4 Reaction Mechanisms Part (iii) Free-radical Reactions By R. J. FLETCHER and J. A. MURPHY Department of Chemistry University of Nottingham Nottingham NG7 2RD This field has continued a very rapid expansion. Each year the principal themes vary markedly. This year’s highlights focus on the following areas (1) radicals centred on hydrogen nitrogen and oxygen; (2) metals and their roles in radical reactions; (3) stereochemistry of radicals and of their reactions; (4) cyclizations and fragmentations; (5) radical chemistry of molecules of biological interest; (6) three electron bonds. 1 Radicals Centred on Hydrogen Nitrogen and Oxygen Hydrogen.-A suitable place to start a review of this year’s contributions is with the simplest of radicals the hydrogen atom.Crabtree reports that hydrogen atoms’ are generated by mercury photosensitization of hydrogen gas in an unexceptional apparatus that makes them available for chemistry on a preparatively useful scale at 1 atmosphere of pressure and temperatures from 0-150°C. For example the reaction with diallyl ether (1) is reported to produce 40% of the dimeric product (2). This promises to lead to same very interesting investigations in the near future H 2 ,f \c ,f _. He-Dirncrization 0 0 0 0 0 (1) (2) as the hydrogen atoms react with excess energy due to their method of production; accordingly the radicals produced on reaction with hydrogen atoms are ‘hot’ radicals and their chemistry does not always parallel that of radicals produced in conventional ways.Thus besides the expected products (4) and (6) products of alkyl group migration (5) and (7) formed in the reaction of 3,3-dimethyl-l-butene (3) are ascribed to the intermediacy of ‘high energy radicals’ which promote 1,2-alkyl group migrations. Nitrogen.-The reactions of nitrogen-centred radicals are reported by Newcomb and others. In last year’s report his work with secondary aminyl radical cations C. A. Muedas R. R. Ferguson S. H. Brown and R. H. Crabtree J. Am. Chem. SOC.,1991 113 2233. 83 R. J. Fletcher and J. A. Murphy was discussed. He has carried these studies further noting the comparative effects of a range of Lewis acids on the cyclizations of aminyl radicals2 (8). As an alternative source of nitrogen-centred radicals the imidates (9) have been introduced.The delocalized amidyl radicals ( 10) created on irradiation of thiohydroxamates (9) could in principle react as nitrogen- or oxygen-centred radicals. However only the former reactivity is ~bserved.~ The amidyl radicals are masked forms of primary aminyl radicals. fd* MU+ MI+ X M *+ !d !d BuK* M"+_ BUN. BUG BUN + ._* (8) X=H SePh S-( 2-pyridyl) A very useful synthesis of aminyl radicals has been published by B~wman;~ the initial paper describes solely secondary amines. Here sulfenamides (13) are formed from reaction of the amine (11) with N-benzenesulfenyl phthalimide (12). Treatment of (13) with tributyltin radicals leads to formation of the aminyl radical (14). If ' M. Newcomb and C.Ha Tetrahedron Lett. 1991,32 6493. M. Newcomb and J. L. Esker Tetrahedron Lett. 1991,32 1035. W.R. Bowman D. N. Clark and R. J. Marmon Tetrahedron Lett. 1991,32 6441. Reaction Mechanisms -Part (iii) Free-radical Reactions 0 0 R' -R' R' \/N-H PhS-Nm H N D \/N-SPh \,N= + R2 R2 R2 adjacent to a cyclopropane or a cyclobutane ring opening results in the manner predicted for such radicals. The authors note the more ready cleavage of the PhS group from nitrogen than from carbon. This method of forming aminyl radicals is similar to Zard's method' of forming iminyl radicals from S-phenyl sulfenylimines. His studies have been extended6 to ring-opening of cyclobutane rings by cyclo- butyliminyl radicals. Evidence has been presented for a free radical azido-phenyl~elenenylation~ of alkenes.This reaction occurs when alkenes are treated with (diacetoxyiodo)benzene sodium azide and diphenyl diselenide in dichloromethane at room temperature. The proposed scheme features addition to the alkene triggered by azidyl radicals (15). The addition occurs not only to terminal alkenes but also to (E)-4-octene cyclohexene the electron-rich dihydropyran and the electron-poor methyl crotonate. Moreover more complicated reactions can also be achieved as seen in the formation of the cyclopentane (17) from diene (16). Oxygen.-Beckwith Zard and Newcomb' have investigated the intramolecular reactions of alkoxycarbonyloxy radicals (18a) onto appropriately placed alkenes to form 5-and 6-membered cyclic carbonates.These radicals have previously been shown to be reluctant to decarboxylate in contrast to the behaviour of their nitrogen- containing analogues (18b). Some interesting facts emerge for example exclusive 0 (18) (a) X = OR (b) X = NR J. Boivin E. Fouquet and S. Z. Zard Tetrahedron Lett. 1990 31,85 and 3545. J. Boivin E. Fouquet and S. 2. Zard J. Am.Chem. SOC.,1991 113,1055. ' M. Tingoli M. Tiecco D. Chianelli R. Balducci and A. Temperini J. Org. Chem. 1991 56 6809. M.Newcomb M. Udaya Kumar J. Boivin E. Crepon and S. 2. Zard Tetrahedron Lett. 1991 32,45. A.L.J. Beckwith and I. G. E. Davidson Tetrahedron Lett. 1991 32 49. R. J. Fletcher and J. A. Murphy 5-exo addition occurs onto the hindered alkene in (19) in contrast to the correspond- ing hexenyl radical (i.e.the carbon analogue) for which 6-endo reaction plays a significant role. The 5-membered ring formation can only be used for derivatives of primary allylic alcohols. 6-Membered ring formation occurs for derivatives of many homoallylic alcohols. An example is the conversion of (20) to (21). Note that no intramolecular hydrogen atom abstraction reactions are reported for these oxygen- centred radicals. The balance between ring closures and openings of carbon-centred and oxygen- centred radicals are highlighted in a full report by Fraser-Reid" on the reversibility of cyclizations of radicals (22). The initially bizarre result that cyclizations of carbon-centred radicals (22b) onto aldehydes to give six-membered rings work efficiently but that cyclization is not so facile for the corresponding cyclization of (22a) to give cyclopentanols is well explained by bringing together the relevant kinetic data and understanding that the successful formation of the cyclic alcohols is crucially dependent on the more rapid trapping of the oxygen-centred radicals (k = 3.7 x 10' M-' s-') than carbon-centred radicals (k = 3.2 x lo5 M-' s-l) by tributyltin hydride.kc yclizat ion(set-kS-scission(SeC-l) (a) n = 1 8.7 x lo5 (a) n = 1 4.7 x lo8 (b) n = 2 1.0 x lo6 (b) n = 2 1.1 x lo7 2 Metals and their Roles in Radical Reactions Snider reports the formation of 7- and 8-membered rings by Mn"'-induced radical cyclizations of P-keto esters." The success of particular substrates depends on the kinetics of competing reactions.This paper gives an insight into these parameters R.Walton and B. Fraser-Reid J. Am. Chem. Soc. 1991 113 5791; A. L. J. Beckwith and B. P. Hay J. Am. Chem. SOC,1991 111 230 and 2674. I' B. B. Snider and J. E. Memtt Tetrahedron 1991 47 8663. Reaction Mechanisms -Part ( iii) Free-radical Reactions 87 so that predictions about the synthetic viability of future schemes can be made. The cyclizations are sometimes surprisingly successful. For example oxidation of (23a) affords (24a) in 68% yield while (23b) affords (24b) in 70%yield! Further investiga- tions have shown that the radicals produced in the manganese method'* behave effectively as free radicals i.e. they do not behave in anomalous ways which might happen if a manganese-complexed radical rather than a free radical were involved.P-Ketotulf~xides'~ can be used instead of P-ketoesters in these oxidative radical reactions. The stereochemistry of the sulfoxide very usefully controls the stereochemistry of new centres created during cyclization reactions of these radicals. The scope of the reaction in solvents other than acetic acid has also been pr~bed,'~ and ethanol has been found to be a suitable solvent but to lead to certain differences compared with acetic acid. These principally result from a more efficient quenching of cyclic radicals by hydrogen atom donation from ethanol. The extension of this type of reaction to solvents other than acetic acid is very desirable. C02Me C02Me C02Me d ( :i (Q \ (23) (a) n = 1 (b) n = 2 Pattenden has extended the initial studies on 4-exo-trig cyclization of carbamoyl cobalt reagents to give p-lactams reported here in the 1989 review to a formal total synthesis of thienamy~in.'~ Thus the cobalt salophen reagent (25) was heated in toluene to give the vinyl p-lactam (26) which was converted easily into (27) an established synthetic intermediate on the path to thienamycin.Still on the subject of p-lactams a cobalt-mediated insertion-expansion-P-eliminationof iodomethyl- (25) (26) [Co] = cobalt(II1) salophen 3 =Gokph 0 (27) 12 D. P. Curran T. M. Morgan C. E. Schwartz B. B. Snider and M. A. Dombroski J. Am. Chem. SOC. 1991 113 6607. l3 B. B. Snider B. Y.-F. Wan B. 0. Buckman and B.M. Foxman J. Org. Chem. 1991 56 328. l4 B. B. Snider J. E. Merritt M. A. Dombroski and B. 0.Buckman J. Org. Chem 1991 56 5844. 15 G. Pattenden and S. J. Reynolds Tetruhedron Lett. 1991 32 259. 88 R. J. Fletcher and J. A. Murphy penams (28) has been shown16 to yield the commercially important 3-exomethylene cephalosporin. The particular benefit of using cobalt is evident in the final step of this reaction where a dehydrocobaltation occurs producing the alkene in the product. The ratio of exocyclic alkene (29) endocyclic alkene (30) in the product is dependent on the nature of the ligands around the colbalt. A 95 :5 ratio is achieved when vitamin B,, is used. H H H RN 0 0 0 -"pJ+ RNE& I I I I I C02Me C02Me C02Me (28) (29) (30) Harrowven and Pattenden17 have utilized the nucleophilic properties of cobalt(1) to give the P-hydroxycobaloximes (31).These molecules can be directed along two different pathways depending on whether thermal or photochemical activation is employed. Dehydrocobaltation occurs on heating the hydroxycobaloximes giving the enol of the final product (32) whereas cobalt-carbon bond homolysis occurs on irradiation. Cyclization of the resulting carbon radical trapping with Co" and dehydrocobaltation affords the diol (33). -% [co]' + A -OH -OH -OH (31) lhU An investigation of the mechanism of oxidation of dicarbonyl 7'-cyclopentadienyl iron derivatives of carboxylic acids (Fp-acyl complexes)'8 with ceric ammonium nitrate (CAN) suggests that the initial oxidation product a 17-electron radical cation (34) undergoes a fast homolytic dissociation leading to acyl radicals which (34) I co R-products l6 J.E. Baldwin R. M. Adlington and T. W. Kang Tetrahedron Lett. 1991 32,7093. " D.C. Harrowven and G. Pattenden Tetrahedron Lett. 1991 32,243. C. Amiens G. Balavoine and F. Guibe J. Chem SOC.,Chem. Commun. 1991 1458. Reaction Mechanisms -Part ( iii) Free-radical Reactions 89 have been trapped or following decarbonylation to alkyl radicals which have also been trapped. In an exciting development Merlic" has performed the first radical additions to Fischer carbene complexes. Reasoning that the carbenes are quite electrophilic and that alkyl radicals are nucleophilic he reacted the complexes (35) and (36) with radicals formed from epoxide opening by the Cp,TiCl method of Rajanbabu and Nugent.20 The carbon radicals formed added exclusively to the &carbon of the alkene rather than directly to the carbene carbon reminiscent of the reaction of radicals with Michael acceptors.(,,,sw< ClCH,CH,CI H Q (CO15W(pH (co)swr" [CP,TiC~I Ph \ Ph Ph (35) 43% 12% Ph ,Ph (36) 50% 33% Perhaps the most notable development of this year has been the spread in popularity of samarium diiodide. Inanaga et al. have performed reductive dimeriz- ations of derivatives of conjugated acids. The conditions are quite mild reactions frequently requiring less than one minute at room temperature but some interesting reactions emerge.Thus reductive reaction of the skipped diene (37) produces a cyclopropane (38) in 80% yield. The disadvantage of this and of many reactions featuring this reagent is that they greatly benefit from the addition of toxic HMPA. SmI THF HMPA ROH 80% Me02C-x.=.Jco2Me Meo2CTco2Me (37) (38) In some reactions it is possible to replace HMPA with DMPU. For example Motherwel121 has used a mixture of THF and DMPU in the samarium iodide radical ring openings of cyclopropyl ketones. He has demonstrated that the products (which are a carbon radical and a samarium enolate) can both be trapped as in the conversion of (39) to (40). One of the interesting features of samarium iodide chemistry is that the initial reduction can occur at a number of functional groups.Thus Shibuya et al. have reduced22 the allylic formates (41) to ally1 radicals which are trapped with a second equivalent of samarium iodide. The organosamarium adduct then adds to a hornoallylic formate group to give the hemiacetal (42). l9 C. A. Merlic and D. Xu J. Am. Chem. Soc. 1991 113,9855. 20 T. V. Rajanbabu and W. A. Nugent J. Am. Chem SOC.,1988 110,8561. 21 R.A.Batey and W. B. Motherwell Tetrahedron Lett. 1991 32,6211. 22 K. Shibuya H. Nagaoka and Y. Yamada J. Chem. SOC Chem. Commun. 1991 1545. R J. Fletcher and J. A. Murphy 0 OAc H M01ander~~ has reported the cyclization of P-ketoesters. Thus,the ester (43),on treatment with SmI in the presence of a variety of aldehydes or ketones undergoes an initial radical cyclization.Coupling of the product radical with a further equivalent of SmI gives an organosamarium( 111) intermediate which was trapped by aldehydes or ketones giving (44). One of the attractive features of this type of reaction is the control of relative setereochemistry observed. An interesting full report has also appeared on stereochemical control of intramolecular Reformatsky reactions24 pro- moted by samarium iodide. Although the initial step in these reactions involves electron transfer to an a-halocarbonyl group to produce an enolyl radical trapping of this species by a second equivalent of samarium iodide generates the samarium enolate which actually mediates the reaction. The duality of radical and polar chemistry associated with this reagent is finally illustrated by Curran's discovery25 Sm12 / 23 G.A. Molander and C. Kenny J. Org. Chem 1991,56 1439. 24 G. A. Molander J. B. Etter L. S. Harring and P.-J. Thorel J. Am. Chem Soc. 1991 113 8036. 25 D. P. Curran and R. L. Wolin SYNLEm. 1991 318. Reaction Mechanisms -Part (iii) Free-radical Reactions of a vinylogous Barbier reaction on substrate (45). Although the initial radical cyclization chemistry is uncomplicated the product samarium enolate was shown to give an aldol product with some aldehydes as shown here but to give a Tishchenko reaction with others. 3 Stereochemistry of Radicals and of their Reactions Control of stereochemistry in radical reactions has come strongly to the fore in recent publications.Curran and Rebek26 have employed Kemp’s triacid as the basis for their auxiliary. Radical addition to the maleate (46)occurs with high regioselec- tivity to the carbon p to the auxiliary and with interesting stereoselectivity. Based on the observed products the addition of the radical is suggested to occur preferen- tially to the maleate having the geometry shown -exposing only one face of the alkene to the incoming radical. Further experiments are in hand to understand the surprising regioselectivity of the radical addition. OEt (46) 97:3 Giese and Curran report a ‘Cram’s Rule’ for free radical reactions.” Here a comparison is made between the ratios of isomers formed by (i) hydrogen atom abstraction reactions of the radical (49) giving (48a) and (48b) after work-up and (ii) the addition of lithium aluminium hydride to (47).As seen a remarkable similarity in product ratios exists suggesting similarities in the geometries of the transition states for these reactions. Other examples also support the principle. 26 J. G. Stack D. P.Curran J. Rebek,Jr. and P. Ballester J. Am. Gem. Sor 1991 113 5918. 27 B. Giese W. Damm J. Dickhaut F. Wetterich S.Sun and D. P. Curran Tetrahedron Lett. 1991,32,6097. R. J. netcher and J. A. Murphy R (484 (48b) Me 2.9 1 Pr' 13.3 1 But 8.3 1 (47) (Me,Si),H Bu'ONNOBut I OH OH OH R (484 (48b) Me 2.9 1 Pr' 12.6 1 But 5.3 1 In a separate study on the cause of diastereoselection in radical additions Giese has investigated28 the original suggestion by Porter that allylic strain29 may cause diastereoselection in appropriate test molecules.The cases presented look very convincing. In molecules where hydrogen on Cyand a trigonal carbon at C" are present the preferred conformation of the molecule shows an eclipsed interaction for these substituents. On the other hand if C" is part of a linear group e.g. a nitrile then no such preferred conformation exists. Thus for example the ester (SO) undergoes selective hydrogen-atom abstraction giving a 20 :1 ratio of erythro :threo products whereas the nitrile (51) gives a complete lack of selectivity (1:1). erythro threo ButMe2Si0 \,Y fl-..CH2But H+-&, M'e *N 51 Guindon3' has examined the effect of chelation control on radical reductions of alkyl iodides.The acyclic iodide (52) is reduced by tributyltin hydride and AIBN 28 M. Bulliard H.-G. Zeitz and B. Giese SYNLEn 1991 423 425. 29 R. W. Hoffmann Chem. Reu. 1989 89 1841. 30 Y. Guidon J.-F. Lavallee M. Llinas-Brunet G. Homer and J. Rancourt J. Am. Chem. SOC.,1991 113 9701. Reaction Mechanisms -Part (iii) Free-radical Reactions Ph to (53) [erythro threo ratio 1:>25]. In the presence of MgIz MgBr,.Et,O or AlC13 however the ratio is switched to 25 1 i.e. exactly reversed and this is attributed to the chelation shown. Interestingly the Lewis acid-mediated reactions require no initiation and are proposed to be initiated by electron transfer from tributyltin hydride to the complex or to the Lewis acid.(53) threo As far as synthetic applications of stereoselective radical reactions are concerned Shibasaki has published3' intriguing syntheses of trans-hydrindanes synthetic pre- cursors of molecules of biological interest. Thus the vinyl radicals generated from the vinyl bromides (54) were allowed to react in the presence of low concentrations of tributyltin hydride allowing the intermediate cyclized radicals to equilibrate to the more stable products of 6-endo addition (55). The stereoselectivities for the trans-fused products were excellent. OSiMezBu' OSiMezBut &-OH -3OoC,1hr -OH Et,B,O,,Bu,SnH b Me02C Br Me02C 'H (54) (55) 97% yield 100%trans It is known that the geometry of attack of electrophiles on trigonal carbons (e.g.in alkenes) is very different from the geometry of attack of nucleophiles on trigonal carbons (e.g. in carbonyl groups) so Giese and Ho~k~~ have investigated whether a large difference exists between the transition state geometries for attack of electro-31 S. Satoh M. Sodeoka H. Sasai and M. Shibasaki J. Org. Chem. 1991 56 2278. 32 H. Zipse J. He K. N. Houk,and B. Giese J. Am. Chem. SOC.,1991 113 4324. 94 R. J. Fletcher and J. A. Murphy philic and nucleophilic radicals on trigonal carbons. They calculate that the attack of nucleophilic methyl radical (56a) and of electrophilic malononitrile radical (56b) on ethene show similar transition state geometries. .N ~109.1°173.4"~ 't108.9' 172.5"~ 154.7" 149.9" The nature of the transition state used in 5-hexenyl cyclizations has been examined over many years by many authors but notably by Beckwith.His guidelines based on a pseudo-chair transition state have allowed the prediction of stereochemistry of disubstituted cyclopentanes for many cyclizations. Houk has previously suggested that boatlike transition states may be involved in some such cyclizations; Be~kwith~~ has now investigated a case (57) -+ (58) where the boatlike transition state would betray its presence by the stereoselectivity of the reaction. Thus if only the chair forms of the transition states (59) and (60) were to be used there would be a large difference in energy between that which had the t-butyl group axial and that which had it equatorial so large that the ratio of cis :trans products would be ca.500 1. The observed ratio of 4.5 1 is much closer to that predicted if the trans isomer forms through a boat-like transition state (61). The calculated rate constant for the formation of the cis isomer is 1.6 x lo8sec-' at 80 "C and is the highest rate constant yet recorded for cyclization of a monosubstituted hexenyl radical. But (59) I I I cis (58) trans (58) trans (58) Formation of six-membered rings by radical cyclization is normally slower than formation of five-membered rings but an exceptionally fast 6-membered ring cycliz- ation has appeared (62) + (63) which is super fast with a suggested rate constant for formation of the predominant cis product of 1.1 x 10'" sec-' at 80 "C. The high 33 A.L. J. Beckwith and J. Zimmermann .IOrg. Chern. 1991 56 5791. Reaction Mechanisms -Part (iii) Free-radical Reactions rate34 is suggested to arise since transannular cyclization of (62),in which the radical centre and the double bond are already rotationally constrained imposes a relatively small loss of rotational freedom. There is as yet no satisfactory explanation for the stereoselectivity of the reaction. Bordwell and Gallaghe?' have examined the stability of radicals (64) derived from a-dialkylaminoketones and conclude that an electrostatic stabilization can arise in a particular conformation in such radicals; the reason for the effect is seen if the radicals are drawn in their zwitterionic form. Here the Z-geometry (65)shows the electrostatic stabilization and where the 2-form cannot exist e.g.(66) the radicals have considerably less stabilization. T 4 Cyclizations and Fragmentations A number of full papers describing the formation and fragmentation of large and medium sized rings giving details of work previously described in Communication fOrm36,37 appear this year. Thus full details of Suginome's work on photofragmenta-tion of hypoiodites leading uia alkoxyl radicals to phthalide~~',~~ shows the rather complicated pathway leading to phthalides from benzocyclobutanols. It is still not certain whether the formation of the five-membered ring occurs by a radical or ionic mechanism. Alkoxyl radicals also feature in Pattenden's syntheses4' of hydrazulenones. Fragmentation of the alkoxyl radicals (67) followed by cyclization 34 A.L. J. Beckwith V. W. Bowry and C. H. Schiesser Tetrahedron 1991 47 121. 35 F. G. Bordwell T. Gallagher and X. Zhang J. Am. Chem. SOC.,1991 113 3495. 36 J. E. Baldwin R. M. Adlington M. B. Mitchell and J. Robertson Tetrahedron 1991 47 5901. 37 P. Dowd and S.-C. Choi Tetrahedron 1991 47 4847. 38 K. Kobayashi A. Sasaki Y. Kano and H. Suginome Tetrahedron 1991,47 7245. 39 K. Kobayashi M. Itoh A. Sasaki and H. Suginome Tetrahedron 1991 47 5437. 40 C. Ellwood and G. Pattenden Tetrahedron Lett. 1991 32 1591. R. J. Fletcher and J. A. Murphy leads to the radicals (68) which in the presence of iodosylbenzene diacetate and the hypoiodite form the quenched products (69). Ring expansion is possible with these products on treatment with tributyltin hydride as shown.I Bu,SnH e-- A number of products resulting from bizarre and interesting rearrangements are presented by Bald~in.~' Thus for example it was proposed to synthesize the macrocyclic ketone (71) from the iodoketone (70) by a cyclization fragmentation and elimination sequence. The actual pathways followed did indeed produce some of this product; however the cyclopentenones (72) and (73) were also produced. Proposed pathways are shown. The crucial point is that steric crowding of the ketone carbonyl slows addition to such an extent that alternative processes here -hydrogen atom abstraction -are observed. 0 fl**.vvl -cat. Bu,SnH AIBN -SnBu3 (70) (71) 33% 01 (72) E-isomer 26% (73) Z-isomer 23% :*--(72)+ (73) yJ--)-a*-* SnBu3 SnBu3 SnBu3 41 J.E. Baldwin R. M. Adlington and J. Robertson Tetrahedron 1991,47 6795. Reaction Mechanisms -Part (iii) Free-radical Reactions Ryu Sonoda and co-workers have produced three report^^^-^ featuring the radical chemistry of CO. Formation of an acyl radical by carbonylation of an alkyl radical is followed by trapping with an activated alkene an allylstannane or [intramolecularly] with an unactivated alkene giving products (74)-(76).The relative rates of the reactions and the relative concentration of carbon monoxide to the Bu SnH -RCO. co @C N RI 3 Re AIBN 80 atm Bu,SnH R R = nC,H, (74) 74% 0 CO 10 atm. AIBN other trapping agent are of crucial importance in obtaining high yields.In a related study C~rran~~ has probed the ability of aryl isonitriles (77) to trap pentynyl radicals (78). Here the isolated products are cyclopenta-fused quinolines. Two isomers are produced one the product of ortho attack (79) the other the product of ips0 attack (79) I R R R (80) 42 I. Ryu K. Kusano H. Yamazaki and N. Sonoda J. Org. Chem. 1991 56 5003. 43 I. Ryu H. Yamazaki K. Kusano A. Ogawa and N. Sonoda J. Am. Chem. SOC.,1991 113 8558. 44 I. Ryu K. Kusano M. Hasegawa N. Kambe and N. Sonoda J. Chem. SOC.,Chem. Commun. 1991,1018. 45 D. P. Curran and H. Liu J. Am. Chem. SOC.,1991 113 2127. R. J. Fletcher and J. A. Murphy 5 Radical Chemistry of Molecules of Biological Interest The cleavage of nucleic acids by radical chemistry is again a subject of much research.In a studyM on the cleavage chemistry of the anti-tumour antibiotic neocarzinostatin it has been discovered that the rate-determining step for the cleavage can involve hydrogen atom abstraction from C-4’ or C-5’ of a thymidine residue. The interesting fact is that this does not hold for all thymidines but appears to occur for all thymidines in a GT sequence. It had been that some of the synthetic ‘mimics’ of the enediyne antibiotics may effect their action on DNA by polar routes ie. routes which do not involve the formation of an arene diradical. Tats~ta~~ has examined the cyclization of the simple enediyne (81) (a) in the presence of KOH/DMSO/MeOH and (b) in carbon tetrachloride/DBU either under argon or in air.The product of the first reaction (82) suggests a non-radical cyclization. The second reaction however gives the compounds (83)-(85) in argon and the additional ketone (86) in air. The DBU pathway is proposed to involve free radical intermediates. This re-emphasizes the subtlety and diversity of the chemistry of the enediynes. c1 c1 c1 Interest in methods of producing the hydroxyl radical artificially for the purpose of DNA cleavage is widespread the Saito4’ has found an efficient new photochemi- cally activated source the bis-hydroperoxynaphthaldiimide(87) [activated with UV radiation at 366 nm]. 46 B. L. Frank L. Worth Jr. D. F. Christner J. W. Kozarich J. Stubbe L. S.Kappen and 1.H. Goldberg J. Am. Chem. Soc. 1991 113 2271. 41 K. C. Nicolaou G. Skokotas S. Furuya H. Suemune and D. C. Nicolaou Angew. Chem. Engl. Int. Ed. 1990 29 106/4. 48 K. Toshima K. Ohta T. Oktake and K. Tatsuta Tetrahedron Lett. 1991 32 391. 49 H. Sugiyama T. Sera Y. Dannoue R. Marumoto and I. Saito J. Am. Chem. Soc. 1991 113 2290. Reaction Mechanisms -Part ( iii) Free-radical Reactions 0 OMe (87) To probe the mechanism of action of the anti-tumour antibiotic bleomycin Saito and colleagues” synthesized an oligonucleotide d( GGAriAGG)-d( CCTTCC) where Ari is 2’-deoxyaristeromycin (88). The products derived from interaction of peplomy- cin a derivative of bleomycin with this DNA featured two modifications of the aristeromycin the dehydro product (89) and the alcohol (90).The alcohol (90) was formed with complete stereoselectivity but the source of the oxygen of the alcohol group remains to be established. The formation of these products is consistent with the formation of an intermediate radical at the 4‘position which is subsequently oxidized to a cation. Parallels are frequently drawn between the chemistry of cytochromes P-450and bleomycin since both are thought to involve high valence iron-oxo species. Saito points out that the formal dehydrogenation seen in the formation of the alkene (89) has not been observed before for either bleomycin or P-450. S. Matsugo S. Kawanishi K. Yamamoto H. Sugiyama T. Matsuura and I. Saito Angew. Chem. Znt. Ed. EngL 1991,30 1351. R.J. Fletcher and J. A. Murphy The chemistry of high-valent iron also features in studies on penicillin and cephalosporin biosynthesis. In Baldwin's proposal,*' the high-valent iron oxene in (91) inserts stereospecifically into the methine carbon-hydrogen bond to form an iron-carbon bond. Homolysis of this bond is followed by displacement at sulfur to Enz H H I H Enz RN,S-Fe=o CO2H t COZH CO2H (92) give the penicillin (92). Among the new s-ibstrates examined have been (93) and (94). No methine hydrogen is present for reaction in these substrates. Instead the methyl group undergoes reaction via the modified intermediate (97). The two dideuteriocyclopropanes were found to be regiospecifically converted into the prod- ucts (95) and (96).This implies that the cyclopropane is opened by homolysis with Enz H _. H H I ,OH RN .NuSH Me 0UXy2 CXZ C02H zC02H HO2C (93) X = H Y = DZ = D (95) X = HY =DZ=D (94) X = D Y = H Z = H (96) X = D Y = H Z = H (97) complete stereoselectivity and that the subsequent trapping of the radical by sulfur is faster than the equilibration process shown for (98). If this equilibration process were allowed to occur for a free radical one would expect a lack of regioselectivity in the labelling of the product. (98) 5 \\ J. E. Baldwin R. M. Adlington D. A. Marquess A. R. Pitt and A. T. Russell J. Chem SOC. Chem. Commun. 1991 856. 101 Reaction Mechanisms -Part (iii) Free-radical Reactions Soluble methane monooxygenase converts methane to methanol.In addition it catalyses a broad range of other oxidations. Studies by Daltons2 in Warwick and colleagues from Leicester have probed the mechanism of these enzymatic conver- sions. Specifically it had been previously suggested that the initial step in the oxidation would feature a hydrogen atom abstraction with a mechanism similar to that used by cytochromes P-450. The question arose of whether the hydrogen atom abstracting species was a free hydroxyl radical or a metal-bound oxyl (i.e.a ferryl species). This study found that the CH3' HOCH,' NCCH2' radicals produced in the oxidations of methane methanol and acetonitrile were easily trapped either by 5,5-dimethylpyrroline-l-oxide (DMPO) or by cu-(4-pyridyl-l-oxide)-N-t-butyl-nitrone (POBN).The authors also found that hydroxyl radicals could easily be trapped by their spin-traps also but detected none. They therefore suggest that it is an iron-bound oxyl which abstracts the hydrogen and that no hydroxyl radical is produced during the rection. This in turn implies that heterolysis of an intermediate iron hydroperoxide leads to the ferryl species. 0-11 OCH2Bu' \ Et-P/ OCH2Bu' - OCH~BU' ICHpBu'+I c -Et-P/ THF Bu.N+-OH (Bu'CH,O),P=O I Ft 26% (99) (Bu'CH,O),P=O 73% C2H6 4% C2H4 4% C,H,o 0.3% (Oy OOH 0 n-n .O. *OEt . Et (100) (101) (Bu'CH,O),P=O The degradation of organophosphonates by micro-organisms has been shown to involve cleavage of the C-P bond. All inorganic phosphate needed by the microbe is derived from the phosphonate phosphorus.In a continuation of his pioneering work on organophosphonate degradation Frosts3 has examined the effect of homolytic hydrolysis of organophosphonates. The conditions exposed (99) a model for an enzymatically protonated phosphonate to tetra-n-butylammonium hydroxide THF and either (a) t-butylhydroperoxide and ferrous ions or (b) 2-hydroperoxy- tetrahydrofuran without metal ions. In these cases the isolation of trineopentyl phosphate and alkanes established the homolytic pathway. A curious reaction was observed during this study in which the peroxide (100) was converted into ethoxytetrahydrofuran (101); the mechanism of formation of this compound is currently under investigation. 52 N. Deighton I.D. Podmore M. C. R. Symons P. C. Wilkins and H. Dalton J. Chem. Soc. Chem. Commun. 1991 1086. 53 L. Z. Avila P. A. Bishop and J. W. Frost J. Am. Chem. SOC.,1991,113,2242. R. J. Fletcher and J. A. Murphy 6 Three Electron Bonds The chemistry of radical cations centred or partly centred on sulfur has been the subject of a number of investigations. Asmus has fostered the development of much of this area and has now produced evidence for the first three electron bond54 featuring a phosphorus and a sulfur atom (102). Pulse radiolysis (1 ps pulse) of a solution of (103) saturated with nitrous oxide gave the transient species (102) which absorbed at 385 nm. The lifetime was pH dependent with half-lives of 150 ps and 70 ps at pH 4 and 8.6 respectively.(102) (103) Three electron S-S bonds have been described before but Bushby” has observed an unexpected example. Whereas the arene (104) is oxidized to the wdelocalized radical cation (105) the triphenylene derivative (106) is oxidized to the radical cation (107). This is demonstrated by examination of the ESR spectra of (107a) and its deuteriated analogue (107b). SR SR R .’+ / SR (107) (a) R = (CH2)JH3 (b) R = CHD(CH2),CH 54 H. Hungerbuehler S. N. Guha and K. D. Asmus J. Chem SOC Chem Commun. 1991,999. 55 N.Boden R.Borner R. J. Bushby and J. Clements Tetrahedron Lett. 1991 32 6195.