4 Reaction Mechanisms Part (iii) Free- Radical Reactions By S.A. HEWLINS and J.A. MURPHY Department of Chemistry University of Nottingham Nottingham NG7 2RD UK 1 Oxidative Initiation/Termination of Radical Chemistry The advantages of performing radical reactions without loss of functionality have led to numerous investigations of both oxidative and reductive terminations of free radical reactions. Now results of attempts to effect stereochemical control in manganese(rI1) induced cyclization'"Yb of dicarbonyl derivatives (la-) inter alia are reported." Of these substrates (la) gave the best results with a 90% yield of (2a) formed with 86% diastereomeric excess. Notably (1b) did not react under these conditions. 00 XivI *% Studies on regiochemical control have also appeared from Snider's laboratories.The regiochemistry of these cyclizations is very sensitive to the substituents on the alkene. Thus the chloroalkene (3) directs the exclusive formation2 of the endo-cyclization (a) B. B. Snider and Q. Zhang Tetrahedron Lett. 1992,33,5921; (b)P. A. Zoretic X. Weng C. K. Biggers M. S. Biggers and M. L. Caspar Tetrahedron Lett. 1992 33 2637. B. B. Snider Q. Zhang and M. A. Dombroski J. Org. Chem. 1992 57 4195. 75 S. A. Hewlins and J. A. Murphy intermediate (4) while the isomer (5) causes exclusive exo-cyclization to (6). Whereas alkyl substituents have also been found to control regiochemistry the easy displace- ment of the chlorine group makes it uniquely useful. A study of intermolecular reactions between dicarbonyl radicals and haloalkenes shows that the chlorine acts by decelerating the addition of a radical to the atom bearing the chlorine rather than by accelerating the addition to the neighbouring atom.0 C02Me Cl c1 c1 -products GHCl (6) The use of chloroalkenes is also advantageous in other ways.3 In attempts to form the tricyclic ring system (8) the reactions of the alkynes (7) with Mn"'/Cu" were explored. In the terminally unsubstituted alkyne (7a) no desired product was seen. With the silyl substituted alkyne (7b) only the dimer (9) was obtained. The desired product is thought to form but the speed of desilylation was so rapid that it left (8 R = H) exposed to the oxidizing conditions for a sufficient length of time to produce the oxidatively coupled product (9).Attempts to slow the desilylation by buffering the reaction were unsuccessful. However use of the chloroalkene (10) produced (8a) in 86% via the chlorotricycle (11). The method was subsequently extended to the preparation of the dimethoxy derivative (8b) which bears the aromatic substitution pattern present in the aureolic acid antibiotics. ' B. B. Snider and T. Kwon J. Org. Chem. 1992 57 2399. Reaction Mechanisms -Part (iii) Free-Radical Reactions (8) a R = H b R=OMe Extending this oxidative chemistry to molecules other than fl-dicarbonyl com- pounds is an obviously worthwhile goal and several investigations have moved in this direction. Snider,3 Matta~,~ and Lopez' have studied the oxidative radical chemistry of enol ethers of monocarbonyl compounds but from different perspectives.In Snider's case the aim was initially to see if enolyl radicals (i.e. radicals derived from enols themselves) could cyclize onto alkenes. This reaction could not compete with intermolecular couplings but oxidation of silyl enol ethers such as (12) was successful. Cyclizations to cyclopentanones or cyclohexanones (15)were observed with appropri-ate substrates but the phenyl group is crucial to the success of the reaction. From clever mechanistic investigations Snider has concluded that it is the cation radicals such as (13)and (14)which are performing the cyclizations. Although aminium salts have also been studied in this work oxidation with copper salts and ceric ammonium nitrate (CAN)are preferred.Mattay has investigated a similar reaction but with photochemi- cal initiation of the reaction by electron transfer to the excited singlet state of 9,lO-dicyanoanthracene. In this case it is not necessary to use phenyl ketones. 0SiBu'Me~ OSiBu'MeZ OSiBu'Mez I I I (12) n= 1,2 t . \ A. Heidbreder and J. Mattay Tetruhedron Lett. 1992 33 1973. L. Lopez and L. Troisi Tetrahedron 1992 48 7321. S. A. Hewlins and J. A. Murphy Lopez however has studied the reaction of vinyl ethers (16) with dioxygen to form dioxetanes (17) catalysed at -78 "C in dichloromethane with tris-p-bromo-phenylaminium hexachloroantimonates. The dioxetanes form in good yield. In similar chemistry,6 but induced by ceric salts enamines (18) have been coupled to silyl enol ethers (19)in good yield.The enamine is thought to be converted into its radical cation to trigger this process; notably however the chemistry also works on the less electron rich enamines (20). Aminium salts have been extensively used in the pioneering studies of Bauld on the cation-radical cycloaddition reactions mentioned in previous reports. This work is now being extended7 with kinetic studies to see how fast such intramolecular cyclizations occur. A number of probes were synthesized and their cyclization rate constants determined. For example the rate constant for the cyclobutanation of (21)to (22) via cation radical chemistry is 23 x 109s-'. This probe was then used to investigate the mechanism of metalloporphyrin-induced epoxidations of alkenes which have previously been reported to occur via the intermediacy of cation radicals.The fact that no cyclobutanation was seen in the epoxidation of (21) allowed Bauld to state a maximum lifetime for a free radical cation in this reaction. According to his calculations a cation-radical process with a decay rate-constant up to 6 x 10' s-would have been detectable and concludes that cation-radicals are not intermediates in these epoxidations. On the other hand cation radicals are proposed as intermediates in a novel oxidative cleavage of carbon-tin bonds developed by Hanessiam8 Here trimethyltin radicals add to an alkene (23) triggering a 5-exo cyclization. The product (24) is oxidized with K.Narasaka T. Okauchi K. Tanaka and M. Murakami Chem. Lett. 1992 2099. ' G.A. Mirafzal T. Kim J. Liu and N. L. Bauld J. Am. Chem. SOC. 1992 114 10968. a S. Hanessian and R. Leger J. Am. Chem. SOC. 1992 114 31 15. Reaction Mechanisms -Part (iii) Free-Radical Reactions ceric ammonium nitrate in methanol giving a dimethyl acetal (25). The mechanism of this transformation is speculative but it is proposed that the tin-carbon bond in (24) fragments after formation of the radical cation to give a carbon radical which is trapped by ceric ammonium nitrate. Hanessian proposes that this nitrate undergoes (presumably heterolytic) fission giving an aldehyde which is then converted under the conditions of the reaction to the dimethyl acetal.Mention must be made of a radical cation with 'non-classical cation' character.' The benzhydrylidenenorbornene (26) behaves as an electron donor to an excited state of phenanthrene. The resulting radical cation (27) is trapped with complete stereoselectiv- ity by methanol ultimately yielding (28). The authors propose that this selectivity is due to a strong homoconjugative interaction between the 3n electrons of the radical cation. $ Me3SnvPh MeOvPh OMe \/ -MesSnCl. NaBH3CN AlBN EtO2C C02Et Et02C CO2Et EtO2C CO2Et 2 Reductive Initiation/Termination of Radical Chemistry Electrochemical reductive coupling between ketones and nitriles has been announced by Shono." The chemistry works both intra- and intermolecularly. In an elegant demonstration of its potential he has synthesized guaiazulene (29) ( -)-valeranone (30) polyquinanes e.g.(31) methyldihydrojasmonate (32) and rosaprostol (33). T. Hitano S. Shiina and M. Ohashi J. Chem. SOC..Chem. Commun. 1992 1544. lo T. Shono N. Kise T. Fujimoto N. Tominaga and H. Morita J. Org. Chem. 1992 57 7175. S. A. Hewlins and J. A. Murphy The reactions of samarium iodide continue to excite interest. A Japanese report' details the reductive coupling of N,N-disubstituted amides (34) to form 1,2-diarninoal- kenes (35) using a mixture of samarium and samarium iodide. It is proposed that the reduction starts with formation of the radical anion of the amide. Stereospecific trapping of the reactive intermediate by an alkene suggests that a carbene is formed on further reduction of this intermediate.Both samarium and samarium iodide are required so this is a very potent reducing agent. I 0 The intramolecular reactions of ketyls with alkenes have been previously reported however a current study12 shows strange events can take place in certain such reactions. Thus the products of reaction of substrates (36) and (37) include (38) and (40) resulting from trapping of ethene. The ethene arose from the original generation of samarium iodide from reaction of samarium metal with diiodoethane. These products were not seen when the solution was degassed after generation of samarium iodide. Also seen in the reaction of (36) is 1-methylcyclooctanol (39)from a curious 8-endo trig cyclization. A range of electrophiles has been tested with the organosamarium intermediates generated by this method.Curran has also treated' 3,14 organosamarium reagents with a range of electrophiles to test the scope of their chemistry. In this case the samarium reagents were generated from iodides such as (42). The organosamarium l1 A. Ogawa N. Takami M. Sekiguchi I. Ryu N. Kambe and N. Sonoda J. Am. Chem.SOC.,1992,114,8729. l2 G.A. Molander and J.A. McKie J. Org. Chem. 1992 57 3132. l3 M. J. Totleben D. P. Curran and P. Wipf J. Org. Chem. 1992 57 1740. l4 D. P. Curran and M. J. Totleben J. Am. Chem. SOC. 1992 114 6050. Reaction Mechanisms -Part (iii) Free-Radical Reactions compounds (43) could be converted into organ~copper'~ compounds (44)which are capable of performing conjugate additions in the manner of more traditional cuprates.[~-~CuSmX12]-cux (44) 3 Atom Transfer Reactions Linking into the theme of the previous section remote functionalization has been achieved by treating suitable iodoarenes (45) with samarium iodide.' Radical formation followed by hydrogen atom abstraction occurs prior to trapping by a second equivalent of samarium iodide; the organosamarium intermediate (46) was then successfully reacted with ketones isocyanates and isonitriles. M. Murakarni M. Hayashi and Y. Ito J. Org. Chern. 1992 57 793. S. A. Hewlins and J. A. Murphy f f f &I (LjJ"Y s'-bNL2 0 -0 -0 -0 (45) (46) The above reactions involve the expected 1,Shydrogen atom transfer to aryl radicals. However a number of investigations have been launched to determine the detailed regioselectivity of such abstractions.Thus naphthyl amides such as (47) have been subjectedI6 to reaction with tributyltin deuteride and AIBN. Extensive amounts of 1,6 and 1,7-hydrogen atom transfer occurred as judged by the sites of deuteration in the products (48).This is curious since earlier studies had demonstrated an absence of 1,6 and 1,7-transfers in the amines (49). Curran has sought to functionalize by 1 $transfer from cleavable protecting groups. Hence the silyl ethers (50) and (51) gave F3cY0 F3cY0 -I I (1.5) 40% 1.6) 30% (1,7) 10% (1.7) (13 (47) (48) H R R' 24% 65% I Bu3SnD ____) Ph-OSiRR'Ph I 11% (49) Ph (50) R,R' = Me (51) R,R' = (CH2)4 deuterated products on treatment' with tributyltin deuteride and AIBN.Again the regioselectivity was only moderate. The figures shown represent the extent of deuterium incorporation starting from (50). In a different application" of hydrogen atom transfer the bromotrityl ether was used as a self-oxidizing protecting group. This has particular uses; for example selective oxidation of the primary alcohol was achieved in (52). l6 D. Denenmark T. Winkler A. Waldner and A. De Mesmaeker Tetrnhedron Lett. 1992 33 3613. D. P. Curran K.V. Somayajula and H. Yu Tetrahedron Lett. 1992 33 2295. Is D. P. Curran and H. Yu Synthesis 1992 123. Reaction Mechanisms -Part (iii) Free-Radical Reactions Parsonslg has used 1,5-hydrogen atom abstraction of an allylic hydrogen followed by cyclization for the construction of hydrindanes (54).This full paper details the benefits of starting with vinyl iodides and having an electron-poor alkene as in (53),as the radical acceptor in the final step.A fresh example of the rare 1,4-hydrogen atom abstraction is witnessed in the complicated chemistry of the vinyl radical (55)studied by Malacria.20 Thus following the transfer ring expansion results giving (56). Both alkenes (57)and (58)formed in this case are stereochemically pure. i BySnH c ii. Hz% HO Hok (57) (55) (56) A most unusual regiochemistry was observed in some hydrogen atom abstraction reactions studied by Kraus2' Substrate (59) on photoexcitation underwent 1,9-I' A. D. Borthwick S.Caddick and P.J. Parsons Tetrahedron 1992,48 10655. 2o M. Journet and M. Malacria Tetrahedron Lett. 1992 33 1893. S. A. Hewlins and J. A. Murphy hydrogen atom abstraction. Although it might immediately be supposed that the product results from two sequential 1 ,Sabstractions labelling studies have disproved this. This study on a-ketoesters raises a number of interesting questions particularly concerning the rates of the competing processes. Thus the cyclopropane (60)is cleanly transformed via a 1,5-transfer into (61) indicating that the fragmentation of the acyl-oxygen bond is extremely rapid. Changing from hydrogen transfers to iodine atom transfers an intriguing revelation has been made about the differences between the atom-transfer additions to alkenes of iodomalonates and iodomalononitriles.It has previously been reported that iodo- malonate (62) undergoes addition to alkenes when irradiated in the presence of hexabutylditin giving the cyclic product (63).However the corresponding annulation with the dinitrile (64)went in very poor yield,22 with the iodide (65) being isolated as the principal product of a dirty reaction. Omitting the ditin however and using thermal activation in the dark for prolonged reaction times led to good yields of the desired cyclic compound (66).Evidence suggests that this is indeed a radical process. It is proposed that the reason that the reaction does not proceed well with irradiation is that photochemical activation generates too high a concentration of reactive inter- mediates which react with each other rather than with the alkene.The hexabutylditin is thought to thwart successful reaction by reacting directly with the iodomalononitrile -by a polar mechanism. The dinitrile reaction appears to generate iodine from the start < + hv. A fBU %Sn2 * Me02C Me02C C02Me C02Me '' G. A. Kraus and Y. Wu J. Am. Chem. SOC. 1992 114 8705. 22 D. P. Curran and C. M. Seong Tetrahedron 1992 48 2157. Reaction Mechanisms -Part (iii) Free-Radical Reactions of the reaction and it is possible that iodine (as well as iodomalononitrile) may act as a very efficient quencher for the intermediate radical (67) before cyclization can occur. The dinitrile radical successfully adds to di- and trisubstituted alkenes unlike the diester radical.This considerably extends the synthetic potential of electrophilic radicals. Bu NCGBU-'b CN NCQBu CN NC CN Ally1 iodomalononitriles23 also lead to two stage annulation reactions. However in this case a nitrile transfer can occur. The geometric constraints imposed on intermediates in these reactions can lead to remarkable control of stereochemistry. For example cyclization of (68) with cyclopentene leads to the single isomer (69) being formed in 68% yield. The only other isolated product is the reduced bicyclic product (70). Starting from two simple achiral precursors five stereogenic centres have been constructed with excellent selectivity! NC CN 'Cr> NcY Nd NC A variant of these atom-transfer ideas is seen in the one-pot syntheses of azabicyclo~ctanes.~~ Here the initial radical cyclization of (71) is coupled with a subsequent polar displacement on the intermediate (72) giving good yields of the bicyclic products (73).These products have been used to develop agonists at a newly discovered serotonin receptor. 23 D. P. Curran and C. M. Seong Tetrahedron 1992. 48 2175. 24 D. L. Flynn D. P. Becker R. Nosal and D. L. Zabrowski. Tetrahedron Lett.. 1992 33 7283; D. L. Flynn. D. L. Zabrowski and R. Nod ihid. 1992 33 7281. S. A. Hewlins and J. A. Murphy H I + eNR BbSn2 &-IQ hv t Me02C C02Me Me02C C02Me Me02C C02Me 4 Large and Small Rings Large Rings.-Three communications and one full paper have appeared this year describing reactions which take advantage of the reluctance of amides and esters to adopt suitable conformations for cyclizations to 5- and 6-membered rings.The slowing of this process allows cyclization to larger rings to compete. Amide25 (74) has been used for a 10-endo dig cyclization in a very novel approach to isoquinoline alkaloid synthesis. Expoxidation with rn-chloroperoxybenzoic acid followed by acid treatment led via the ketoamide (79 to the tetracycle (76). H WNV0 n (74) li mCPBA 14-to 16-endo cyclizations onto propiolate alkynes are witnessed in the work of Baldwin.26 Whereas in many cases direct reduction of the halide competes with the cyclization a 63% yield of (78) was obtained from the iodoester (77). This method does not produce smaller rings.A cuprous chloride/2,2’-bipyridine complex has been used2’ to trigger formation of 8-membered ring lactones e.g. (79) by an atom-transfer reaction in moderate to excellent yields. The authors suggest that the terminal alkene may be coordinated to the copper during the reaction favouring the reaction by a template effect. 25 C. Lamas C. Saa L. Castedo and D. Dominguez Tetrahedron Lett. 1992 33 5653. 26 J. E. Baldwin R. M. Adlington and S. H. Ramcharitar Tetrahedron 1992 48 3413. 27 F.O.H. Pirrung W. J. M. Steeman H. Hiemstra W. N. Speckamp B. Kaptein W. H. J. Boesten H. E. Schoemaker and J. Kamphuis Tetrahedron Lett. 1992 33 5141. Reaction Mechanisms -Part (iii) Free-Radical Reactions 0 (77) When monobromoacetates e.g. (80),were treated with tributyltin hydride a similar 8-end0 cyclization28 to (82) was found.Direct reduction product (81) was also observed. One remarkable observation was the isolation of 25% of lactone (84) from the reaction of (83).The mechanism proposed for the formation of this bicyclic lactone features a carbon radical attack upon a lactone carbonyl group as shown. Whereas attack on ketone and aldehyde carbonyl groups are now common this reaction at a lactone carbonyl is to our knowledge unique. The proposed conformations required for 5-exo (85) and 8-end0 (86)attack for radical (83) are shown. Once again it is the reluctance to adopt the s-cis conformation that leads to the eight-membered rings. The attack by a carbon radical on a ketone carbonyl group to give an alkoxyl radical (87) that then fragments with cleavage of the ring-junction to leave a larger ring continues to be e~plored.~~-~~ Dowd long an exponent of this chemistry has reported the fragmentation of radicals derived from spiro- and fused-cycl~butanones~~*~~ such as (88) and (89).The fused compounds are easily synthesized by photochemical ” E. Lee C. H. Yoon and T. H. Lee J. Am. Chem. Soc. 1992 114 10981. 29 P. Dowd and S.-C. Choi Tetrahedron 1992 48 4773. 30 J. E. Baldwin R. M. Adlington and R. Singh Tetrahedron 1992 48 3385. 31 W. R. Bowman and P.J. Westlake Tetrahedron 1992 48 4027. 32 D. Batty and D. Crich J. Chem. Soc.. Perkin Trans. I 1992 3205. 33 W. Zhang and P. Dowd Tetrahedron Letr. 1992 33 3284. 34 P. Dowd and W. Zhang J.Orq. Chem. 1992 57 7163. S. A. Hewlins and J.A. Murphy Oy' 0 0-S-~~UU (86) s-cis (85) (89) 0 87 % Reaction Mechanisms -Part (iii) Free-Radical Reactions cycloaddition. The presence of the chlorine atom in these molecules is no real disadvantage since it is ultimately removed by tributyltin radical reduction. It has the beneficial effect of preventing hydrogen atom abstraction in those cases where this would otherwise be likely. Some curious rearrangements are seen in this work. The formation of decalone (92),from (90),is rationalized as fragmentation/recyclization of the radical (93). (93) (92) A great deal of synthetic effort has been dedicated to synthesis of taxol (94).In a novel appr~ach,~’ Pattenden has shown that selective macrocyclization (12-endo)of (95) followed by transannular ring formation (6-exo) leads to the tricyclic taxane skeleton as a mixture of epimers (96)and (97).Byproducts were also produced in this reaction from reduction of radical intermediates on the path to these epimers but the direct assembly of the carbon framework is appealing. AcO n 0 Ph OH (94)(94) OAc In another synthetic venture36 from the same laboratories an approach to lophotoxin (98) via a macrocyclization using acyl radicals was employed. Selenoester (99)gave successful cyclization to the diketone (loo),which could be converted into the furan (101).Interestingly an alternative approach to the acyl radical using acylcobalt precursors proved problematic.A final intriguing example37of macrocyclization with radicals is in the formation of macrocycle (102) in 74% yield. The authors report little if any uncyclized material. In 35 S. A. Hitchcock and G. Pattenden Tetrahedron Letr. 1992 33 4843. 36 M.P. Astley and G. Pattenden Synrhesis 1992 101. 37 K. J. Shea R. O’Dell and D.Y. Sasaki Tetrahedron Lett. 1992 33 4699. S. A. Hewlins and J.A. Murphy 0 0 (95) I $0 ??@ 0 0 (97) CHO / BugSnH TsOH AlBN this case molecular conformation must play a significant role in boosting the efficiency of the cyclization. Small Rings.-We have previously made mention of the polarity-reversal catalysis developed by Roberts. This year sees a useful application of this phenomenon.t-Butoxyl radical abstracts hydrogen atoms from esters either adjacent to the carbonyl group or adjacent to the alkyl oxygen atom but shows little selectivity between these Reaction Mechanisms -Part (iii) Free-Radical Reactions Bu3SnH AlBN sites. However by adding an amine-borane complex R3N -+ BH,R hydrogen is rapidly abstracted from the borane. The resulting radical is highly nucleophilic and will show a kinetic preference for reaction at a site which will leave an electrophilic radical. In the case of an ester this means reaction ct to the carbonyl. Thus when methyl acetate and the ally1 t-butylperoxide (103)react in the presence of this catalyst the ester is specifically fun~tionalized~~ to (104). The synthetic utility of epoxide cleavages has been further explored.39 Rawal has developed a photochemical conversion of epoxy enol acetates (105)to alcohols (106) utilizing diphenyl disulfide as the catalyst.Many of these reactions proceed in good yield. One strange fact emerges. Although it might be imagined that photocleavage of diphenyl disulfide would permit the reactions to be well initiated they do not proceed cleanly unless 10% AIBN is added to the reaction. Similar results using silyl enol ethers in place of enol acetates and tributyltin hydride plus AIBN as a thermally activated radical source have also been rep~rted.~’ It is not necessary to use enol acetates or silyl enol ethers in these reactions h~wever;~”~~ Kim has shown that keto-epoxides (107) also perform well in the 38 H.S.Dang and B. P. Roberts Tetrahedron Lett. 1992 33 4621 6169. 39 V. H. Rawal and V. Krishnamurthy Tetrahedron Lett. 1992 33 3439. 40 S. Kim and J.S. Koh Tetrahedron Lett. 1992 33 7391. 41 S. Kim and J. S. Koh J. Chem. Soc. Chem. Commun. 1992 1377. 42 E. Hasegawa K. Ishiyama T. Kato T. Horaguchi T. Shimizu S. Tanaka and Y. Yamashita J.Org.Chem. 1992 57 5352. S. A. Hewlins and J. A. Murphy presence of tributyltin radicals. This reaction is very interesting since the initially formed radical can react as above i.e. as an oxyl radical (108) or it can undergo a tributyltin shift to an enolyl radical (109). Kim has demonstrated that the latter reacting through the carbon atom can undergo cyclizations onto alkenes. The mechanism of radical-induced epoxide cleavage has inspired further study.Here,43 the effect of stereoelectronic factors on the regioselectivity of cleavage has been addressed. The rigid spiroepoxides (110) and (111) were subjected to cleavage with tributyltin radicals. If stereoelectronic factors operate and if the reaction is irreversible one would expect to see C-0 bond cleavage for (11 1) and C-C bond cleavage for (1 lo) because of the overlap between the carbon radical orbital and the adjacent epoxide bonds. However only products (112) and (1 13) resulting from C-C bond cleavage were observed. Hence this reaction either disregards stereoelectronic factors or it is reversible. One curious side-reaction seen in this study was the formation of methyl ether (114) and thioacetal (1 15) as a result of a reluctance of an intermediate radical to fragment..-"r..4Lr-Ar 43 W. R. Bowman D. S. Brown C. A. Burns B. A. Marples and N.A. Zaidi Tetrahedron. 1992 48 6883. Reaction Mechanisms -Part (iii) Free-Radical Reactions 93 Epoxides are not the only small rings which have received wide attention. Cyclobutanes and cyclopropanes are widely represented in the literature of 1992. An excellent study of the homolytic reactions of cubanes attests to the strange chemistry of these carbo~ycles.~~ Cubyl bromides (116) react with triethylsilyl radicals to give the cubyl radicals (117) which were expected to abstract hydrogen from the Si-H bond of triethylsilane; this does happen but it is in competition with abstraction from the carbon adjacent to silicon giving radical (118) as attested by ESR studies.Another notable observation was that t-butoxyl radicals appear from ESR to selectively abstract hydrogens from the ring carbons in methylcubane in preference to the methyl group hydrogens. This is counter to expectations based on bond strength calculations. When an equilibrium exists between a ring-opened form and a cyclic molecule for a three- or four-membered ring the ring-opened form normally predominates. Counter- examples have been highlighted in previous Annual Reports but more have come to light recently. Thus the bromoketal(ll9)cyclizesin72% yield to thecyclobutane (120); the authors suggest an enhanced Thorpe-Ingold effect4' operates for the two alkoxy substituents since the effect is stronger than for the case with two alkyl substituents.On the other hand it is the stability of the cyclic radical (122) which drives the equilibrium with (121) to the side of the cy~lobutane,~~ allowing good conversion to (123). EtO OEt EtO OEt 'C02Et %.,,I SMe Unexpected formation of cyclopropane (124) occurred during Dowd's investiga- tions of cyclobutanone rearrangements4' (fiide supra) when the precursor side-chains contained vinyl bromides. " E. W. Della. N. J. Head P. Mallon and J. C. Walton. J. Am. Chem. Sot,. 1992. 114. 10 730. 45 M. E. Jung. I. D. Trifunovich. and N. Lensen Tetrahedron Lett.. 1992 33 6719. 46 K. Ogura. N. Sumitani A. Kayano. H. Iguchi. and M. Fujita. Chern. Lett.. 1992 1487. '7 W. Zhang and P.Dowd. Tetrahedron Lett. 1992. 33 7307. S. A. Hewlins and J. A. Murphy Cyclopropane cleavages were probed from a novel viewpoint by Kilb~rn.~~ Methylenecyclopropanes (125) with reactive side-chains undergo initial exo-cycliz- ation and then fragmentation leading to cyclohexane (126) derivatives. (124) 40% 5 Probes The chemistry of small ring cleavage has been deployed in synthetic studies but it has also found application in kinetics in particular in the use of probes to determine the mechanisms of reactions. Both Tanner49 and Tanko” have proved false the assumption that use of the ring opening reaction of cyclopropylcarbinyl radicals as a mechanistic probe can be extended to cyclopropylketyls (127). They have shown that the ring opening is reversible and Tanko has estimated the equilibrium constant for ring opening of the ketyl(l27) as 2 x Hence ketyl can be present in test solutions without ring opening being observed.The reactions of the ketyls have been further investigated. The principal product isolated when phenylcyclopropyl ketone is subjected to bulk electrolytic reduction is the alcohol (128) arising by a coupling of ring-opened and ring-closed forme and an intramolecular hydride transfer. Newcomb’ has discovcxd ultra fast radical probes. The mono- and diphenylcyc- 48 C. Destabel and J.D. Kilburn J. Chem. SOC. Chem. Commun. 1992 596. 49 D. D. Tanner J. J. Chen C. Luelo and P. M. Peters J. Am. Chem. SOC.,1992 114 713. J.M. Tanko and R. E. Drurnright J. Am. Chem. SOC. 1992 114 1844.M. Newcornb C.C. Johnson M.B. Manek and T.R. Varick J. Am. Chem. SOC. 1992 114 10915. Reaction Mechanisms -Part (iii) Free-Radical Reactions lopropylcarbinyl radicals (129) all undergo ring opening with rate constants in excess of 10" s-Appropriate prescursors can therefore probe the extremely fleeting existences of these radicals in chemical or biological reactions. (127) I *<R1 R2 (129) a R' = H R2= Ph b R' = R2= Ph 6 Biological Applications of Radicals Cyclopropane cleavages have been exploited by Silverman in investigations of the mode of action of monoamine oxidaseS2 (MAO). This enzyme causes an initial electron transfer from amines (130). The question is what happens next? There are two possibilities (a) proton loss occurs from the neighbouring carbon,.followed by electron loss or (b) loss of hydrogen atom occurs.The product of a proton loss would be a carbon radical (131) and to test for this the phenylcyclopropyl amine (132) was treated with the enzyme. It acted as a good substrate and only one product emerged from the reaction; this was phenylcyclopropylcarboxaldehyde(134). Silverman deduced that if a carbon radical (133) was present it either had a very short lifetime or the cyclopropylcarbinyl radical was more stable than the ring-opened form (vide supra) due to stabilization from nitrogen. To probe these questions further the radical was produced by non-enzymatic routes; here the isolated product was exclusively the result of ring cleavage demonstrating that the cyclopropane cleavage was thermodynami- cally favourable.The mechanism of action of general acyl coenzyme A dehydrogenase (GAD) is now 52 R. B. Silverman and Y. Zelechonok J. Org. Chem. 1992 57 6373 5787. S. A. Hewlins and J.A. Murphy -H' I ,p- -,m -,b-,D-c;" 0 P~ NH2 Pi NH2 Pi yH2 Pi thought to proceed through radical chemistry.53 This enzyme is responsible for converting coenzyme A esters of medium length fatty acids to their @-unsaturated counterparts. It is inhibited by (methylenecyclopropy1)acetyl coenzyme A (MCPA- CoA) (135) a metabolite of hypoglycin A that causes Jamaican vomiting sickness. The realization that radicals are involved comes from isolation of the epoxide turnover product (136) from MCPA-CoA.The proposed mechanism of formation of this compound is shown. (135) &SCoA c--~SC A SCoA O -0-0 0 0-0. 0 0 FAD' 1 Dowd has revisited the chemistry of methylmalonyl-CoA rn~tase.~~ He has prepared the first model compound in which the rearrangement to a succinate (138a) is triggered by hydrogen atom abstraction from an unactivated carbon in (137) as in the enzyme reaction. The investigation has produced some uncertainty over the mechanism however since performing the reaction in MeOD led to incorporation of deuterium giving (138b) and (139b and 139c). If the process were purely radical in nature no s3 M.T. Lai and H.-W. Liu J. Am. Chem. Soc. 1992 114 3161. s4 P. Dowd B. Wilk and B. K. Wilk. J. Am. Chem. SOC. 1992 114 7949.Reaction Mechanisms -Part (iii) Free-Radical Reactions incorporation of deuterium would be expected; the observed levels of deuterium incorporation were very high. f yosE3 C02Et - EtO2C C02Et COzEt R C02Et (138) a R= H I b R=D (137) EtO2C LEt SOC The mechanism of DNA fragmentation resulting from hydrogen atom abstraction from C-4’ of deoxyribose has been studied by Giese.” By adding phenylthiyl radicals to the alkene (140) Giese has generated an intermediate resembling the resulting radical. Loss of phosphate occurs to leave a radical cation (141). Although similar radical cations have been seen in previous work by other authors the loss is surprising here since the reaction is performed in the presence of thiophenol which is one of the most effective hydrogen atom donors known.This shows that the phosphate loss is extremely rapid. 0 0 I I 0=P(OR)2 0=P(OR)2 One of the most enlightening investigations published this year concerned the action of the anti-tumour antibiotic quin~carcin.~~ When pure quinocarcin (142) was allowed to stand in the absence of oxygen it underwent a spontaneous disproportionation to 55 B. Giese J. Burger T. W. Kang C. Kesselheim and T. Wittmer J. Am. Chem. Soc. 1992 114 7322. 56 R. M. Williams T. Glinka M. E. Flanagan R. Gallegos H. Coffman and D. Pei J. Am. Chem. Soc. 1992 114 733. 98 S. A. Hewlins and J. A. Murphy two products (143) and (144).When quinocarcin was allowed to stand in the presence of oxygen it was noted that superoxide was produced.Williams proposes a mechanism consistent with the available facts where quinocarcin undergoes an electron transfer to its ring-opened form (145) giving product radicals (146) and (147). A second electron transfer from (146) leads via the cation (148) to the amide (144) while (147) is either converted into quinocarcinol (143) or reacts with dioxygen ultimately yielding superoxide and quinocarcin. DNA cleavage mediated by this drug is completely inhibited by superoxide dismutase and also strongly inhibited (84%) by conducting the reaction under anaerobic conditions. The superoxide produced is proposed to undergo reduction with traces of adventitious redox metals to produce a diffusible oxidant thought to be hydroxyl radical which cleaves DNA by established routes.The authors produce evidence of C-4’ hydrogen atom abstraction from deoxyriboses as the trigger for at least one of the operative mechanisms. Bleomycin also acts as an anti-tumour agent by abstraction of the (2-4’ hydrogen from deoxyriboses. A great deal of research has been conducted into the chemistry of modified bleomycins but it is particularly pertinent to mention one of these in this review of radical mechanisms. This concerns the report by Mascharak5’ on DNA cleavage reactions of cobalt(Ii1) analogues of bleomycin. Here under anaerobic conditions and under photochemical activation DNA cleavage is observed. ESR analysis of such solutions in the presence of dimethylpyrroline-N-oxide (DMPO) and in the absence of DNA shows a spectrum reminiscent of the hydroxyl radical adduct of DMPO.Further investigation suggests that the hydroxyl radical is not the first radical formed in solution but is preceded by another radical which gives a more complex spectrum. This is therefore assigned to a carbon- or nitrogen-centred radical. However the authors suggest that this initial radical abstracts hydrogen from water to give the hydroxyl radical which is then trapped. This is very strange as water features very strong 0-H bonds. Acetonitrile (with varying amounts of added water) is the solvent in which these ESR studies were conducted and if simple hydrogen atom abstraction were occurring one would expect that acetonitrile would be a better donor. Hence the possibility cannot be ruled out that the formation of the hydroxyl radical occurs through a less direct mechanism.Several metal-dependent oxidation systems have appeared over the last few years which are capable of destroying nucleic acids by attack on the deoxyribose hydrogens. A question which always arises is whether the intermediate is a freely diffusing oxyl radical or a metal-bound oxidant. With a molecule which shows recognition of DNA e.g.bleomycin the sequence selectivity indicates that free hydroxyl radicals are not the principal oxidants. However for molecules which do not show DNA recognition alternative methods must be found. A method which has been used by a number of authors is the quenching of the hydroxyl radical with dimethyl s~lfoxide.~~,~~ The nature of the intermediates formed in the presence of metals and oxidants is of interest also in the chemistry of methane monooxygenase which has many facets in common with Gif oxidants developed by Barton.In investigations into the nature of ’’ J. D. Tan S. E. Hudson S.J. Brown M. M. Olmstead and P. K. Mascharak J. Am. Chem.SOC.,1992,114 3841. 58 S. Hashimoto R. Yamashita and Y. Nakamura Chem. Lett. 1992 1639. 59 R. E. Shepherd T.J. Lomis and R.R. Koepsel J. Chem. SOC.. Chem. Commun. 1992 222. Reaction Mechanisms -Part (iii) Free-Radical Reactions O2+ (142) S. A. Hewlins and J.A. Murphy intermediates in these oxidations Barton6' has produced evidence that one of his intermediates can be trapped by tetramethylpiperidinoxyl (Tempo) even though it is not a free radical.This provides a caveat for those seeking to prove radical mechanisms in organometallic chemistry. One of the fastest biological reactions known is the oxygenation carried out by cytochromes P-450. This year has seen the first report of dioxygen cleavage followed by hydrocarbon hydroxylation with a strapped porphyrin acting as a P-450model bearing a thiolate ligand (149a -+ 149b). Although the site of oxygenation in the polymethylene strap is not yet established this is a very useful advance for future understanding of these important cytochromes.61 X (149a) X=H (149b) X=OH 7 Radical Chemistry of Silicon Sulfur and Selenium The search continues for new radical reducing agents with advantageous properties.The toxicity of tin derivatives has encouraged investigations into silanes and the new silane bis(trimethylsilyl)methylsilane62(Me,Si),SiMeH is now reported. It functions as expected but delivers a hydrogen atom to carbon radicals some ten times more slowly than tris(trimethylsily1)silane and some one hundred times more slowly than tributyltin hydride. The variation in chemoselectivity of different silanes is considerable and the more electrophilic tris(methylthio)silane (MeS),SiH shows a substantial tendency to add to alkenes in competition with abstraction of a bromine atom by a (MeS),Si' radical. Thus reaction with 6-bromohexene yielded 25% of tris(methy1thio)hexylsilane as well as 62% methyl~yclopentane.~~ The addition of tris(trimethylsily1)silane to the alkene in (150) was also competitive with attack on the acid chloride.64 A further difference between tributyltin hydride and silane reducing reagents is the inability of the silanes to reduce tertiary alkylnitro R,CNO compounds to the 6o D.H. R. Barton. S. D. Beviere W. Chavasiri E. Csuhai D. Doller and W.-G. Liu J.Am. Chem.SOC.,1992 114 2147. 61 H. Patzelt and W.-D. Woggon Helv. Chim. Acta 1992 75 523. 62 C. Chatgilialoglu A. Guerrini and M. Lucarini J. Org. Chem. 1992 57 3405. 63 C. Chatgilialoglu M. Guerra A. Guerrini G.Seconi K. B. Clark D. Griller J. Kanabus-Kaminska and J.A. Martinho-Simoes J. Org. Chem. 1992 57 2427. 64 M. Ballestri C. Chatgilialoglu N. Cardi and A. Sommazzi Tetrahedron Lert. 1992 33 1787. Reaction Mechanisms -Part (iii) Free-Radical Reactions 101 corresponding alkanes R,CH.The reason for this has now been el~cidated.~’ A nitroalkane (151) initially forms the nitroxide (152). Whereas the corresponding tin adduct would undergo C-N cleagave for cases where R was a tertiary alkyl group in this case N-0 cleavage occurs. The radical (153) then undergoes an exothermic rearrangement to (154) which adds to the nitroalkane giving (155). (Me3Si)3Si 0-Si(SiMe,) +I? (Me$i)+i’ R-N\ * R-N -R-lV ‘0-Si(SiMe& 0-0’ 0 (153) (151) (152) SiMe I d I O-Si(SiMe3)2 Me3% RNOz 1 R-N\ c-O-?i(SiMe& 0’ (1 54) (155) A different type of rearrangement featuring attack on silicon by a carbon radical has been reported.66 Here cyclization of enyne (156) leads to the formation of three products.The cyclic silane (157) results from vinyl radical attack on the tris(trimethylsily1)silyl group in (1 58). This attack on silicon has parallels in the and chemistry of seleni~m~’.~~ Theoretical interest in the nature of the reaction pathway^^'.^^ has been accompanied by experimental work as in the formation of selenium heterocycles (159) and (160). M~therwell’~ has discovered a novel way of appending a functionalized vinyl group onto an arene. The homopropargyl aryl sulfonates (161) suffer attack by tributyltin radical and (1,6)-ipso substitution of the sulfonate group follows. The resulting sulfur radical (162) then substitutes for the tin on the vinyl stannane giving sultones (163). 65 M.Ballestri C. Chatgilialoglu M. Lucarini. and G. F. Pedulli. J. Orq. Chem. 1992 57 948. 66 K. Miura K. Oshima and K. Utimoto Chem. Lett. 1992. 2477. ‘’ C. H. Schiesser and K. Sutej. J. Chem. Soc.. Chem. Cornmun. 1992. 57. “ C.H. Schiesser and K. Sutej Tetrahedron Lett.. 1992 33. 5137. 69 M. Tada and H. Nakagiri Tetrahedron Lett. 1992 33 6657. ’O K. F. Ferris J. Franz C. Sossa and R.J. Bartlett J. Org. Chem. 1992. 57. 777. J. E. Lyons and C. H. Schiesser J. Chem. Soc.. Perkin Trans. 2. 1992 1655. ’’ W. B. Motherwell A. K. M. Pennell. and F. Ujjainwalla J. Chem. Soc. Chem. Commun.. 1992 1067. S. A. Hewlins and J.A. Murphy EtOzC R OH ! Reaction Mechanisms -Part (iii) Free-Radical Reactions 8 Radical Chemistry of Nitrogen Barton has announced73 a novel means of introducing an amine by radical means.Thus a carbon radical (164)formed by decarboxylation of a Barton ester or by other means reacts with 3-phenyl-3-trifluoromethyldiazirine(165)to yield the imine (166) hydrolysis of which affords the amine (167). F3cxph ,R N-N-N-N ix (165) R F3C Ph Warkentin has examined74 the exolendo cyclization competition in aryl radicals derived from bromides (168). There are cases where 6-end0 does prevail but the products resulting from attack on nitrogen i.e. indolines (169),are also observed. The bizarre fact is that the ratio of products varies with the relative concentration of tributyltin hydride. The relative amount of (169) increases at higher tin hydride concentrations and the proposal is made that tributyltin radicals can attack at the nitrogen of the imine giving N-stannylamine (170) as an intermediate and that this would be followed by homolytic displacement of tin from nitrogen during the formation of the five-membered ring.This proposal is speculative at present although precedents are quoted for addition of triethyltin radicals to the nitrogen of an imine. 73 D. H. R. Barton J. C. Jaszberenyi and E. A. Theodorakis J. Am. Chern. SOC. 1992 114 5904. '' M. J. Tomaszewski and J. Warkentin Tetrahedron Lett. 1992 33 2123. S. A. Hewlins and J. A. Murphy 9 Polycyclization Chemistry One of the most active areas of research in radical chemistry is that of multiple cyclizations; the stereo- and regioselectivity afforded determines the applicability to syntheses of natural molecules.A total synthesis of (k) dihydrois~codeine~~ (172) results from (171) by a tandem 5-exo-trig-6-endo-trig cyclization and elimination with stereocontrol. The ready further conversion of this product to dihydrocodeinone (173) completes an attractive formal total synthesis of (k)-morphine. NMe Ts The tetracyclic anti-tumour agent ( )-camptothecin (175) has been ~ynthesized~~ in short order by reaction of hexabutylditin and phenylisonitrile with vinyl bromide (174).The isonitrile group acts as radical acceptor and then radical donor to form the B and c rings. -9 A/---0 C02Me Br Et (174) C0,Me Et I i ’’ K.A. Parker and D. Fokas J. Am. Chrm.SOC. 1992 114 9688.’‘ D.P. Curran and H. Liu J. Am.Chem. SOC. 1992 114 5863. Reaction Mechanisms -Part (iii) Free-Radical Reactions Multiple cyclizations also result from molecules related to the antibiotic enediynes. 1,6-Didehydro[lO]annulene (176) has been synthesized by the Myers group” at Pasadena. This molecule is highly unstable but can be observed at -90 “Cby NMR. The 13Cspectrum shows two CH signals and one signal due to quaternary carbon indicating that the electron distribution for this molecule lies between the ‘Kekule’ forms (176a) and (176b). Rapid cyclization occurs on warming above -90 “Cto give naphthalene. This rearrangement is therefore analogous to the neocarzinostatin cyclization. Diyl cyclizations are now beginning to be extended to the synthesis of more varied molecules.Thus trapping of one of the radicals from a Bergman cyclization by a pendant alkene has been accomplished by two research gro~ps.’*~’~ Note that the conversion of diyne (177) into (178) requires very forcing conditions. Interestingly the omission of cyclohexadiene leads to formation of no desired product. (176a) (176b) e ‘ I (178) 58% OTBS (177) 77 A.G. Myers and N.S. Finney J. Am. Chem. Soc.. 1992 114 10986. 78 Y. W. Andemichael Y. Gui Gu and K. K. Wang J. Org. Chem. 1992 57 794. 79 J. W. Grissom and T. L. Calkins Tetruhedron Lett. 1992. 33. 2315.