, V. Sridharan and S. Sukirthalingham, Tetrahedron Lett., 1990, 31, 1343; Y. Shi and B.M. Trost, J. Am. Chem. Soc., 1992, 114, 791.

[3] e.g. P. Deslonghcamps, Aldrichim Acta, 1991, 24, 43.

[4] For bibliography see: W.B. Motherwell and D. Crich, Free Radical Chain Reactions in Organic Synthesis, Academic Press, London, 1991; C.P. Jasperse, D.P. Curran and T.L. Fevig, Chem. Rev., 1991, 91, 1237.

[5] N.J.G. Cox, G. Pattenden and S.D. Mills, Tetrahedron Lett., 1989, 30, 621; N.J.G. Cox, S.D. Mills and G. Pattenden, J. Chem. Soc., Perkin Trans. 1, 1992, 1313; S.A. Hitchcock and G. Pattenden, Tetrahedron Lett., 1990, 31, 3641; S.A. Hitchcock and G. Pattenden, J. Chem. Soc., Perkin Trans. 1, 1992, 1323.

[6] See for example: A.Ali, D.C. Harrowven and G. Pattenden, Tetrahedron Lett., 1992, 33, 2851; G. Pattenden and S.J. Reynolds, J. Chem. Soc., Perkin Trans. 1, 1994, 379.

[7] G. Pattenden, A.J. Smithies and D. S. Walter, Tetrahedron Lett, 1994, 35, 2413.

[8] G.J. Hollingworth, G. Pattenden and D. J. Schulz, Aust. J. Chem., 1995, 48, 381.

[9] For recent reference and bibliography see: P.V. Fish, A.R. Sudhakar and W.S. Johnson, Tetrahedron Lett., 1993, 34, 7849.

[10] G. Pattenden and S.J. Teague, Tetrahedron Lett., 1984, 25, 3021; Tetrahedron, 1987, 43, 5637.

[11] For some preliminary results see reference 7.

[12] For some summary of earlier work see: G. Pattenden, 'Polycycle Constructions by Transition Metal Catalysed and Radical Mediated Processes' in Organometallic Reagents in Organic Synthesis, Academic Press, eds. J.H. Bateson and M.B. Mitchell, 1993.

[13] See: N.A. Porter, D.R. Magnin and B.T. Wright, J. Am. Chem. Soc., 1986, 108, 2787; N.A. Porter, V.H.-T. Chang, D.R. Magnin and B.T. Wright, J. Am. Chem. Soc., 1988, 110, 3554; N.A. Porter, B. Lacher, V.H.-T. Chang and D.R. Magnin, J. Am. Chem. Soc., 1989, 111, 8309.

[14] S.A. Hitchcock and G. Pattenden, Tetrahedron Lett., 1992, 33, 4843 (corrigendum Tetrahedron Lett., 1992, 33, 7448).

[15] D.W. Pryde, Unpublished work; Nottingham University.

[16] For some preliminary results see: M.J Begley, G. Pattenden, A.J. Smithies and D.S. Walter, Tetrahedron Lett., 1994, 35, 2417.

[17] For an example of a radical-mediated transannular strategy towards diterpenoid ring systems see: A.G. Myers, K.R. Condronski, J. Am. Chem. Soc., 1993, 115, 7926; idem, ibid, 1995, 117, 3057.

[18] A.L.J. Beckwith and C.H. Schiesser, Tetrahedron, 1985, 41, 3925; K.N. Houk and D.C. Spellmeyer, J. Org. Chem., 1987, 52, 959; K.N. Houk and J.L. Broeker, J. Org. Chem., 1991, 56, 3651.

[19] R. Breslow, E. Barrett and E. Mohacsi, Tetrahedron Lett., 1962, 1207; R. Breslow, S.S. Olin and J.T. Groves, Tetrahedron Lett., 1968, 1837.

[20] M. Julia, Tetrahedron Lett., 1973, 4464.

[21] M.A. Dombroski, S.A. Kates and B.B. Snider, J. Am. Chem. Soc., 1990, 112, 2759.

[22] P.A, Zoretic, X. Wang and M.L. Caspar, Tetrahedron Lett., 1991, 32, 4819; P.A. Zoretic, Z. Shen, M. Wang and A.A. Ribeiro, Tetrahedron Lett., 1995, 36, 2925; P.A. Zoretic, Y. Zhang, and A.A. Ribeiro, Tetrahedron Lett., 1995, 36, 2929.

[23] e.g. M. Hoffman, Y. Gao, B. Pandey, S. Klinge, K.D. Warzecha, C. Kruger, H.D. Roth and M. Demuth, J. Am. Chem. Soc., 1993, 115, 10358.

[24] D.C. Spellmeyer and K.N. Houk, J. Org. Chem., 1987, 52, 959; A.L.J. Beckwith, Tetrahedron, 1981, 37, 3073; A.L.J. Beckwith and C.H. Schiesser, Tetrahedron, 1985, 41, 3925; D.P. Curran, Synthesis, 1988, 417; D.P. Curran, Synthesis, 1988, 489; M. Julia, Acc. Chem. Res., 1971, 4, 386.

[25] cf. E.R. Lee, I. Lakomy, P. Bigler and R. Scheffold, Helv. Chim. Acta., 1991, 74, 146.

[26] D.J. Coveney, V.F. Patel, G. Pattenden and D.M. Thompson, J. Chem. Soc., Perkin Trans. 1, 1990, 2721; T.M. Patrick, Jr., J. Org. Chem., 1952, 17, 1009, 1269; R. Dulou, Y. Chretien-Bessiere and H. Desalbres, C.R. Acad. Sci., Ser. C, 1964, 258, 603; J-P. Montheard, ibid., 1965, 260, 577; M. Chatzopoulos and J-P. Montheard, ibid., 1975, 280, 29; J.A. Kampmeier, S.H. Harris and D.K. Wedegaertner, J. Org. Chem., 1980, 45, 315; M. Julia and M. Maumy, Bull. Soc. Chim. Fr., 1969, 2415; Z. Cekovic, Tetrahedron Lett., 1972, 749; E.J. Walsh, Jr., J.M. Messinger II, D.A. Grudoski and C.A. Allchin, Tetrahedron Lett., 1980, 21, 4409; P. Delduc, C. Tailham and S.Z. Zard, J. Chem. Soc., Chem. Comm., 1988, 308; D.L. Boger and R.J. Mathvink, J. Org. Chem., 1988, 53, 3377; D.L. Boger and R.J. Mathvink, J. Org. Chem., 1989, 54, 1779; D.L. Boger and R.J. Mathvink, J. Am. Chem. Soc., 1990, 112, 4003; D.L. Boger and R.J. Mathvink, J. Am. Chem. Soc., 1990, 112, 4008; D.L. Boger and R.J. Mathvink, J. Org. Chem., 1990, 55, 5442; D.L. Boger and R.J. Mathvink, J. Org. Chem., 1992, 57, 1429; D. Crich and S.M. Fortt, Tetrahedron Lett., 1987, 28, 2895; D.Crich and S.M. Fortt, Tetrahedron Lett., 1988, 29, 2585; D. Crich and S.M. Fortt, Tetrahedron, 1989, 45, 6581; D. Crich, K.A. Eustace and T.J. Ritchie, Heterocycles, 1989, 28, 67; D. Batty, D. Crich and S.M. Fortt, J. Chem. Soc., Chem. Commun., 1989, 1366; f) D. Batty, D. Crich and S.M. Fortt, J. Chem. Soc., Perkin Trans., 1, 1990, 2875; D. Crich, K.A. Eustace, S.M. Fortt and T.J. Ritchie, Tetrahedron, 1990, 46, 2135; D. Batty, D. Crich, Tetrahedron Lett., 1992, 33, 875.

[27] Both Boger et al and Crich et al have also examined aspects of tandem cyclisations of acyl radicals produced from certain unsaturated selenyl esters. See under reference 26.

[28] For preliminary results see: a) L. Chen, G.B. Gill and G. Pattenden, Tetrahedron Lett., 1994, 35, 2593. b) H. Simonian, Unpublished work; Nottingham University.