摘要:

J. Chem. SOC., Faraday Trans. I, 1987,83, 85-89 Formation of Superoxide during the Auto-oxidation of Anthralin (1,8-Dihydroxy-9-anthrone) J. Malcolm Bruce* and Colin W. Kerr Department of Chemistry, The University, Manchester M13 9PL Nicholas J. F. Dodd Paterson Laboratories, Christie Hospital and Holt Radium Institute, Manchester M20 9BX Auto-oxidation of the antipsoriatic drug anthralin (1,8-dihydroxy-9- anthrone) in dimethyl sulphoxide has been shown by spin trapping with 5,5-dimethyl-l-pyrroline- 1-oxide (DMPO) to lead to the formation of superoxide. The auto-oxidation of several analogues of anthralin also results in the formation of superoxide. Anthralin [dithranol; l,S-dihydroxy-9-anthrone (1, R = H)] has been in clinical use for the treatment of the skin disease psoriasis since 1916,l but its mode of action has yet to be established; the chemistry of anthralin and its analogues in vitro is not fully understood.1 -Hydroxy-9-anthrone (2, R = OH) is regarded2 as the simplest 9-anthrone to display antipsoriatic activity; 9-anthrone (2, R = H) is ina~tive.~ 10-Butyryl-1,s- dihydroxy-9-anthrone [(1, R = COPr), butanthrone] has recently been extensively OH 0 OH O R H R screened, but whilst it shows less4v of the major side effects of anthralin, viz. burning and staining of the skin,s it does not appear to be as useful an antipsoriatic as anthralin; it may act, at least in part, by in vivo cleavage of the vinylogous B-diketone moiety to yield anthralin.2 10,lO-Dialkylanthralins are inactive against Auto-oxidation of anthralin affords both the dehydrodimer (3)8 and 1,s-dihydroxy- 9,lO-anthraquinone (4),O9 lo the former predominating when the oxidation is effected in (y-& / \ I OH 0 OH w 0 (3) ( 4) 8586 A uto-oxidation of 1,8-Dihydroxy-9-anthrone acetonell? l2 solution and the latter predominating when the medium is aqueous sodium hydr0~ide.l~ Formation of the anthralin radical (1, R = .)14 is implicated, with the peroxide (1, R = OOH) being a likely precursor of the anthraquinone (4).13* l5 The keto form (1, R = H) of anthralin is stable to auto-oxidation,16 but both the enol(5, R = H) and enolate (5, R = -ve) are susceptible, the enolate particularly so.16 The consumption of oxygen by psoriatic cells is considerably greater than that by normal epidermal cells17 and a possible mechanism for the therapeutic effect of anthralin involves it acting as an oxygen scavenger within the cell, thus reducing the proliferative rate.A consequence of such scavenging would be the generation of active oxygen species,18 including the peroxide (1, R = OOH) and oxygen radicals derived from it. These species could cause cell death.lS They may also be involved in the burning of normal skin following application of anthralin. Central to this hypothesis is the mechanism of auto-oxidation of anthralin and we suggest that the key step is transfer of an electron from the enol (5, R = H) to oxygen to afford the cation radical (6) [cf. ref. (20)] and superoxide (02-); deprotonation of (6) will then afford the radical (1, R = a). We now present evidence relating to the formation of superoxide.Experiment a1 E.s.r. spectra were recorded at room temperature using a Varian E-9 X-band spectro- meter in conjunction with a Nicolet 1170 signal averager and Hewlett-Packard HP 85 microcomputer. Dimethyl sulphoxide was purified by distillation over 4A molecular sieves at 20 mmHg.? DMPO (Aldrich Chemical Co. Ltd) was distilled at 10 mmHg. Other compounds were purified by crystallisation and/or sublimation at 0.01 mmHg, and characterised by lH n.m.r. spectroscopy. The concentration of each compound under investigation was ca. mol dmP3, with DMPO at 10-l mol dmP3. Initial spectra were measured immediately after preparation of solutions. Results and Discussion There are several methods for the detection and determination of superoxide.21 Thus reduction of nitroblue tetrazolium by superoxide generates the dark blue diformazan,22? 23 which can be monitored spectroscopically at 560 nm.Similarly, ferricytochrome c is reduced24 to ferrocytochrome c, detectable at 550 nm. Confirmation of the involvement of superoxide is provided by inhibition of these reductions by the addition of catalytic amounts of superoxide dismutase. However, both reagents can give positive results in the presence of other reductantsz59 26 and in some circumstances superoxide dismutase inhibits the reduction of nitroblue tetrazolium, but does not affect the reduction of ferricytochrome c.~', 28 This discrepancy can be explained by prior reduction (by an alternative reductant) of nitroblue tetrazolium to its radical, which then reduces molecular oxygen to ~uperoxide.~~- 2 7 9 28 In the present work both tests were positive, even under anaerobic conditions, when 1 mmHg % 133.3 Pa.J .M . Bruce, C. W. Kerr and N . J . F. Dodd 87 applied to anthralin in dimethylformamide or aqueous dimethylformamide, solvents in which anthralin is in equilibrium with its enol(5, R = H),l09 l6 thus invalidating them for the detection of superoxide during the auto-oxidation of anthralin; it has previously been claimedz9 that reduction of nitroblue tetrazolium during aerial oxidation of anthralin is indicative of the formation of superoxide. Reduction of tetranitromethane to the nitroform anion has also been used for the detection of s~peroxide,~~ but the monitoring wavelength (350 nm) complicates its use with anthralin which absorbs strongly in this region.Again, enolic anthralin may reduce the reagent directly. Because of these uncertainties, the auto-oxidation of anthralin and some of its analogues were examined by e.s.r. spectroscopy using 5,Sdimethyl- 1 -pyrroline- I -oxide (DMPO) (7) as a spin trap for s~peroxide.~~ Anthralin and its 10-monosubstituted analogues all equilibrate with their enolic forms in dimethyl sulphoxide,l0? l6 which was used as solvent for the e.s.r. experiments. For each experiment a control using DMPO and oxygen in the absence of the anthralin followed by purging with nitrogen did not give a detectable signal. When anthralin (1, R = H) was present, the e.s.r. spectrum (fig. 1) previously to the adduct (8) of protonated superoxide with DMPO was readily detected, indicating the formation of superoxide.Parallel results were obtained for 9-anthrone (2, K = H), l-hydro~y-9-anthrone~~ (2, R = OH), and l0-propy1anthralinl6 (1, R = Pr), consistent with production of superoxide via transfer of an electron from the corre- sponding enols to molecular oxygen. B ~ t a n t r o n e ~ ~ (1, R = COPr), IO-pr~pionylanthralin~~ (1, R = COEt), and 10,lO-di- propylanthralin' gave very weak signals, supporting, for the first two of these compounds, the expectation that oxidation of the corresponding 10-acyl- 1,8,9-trihydroxyanthracenes I I I I I 1 1 32 7 328 329 330 331 332 333 field/mT Fig. 1. E.s.r. spectrum of the DMPO-OOH spin adduct (aN = 1.29 mT, af = 1.04 mT, a? = 0.14 mT), observed after exposure to oxygen, of 10W mol dmP3 anthralin, lo-' mol dmP3 DMPO in dimethyl sulphoxide.Spectrum recorded at room temperature, 5 mW incident microwave power, 0.02 mT modulation amplitude.88 Auto-oxidation of I,8-Dihydroxy-9-anthrone (9, R = COPr and COEt, respectively) will be retarded by the electron-accepting acyl groups. No signal would have been expected from 10,lO-dipropylanthralin, which cannot enolise. IO-Hydr~xyanthralin~~ (1, R = OH) and lO-meth~xyanthralin~~ (1, R = OMe) gave stronger signals than anthralin, consistent with the lower ionisation potential expected for their enols (9, R = OH and OMe, respectively), which are more electron-rich than anthralin owing to the mesomeric effect of the 10-substituent. OH OH OH The dehydrodimerll3 l2 (3) behaved similarly to anthralin and its 10-propyl homologue.The signals due to the spin adduct (8) progressively increased in intensity when the solutions containing 1 -hydroxy-9-anthrone, 1 0-propylanthralin, 1 0-methoxyanthralin and the dehydrodimer (3) were left in contact with air. For anthralin the absorption due to the spin adduct decayed and a 1 : 2: 1 triplet (a = 0.32 mT), similar to that previously reported3' to occur during the auto-oxidation of anthralin in the absence of a spin trap, appeared. An identical triplet developed for solutions of the dehydrodimer, but more slowly. This triplet has not been assigned. Auto-oxidation of anthralin in dimethyl sulphoxide ultimately yields 1 ,8-dihydroxy- 9,lO-anthraquinone (4) as the major product16 and it was therefore appropriate to determine whether the presence of this quinone affected the DMPO spin-trapping experiments.When a dimethyl sulphoxide solution containing both the quinone (4) and DMPO was oxygenated and then purged with nitrogen, a weak signal due to the adduct (8) was observed; its intensity progressively increased when the sample was kept in air. 1 -Hydroxy-9,lO-anthraquinone and 9,lO-anthraquinone at the same concentration behaved similarly, but the signals were, respectively, weaker and much weaker. These results may be explained by electron transfer from DMPO to the quinone to give the semiquinone, which is then oxidised back to the quinone by molecular oxygen, with concomitant formation of superoxide (redox cycling). The progressively weaker e.s.r. signal through the series parallels the first half-wave reduction potentials of the quinones, viz.1,8-dihydroxy-9,10-anthraquinone > 1 -hydroxy-9,lO-anthraquinone > 9, IO-anthraquin~ne.~~ These observations indicate that caution must be exercised in interpreting the results of spin-trapping experiments in the presence of quinones, particularly high potential quinones. However, for separate solutions of anthralin and 1,8-dihydroxy-9,10- anthraquinone at the same concentration in the presence of an excess of DMPO, the initially observed signal due to the spin adduct (8) is much stronger for the anthralin solution. Since at this stage relatively 1ittlel6 of the anthralin has been oxidised to the corresponding quinone, the contribution to the e.s.r. signal from the redox cycling route will be small; the major contribution is therefore due to superoxide produced via electron transfer from the anthralin enol (5, R = H) to molecular oxygen.These results thus provide definitive evidence for the formation of superoxide during the auto-oxidation of anthralin and give a further indication of the likely involvement of anthralin radicals and active oxygen species during the treatment of psoriasis with anthralin, where, although the conditions are formally very different from those obtaining in solvents such as dimethylformamide and dimethyl sulphoxide, the dielectricJ. M. Bruce, C. W. Kerr and N . J . F. Dodd 89 properties of the partly aqueous environment in V ~ U O ~ ~ are conducive to the formation of the reactive enolic form (5, R = H) of anthralin, in an oxygen-rich environment.However, the formation of superoxide from anthralin analogues does not in itself imply that these compounds will be active against psoriasis; drug distribution in the epidermal region must play a vital role. We thank ICI Pharmaceuticals Division and the S.E.R.C. for a CASE Studentship (C. W. K.) and the C.R.C. for financial support. References 1 R. E. Ashton, P. Andre, N. J. Lowe and M. Whitefield, J. Am. Acad. Dermatol., 1983, 9, 173, and 2 A. Krebs, H. Schaltegger and A. Schaltegger, Br. J. Dermatol., 1981, 105, Suppl. 20, p. 6. 3 A. Krebs and H. Schaltegger, Hautarzt, 1969, 20, 204. 4 K. K. Mustakallio and P. J. Kolari, Acta Derm. Venereol. (Stockholm), 1983, 63, 513. 5 A. M. Goransson, P. J. Kolari and K.K. Mustakallio, Acta Derm. Venereol. (Stockholm), 1984,64,134. 6 J. P. Nater and A. C. De Groot, Unwanted Effects of Cosmetics and Drugs Used in Dermatology 7 W. Wiegrebe, A. Gerber, J. Kappler and C. Bayerl, Arzneim-Forsch., 1979, 29, 1083. 8 M. Whitefield, K. Hendrick and P. G. Owston, Acta Crystallogr., Sect. B, 1982, 38, 1248. 9 D. Cavey, J-C. Caron and B. Shroot, J. Pharm. Sci., 1982, 71, 980. references therein. (Excerpta Medica, Amsterdam, 1983), p. 145. 10 T. Sa e Melo, L. Dubertret, P. Prognon, A. Gond, G. Mahuzier and R. Santus, J. Invest. Dermatol., 11 A. Segal, C. Katz and B. L. Van Duuren, J. Med. Chem., 1971, 14, 1 152. 12 P. Hofer and M. Schurch, Pharm. Acta Helv., 1975,50, 202. 13 C. T. Bedford, J. Chem. SOC. C, 1968, 2514. 14 A. G. Davies, J. A-A.Hawari and M. Whitefield, Tetrahedron Lett., 1983, 24, 4465. 15 J. M. Bruce, C. W. Kerr and A. Zare-Shyadehi, unpublished work. 16 J. M. Bruce and C. W. Kerr, unpublished work. 17 H. Hammar and C. Hellerstrom, Acta Derm. Venereol. (Stockholm), 1968, 48, 563. 18 Oxygen Radicals in Chemistry and Biology, ed. W. Bors, M. Saran and D. Tait (de Gruyter, Berlin, 19 Cf. D. Schulte-Frohlinde and C. von Sonntag, in Oxidative Stress, ed. H. Sies (Academic Press, New 20 A. J. Bard, A. Ledwith and H. J. Shine, Ado. Phys. Org. Chem., 1976, 13, 156. 21 Handbook of Methoh for Oxygen Radical Research, ed. R. A. Greenwald (CRC Press, Boca Raton, 22 I. Saito, T. Matsuura and K. Inoue, J. Am. Chem. SOC., 1983, 105, 3200. 23 C. Auclair and E. Voisin, in Handbook of Methods for Oxygen Radical Research, ed.R. A. Greenwald 24 J. Butler, G. G. Jayson and A. J. Swallow, Biochim. Biophys. Acta, 1975, 408, 215. 25 C. Auclair, M. Torres and J. Hakim, FEBS Lett., 1978, 89, 26. 26 P. Morlikre, L. Dubertret, T. Sa e Melo, C. Salet and R. Santus, J. Invest. Dermatol., 1983, 80, 350. 27 V. Mishin, A. Pokrovsky and V. V. Lyakhovich, Biochem. J., 1976, 154, 307. 28 C. Auclair, E. Voisin and H. Banoun, in Oxy Radicals and their Scavenger Systems, ed. G. Cohen and R. A. Greenwald (Elsevier, Amsterdam, 1983), vol. 1, p. 312. 29 K. K. Mustakallio, Br. J. Dermatol., 1981, 105, Suppl. 20, p. 23. 30 J. M. McCord and 1. Fridovich, J. Biol. Chem., 1969, 244, 6049. 31 E. Finkelstein, G. M. Rosen and E. J. Rauckman, Arch. Biochem. Biophys., 1980, 200, 1. 32 J. R. Harbour and M. L. Hair, J. Phys. Chem., 1978,82, 1397. 33 E. J. Cross and A. G. Perkin, J. Chem. SOC., 1930, 292. 34 K. K. Mustakallio, A. K. Pippuri and E. J. Honkanen, Eur. Patent, 1980, 0 017 420 Al. 35 M. d’Ischia, A. Privitera and G. Prota, Tetrahedron Lett., 1984, 25, 4837. 36 R. A. Barnes and W. Holfeld, Chem. Znd. (London), 1956, 873. 37 J. Martinmaa, J. Juselius and K. K. Mustakallio, in Psoriasis: Proc. 3rd Znt. Symp., ed. E. M. Farber and A. J. Cox (Grune and Stratton, New York, 1982), p. 383. 38 A. Ashnagar, J. M. Bruce, P. L. Dutton and R. C. Prince, Biochim. Biophys. Acta, 1984,801, 351. 39 Cf. R. Wolfenden, Science, 1983, 222, 1087. 1983,80, 1. 1984). York, 1985), chap. 2. Florida, 1985). (CRC Press, Boca Raton, Florida, 1985), p. 123. Paper 61822; Received 28th April, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300085

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1987,83,91-94 Paramagnetic Adducts in the Reaction of 4-Substituted Pyridines and Phosphorus-centred Radicals Angelo Alberti I. Co. C. E. A .-C. N . R., 40064 Ozzano Emilia, Italy Andrew Hudson* School of Chemistry and Molecular Sciences, University of Sussex, Brighton BNl 9QJ Gian Franco Pedulli Istituto di Chimica Organica, Universita', 09100 Cagliari, Italy The reactions of a number of 4-substituted pyridines with diethoxyphos- phonyl, diphenylphosphonyl and diphenylphosphinyl radicals have been monitored by e.s.r. spectroscopy in the temperature range 293 to 403 K. Whereas with phosphonyl radicals the corresponding 1 -phosphonyl pyri- dinyls were observed, the spectra detected with phosphinyl radicals proved to be those of the 4-substituted 1-hydropyridinyls.A comparison with data available for related pyridinyls indicates a marked electron withdrawing ability for the phosphonyl group. An analogous reactivity pattern has also been established for pyrazine and 2,6-dimethylpyrazine. Numerous e.s.r. studies have been concerned with the structure and stability of paramagnetic species formed by the addition of organometallic radicals to suitable organic substrates. Thus the addition of Group IVB radicals (silyls, germyls and stannyls) to substituted pyridines produces pyridinyl radicals whose stability is strongly dependent on the nature and position of ring sub~titution.l-~ Phosphorus-centred radicals often undergo addition reactions analogous to those of Group IVB radicals, for example with quinonoid c~mpounds.~ In this note we describe some new pyridinyls formed in the addition of phosphorus-centred radicals to some 4-substituted pyridines and to pyrazine. Experimental The pyridines and pyrazines used as substrates were either obtained commercially or prepared by standard procedures.The phosphorus compounds were purchased from Janssen Chimica and distilled prior to use. Diphenylphosphinyi radicals were generated either by reaction of diphenylphosphine and t-butoxy radicals, by halogen abstraction from diphenylchlorophosphine by 'Re(CO), radicals or by direct photolysis of solutions containing tetraphenyl- biphosphine. Diphenylphosphonyl radicals were generated by hydrogen abstraction from diphenyl- phosphine oxide employing butoxy radicals ; diethoxyphosphonyl radicals were similarly obtained from diethylphosphite or tetraethylpyrophosphite. E.s.r.spectra were obtained by irradiation of deoxygenated dilute solutions of the reactants in t-butylbenzene using a 1 kW high-pressure mercury lamp and were recorded on a Bruker ER 200 spectrometer. In most cases the quality of the spectra improved on raising the temperature to ca. 70 "C. 9192 Paramagnetic Adducts of 4-Substituted Pyridines Results and Discussion A marked difference was found between the behaviour of phosphonyl and phosphinyl radicals and we shall deal with each case separately. Phosphonyl Radicals Both diphenyl- and diethoxy-phosphonyl radicals gave good quality e.s.r. spectra with pyridines substituted in the 4-position with an electron-withdrawing group (-CN, -COOR, -Ph).The hyperfine coupling constants for these adducts (table 1) are consistent with the formation of pyridinyl radicals (I). However, no e.s.r. spectra were observed when the substituent was an electron-releasing group (-NH,, -OH, -SH). The spectra are characterized by rather small 31P splittings (0.7-1.5 G) and 14N couplings of 3-4 G. The latter appear to be somewhat higher for the diphenylphosphonyl adducts than for those formed by diethoxyphosphonyl radicals. A more interesting comparison can be made by combining the data for the new radicals in table 1 with previously published data on Group IVB adducts. The hyperfine coupling constants for a series of adducts formed by 4-cyanopyridine (table 2) exhibit a pronounced substituent effect on the I4N splitting, which is almost halved on replacing methyl by diethoxyphosphonyl.The origin of these variations can be understood by reference to the resonance structures : A high-spin density at the nitrogen would be favoured by substituents which can easily delocalize the positive charge, i.e. electron-releasing groups. The increase in a(N) and a(3), and the concomitant decrease in a(2) for the sequence P(O)R, < SIR, < GeR, < SnR,, H < CH,, is thus an indication of the increased significance of polar struc- tures on moving from electron-withdrawing to electron-donating substituents; the phosphonyl group acts as a strong electron acceptor. There have been many attempts to correlate e.s.r. coupling constants with parameters such as Hammett’s 0 constants’ or the 0.scale.8 With a more extensive set of data it would be interesting to perform a similar analysis for the 4-cyanopyridinyls, particularly since the relative changes in hyperfine splittings in these radicals are much larger than in the systems investigated previously. For the present we note that the sequence of 14N coupling constants in table 2 is the same as that found for the 14N splittings in p-substituted t-butylphenyl nitroxides. In DMSO the literature values are P(O)Ph,, 11.6;9 SiMe,, 12.3; GeMe,, 12.7; SnMe,, 12.7; H, 13.0;1° CH,, 13.6 (in glycol).llTable 1. Hyperfine coupling constants (G) for some N-phosphonylpyridinyl radicals OEt CN 1.57 3.40 5.61 0.72 0.72 5.61 1.90 (N) Ph CN 0.73 3.90 5.46 0.73 0.73 5.46 1.96 (N) OEt C02H 1.55 3.42 5.02 0.32 0.39 5.21 1.99 (H) OEt co, 1.52 3.36 5.50 0.64 0.64 5.50 - a Ph Ph 1.35 4.02 4.35 1.35 1.35 4.35 0.97 (2H), 2.65 (2H), 3.01 (H) b s 2 "2.a OEt COOC,,H,, 0.72 3.50 5.54 0.72 0.72 5.54 0.72 (2H) E, 3 Table 2. Hyperfine coupling constants (G) for some N-substituted 4-cyanopyridinyls 9 N-substituent a@) ref. 3.40 3.90 3.98 5.07 5.35 5.36 6.32 5.61 5.46 4.95 4.48 4.35 4.39 4.14 0.72 0.73 0.40 0.14 n.r. n.r. 0.15 1.90 1.96 1.97 1.96 2.00 2.00 1.94 1.57 ("P) this work 0.73 ("P) this work 2 2 2 4.39 (H) 5 5.47 (3H) 6 - - -94 Paramagnetic Adducts of 4-Substituted Pyridines Phosphinyl Radicals Although diphenylphosphinyl radicals were generated in the presence of a variety of substituted pyridines by three different methods, we have been unable to obtain any definitive evidence for the formation of N-phosphinylpyridinyl radicals.The e.s.r. spectra obtained could be analysed in terms of the expected number of hyperfine splittings with a doublet (4-5 G) which might arise from 31P. However, this doublet was replaced by a small 1 : 1 : 1 triplet on addition of deuterated water and must be assigned to an exchangeable proton. We conclude that in each case the observed spectrum is actually that of the appropriate 1 -hydropyridinyl radical, although their mode of formation remains unravelled. P yrazines Radical adducts have also been observed upon reaction of diethoxyphosphonyl radicals with pyrazine and 2,6-dimethylpyrazine ; both species exhibited very small 31P splittings (table 3) in line with those observed for the l-phosphonylpyridinyls and have therefore been assigned structure (11).As was the case with pyridines, we have been unable to add X H diphenylphosphinyl radicals to pyrazine. Indeed the only spectrum observed in these experiments proved to be that of the 1,4-dihydropyrazine radical cation, with a(2N) = a(2H) = 7.33 and a(4H) = 3.05 G. Table 3. Hyperfine coupling constants (G) for radical adducts (11) H 0.64 4.33 7.17 6.52 0.26 (2H) CH3 0.20 4.42 6.52 6.52 0.74 (6H) References 1 2 3 4 5 6 7 8 9 10 11 B. Schroeder, W. P. Neumann, J. Hollaender and H. P. Becker, Angew. Chem., 1972, 84, 894; T. N. Mitchell, J. Chem. Soc., Perkin Trans. 2, 1976, 1149. A. Alberti and G. F. Pedulli, Tetrahedron Lett., 1978, 3283. A. Alberti, M. Guerra, S. Cabiddu and G. F. Pedulli, Gazz. Chim. Ztul., 1979, 109, 647; L. Greci, A. Alberti, I. Carelli, A. Trazza and A. Casini, J . Chem. Soc., Perkin Trans. 2, 1984, 2013. A. Alberti, A. Hudson, G. F. Pedulli, W. G. McGimpsey and J. K. S. Wan, Can. J. Chem., 1985, 63, 917. J. K. Dohrmann and R. Becker, J . Mugn. Reson., 1977, 27, 371. L. Grossi, F. Minisci and G. F. Pedulli, J. Chem. Soc., Perkin Trans. 2, 1977, 943. E. G. Janzen, Acc. Chem. Res., 1969, 2, 279. S. Dincturk, R. A. Jackson, M. Townson, H. Agrbas, N. C. Billingham and G. March, J . Chem. SOC., Perkin Trans. 2, 1981, 1121. K. Torssell, Tetrahedron, 1977, 33, 2287. K. Torsell, J. Goldman and T. E. Petersen, Ann. Chem., 1973, 23 1. H. Lemaire, Y. Marechal, R. Ramasseul and A. Rassat, Bull. SOC. Chim. Fr., 1965, 372. Paper 6/824; Received 28th April, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300091

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday, Trans. I, 1987,83, 95-105 Selective Formation and Conformational Analysis of Carbohydrate-derived Radicals Reiner Sustmann* and Hans-Gert Korth Institut fur Organische Chemie der Universitat Essen, 0-4300 Essen 1, Federal Republic of Germany Carbohydrate free radicals are generated in non-aqueous solution by abstraction of bromine, iodine or selenophenyl with photolytically generated stannyl radicals from appropriately substituted pyranose derivatives and their e.s.r. spectra are recorded in the temperature range from - 30 to 70 "C. From the hyperfine splitting parameters the preferred conformations of these sugar radicals are derived. The findings are compared with results from investigations in aqueous solution and an attempt is made to identify the factors which determine the conformations.From the a-13C coupling constant of the tetra-acetylglucosyl radical it is concluded that pyranosyl radicals are of the n-type. A novel 1,2-migration of an acetoxy group in tetra-acetylgalactosyl radical is reported. The mechanism of interaction of nitrosugars and related compounds with tin radicals is discussed. -~ The study of radical reactions of carbohydrates has been of considerable interest in recent years. Different areas can be distinguished in which carbohydrate free radicals come into view. Radiation chemistry has contributed much to our knowledge of the interaction of hydroxyl radicals with sugars in aqueous solution' and the subsequent fate of these radicals. This in turn is important with respect to an understanding of radiation damage in biological material.Rearrangements and fragmentation of radicals are believed to be responsible for strand breaks in DNA.3 However, free-radical reactions of carbo- hydrates have also received growing attention in preparative organic chemistry. Under proper conditions radical reactions can be of great value in the stereoselective synthesis of natural products, as they are less sensitive to the presence of functional groups than ionic reactions4 This property has been exploited recently for CC coupling reactions of carbohydrates with alkene~.~ 9 E.s.r. spectroscopical measurements can provide direct insight into formation, con- formation and reaction of radicals. The application of this spectroscopical technique to carbohydrate radicals and related systems in solution began with the work of Norman' who used the radiomimetic TilI1-H,O, couple to generate sugar radicals by abstraction of hydrogen with hydroxyl radicals in aqueous solution at low pH.Under these conditions, P-hydroxy-substituted a-hydroxy radicals form carbonyl-conjugated radicals by elimination of water.8 Thus, in the case of D-glucose and D-fructose it was only possible to identify conjugated carbonyl radicals. The drawback of this experimental technique lies in the low selectivity of hydrogen abstraction by OH radicals, and therefore normally a mixture of different radicals is formed. The indiscriminate nature of hydroxyl radicals towards carbohydrates is also known from product ~tudies.~ More recently, Gilbertl0-l3 has continued these studies.In a systematic investigation he analysed the e.s.r. spectra of radicals which were formed from a number of carbohydrates by hydrogen abstraction using the TilI1-H,O, couple in aqueous solution. Even though attack of a carbohydrate by hydroxyl is rather unselective, he was able to deduce coupling parameters for a number of carbohydrate radicals. As these spectra were obtained under conditions quite different from our work, it will be possible to learn from 4 95 F A R 196 Con formational R&\ Analysis of Carbohydrate-derived Radicals R3 Sn- SnR3, hu R = Me. Bun X =-CO-But R3Sn-SnR3 o r RZSnH, hu X = NO2 Fig. 1. Methods of selective generation of carbohydrate free radicals. a comparison something about the influence of the environment and the different substitution patterns on the structure of carbohydrate derived radicals.The results of the above-mentioned CC-coupling reactions in non-aqueous medium motivated us to engage in an e.s.r. spectroscopical study of carbohydrate radicals. The high diastereoselectivity in the reaction of 0-substituted glucosyl radicals with acrylonitrile or methyl acrylate5? requires detailed knowledge of the conformation of the intermediate radical. A a-type radical was invoked as one possibility to rationalise the observations.g* 14-16 Fig. 1 shows the methods by which carbohydrate radicals were obtained regiospecifically, in collaboration with Giese's group. In non-aqueous solution tin radicals are especially suited for the abstraction not only of halogen but also of other functional groups. To avoid complicating effects due to the chain-reaction character of tri-n-butyltin hydride reductions we used the photolytic cleavage of hexa-alkylditin as source of trialkyltin radicals for the abstraction of halogen in tetrahydrofuran or benzene as solvent.In addition, one may take selenophenyl-substituted sugar derivatives and cleave the carbon-selenium bond homolytically by attack with photochemically generated trialkyltin radica1s.l' A general method also exists in the photolytic a-cleavage of sugar ketones. Because of synthetic difficulties this path was generally avoided. Results on the homolytic removal of the NO2 group will be presented in the last part. For most substrates, temperatures above 0 "C have to be applied to obtain good e.s.r.spectra. Glucosyl radicals were obtained in the temperature range from -30 to 40 "C (fig. 2). The most remarkable feature of the e.s.r. spectra are the relatively small coupling constant of the P-hydrogen atom at C-2 and the Occurrence of two y-hydrogen couplings of relatively large magnitude. From this one may infer either a half-chair conformation (B), a B2,5 boat-like conformation (C), a lv4B boat-like conformation (D) or a lC, chair conformation (E). In the conformations (C)-(E) the a-CH bond adopts a nearly equatorial position, i.e. it assumes an orthogonal arrangement with respect to the singly occupied p-orbital at C-1. The conformations (A) and (B) show an almost parallel alignment of the P-C-H bond and the SOMO. The assignment of the y-couplings rests on deuteration experiments. As consequence of deuteration at C-5 the larger of two y-H couplings is replaced by a deuterium tripletR = Ac R .Sustmann and H-G. Korth u 97 I OR 1- RqH OR OR (A) 4 C i - "chair" (B) 4H - " h a l f - chair" (C) B2,5- "boat" ( D) '' 48- "boat" (El C4 - "chair" 1 4C, chair, ( B ) 4H Fig. 2. 0-substituted glucosyl radical and its possible conformations. (A) half-chair, (C) B2,5 boat, (D) ly4B boat and (E) lC4 chair. E.s.r. coupling constants (in mT) as follows. - 30 2.003 1 1.800 1.364 0.348 0.145 20 2.003 1 1.796 1.407 0.345 0.141 of the expected magnitude (calc. 0.053, obs. 0.047 mT). The assignment of the second observed y-coupling then follows automatically. The magnitude of the hydrogen hyperfine coupling at C-3 is rationalised in terms of a fairly good W-like arrangement of the bonds in the C-l/C-3 subunit.The arrangement of the singly occupied orbital at C-1 and the P-CH bond in the five possible conformations (A)-(E) is shown in fig. 3. In (B) this bond is almost eclipsed with the singly occupied orbital. The dihedral angle 6 increases in (C), (D) and (E), making these conformations more plausible on the basis of the observed P-hydrogen hyperfine coupling constant. In these latter conformations, the 8-C-OAc bond adopts a more pronounced parallel arrangement with the singly occupied orbital. Why does the glucosyl radical abandon an energetically advantageous chair confor- mation in favour of a boat or half-chair conformation? Obviously some other interaction compensating this destabilisation must be present.A molecular-orbital (m.0.) theoretical interpretation can be derived which is similar to the explanation of the anomeric effect. The coplanarity of an oxygen lone pair and the C- 1-OR bond maximizes the interaction of the doubly occupied lone-pair orbital and the a*(C-0) m.0. From equilibrium measurements at 25 "C, the magnitude of the anomeric effect in 1,2,3,4,5-penta- 0-acetyl-D-glucose can be calculated to ca. 1.45 kcal mol-l in favour of the a-form.18 The occurrence of this effect is not only restricted to sugar derivatives. In 4-methyl- 2-methoxytetrahydropyran it amounts to 1.3 1 kcal mol-l in dioxane. However, the effect becomes smaller in more polar solvents (e.g. 0.96-1.03 kcal mol-l in acetonitrile).la 4-298 Conformational Analysis of Carbohydrate-derived Radicals (A) (B) (C) (D), (El Fig.3. Newman projections of the C-1-C-2 region in radicals (A)-(E) of fig. 2. * cJ co Fig. 4. Explanation of the quasi-anomeric effect in C-1 glucosyl radicals. In Fig. 4 the argument for the explanation of the anomeric effect is transferred to the glucosyl radical. Here one additional feature comes in to play. Owing to the neighbouring oxygen atom with its lone pair, the SOMO is raised, in this way having a better opportunity to interact with the LUMO a*-C-OR. One may reason that the magnitude of the stabilization is similar to the normal anomeric effect. Having justified an axial arrangement of the P-C-OR bond, we can now proceed in the attempt to delineate a more accurate conformation of the residual part of the glucosyl radical (fig.2). Especially noteworthy is the difference between conformations (B)-(D) and (E). In the latter conformation all substituents are axial, placing some doubt on its energetic feasibility. Important is the magnitude of the dihedral angle between the a-C-H bond and the singly occupied m.0. Conformations (D) and (E) display nearly identical steric arrangements in the C- 1-C-2 region of the pyranosyl ring. Two arguments have to be considered, both being implemented in the m.0. description of the quasi- anomeric effect. In order to maximize the oxygen lone-pair interaction with the SOMO, both orbitals should be placed parallel to each other. This argument favours (B) and (C) over (D) and (E). Secondly, from the value of ca.1.3 mT for the P-C-H coupling constant one can calculate a dihedral angle 8 of 63f5" from the cos28 relationship of P-H coupling constants. This is in better agreement with conformations (C)-(E) than with a half-chair conformation (B). Here, one expects 8 of ca. 30". The discussion does not provide enough criteria to decide between the lr4B conform- ation (D) and the B2,5 conformation (C). However, a distinction can be made between conformations (C) or (D) and (E). The latter involves an interconversion of all substituents from an equatorial to an axial position. A ring inversion of this kind at C-4 is blocked in the bicyclic derivative of fig. 5 . The agreement of the e.s.r. data for this radical with those of the tetra-acetylglucosyl radical can be reconciled only with a boat-like conformation.Our favoured conformation of the tetra-acetylglucosyl radical is displayed in (1). It consists of a slightly twisted B2,5 boat conformation. This assures an optimalR . Sustmann and H-G. Korth 99 Fig. 5. E.s.r. hyperfine coupling parameters for 2,3-di-O-acetyl-4,6-O-benzylidene gluco- pyranosyl radical. T = 16 "C; a(a-H) = 1.802 mT, a(P-H) = 1.200 mT, a(y,-H) = 0.376 mT, a(y,-H) = 0.142 mT. Fig. 6. E.s.r. hyperfine coupling constants for 4,6-O-benzylidene glucopyranosyl radical and 2,3-O-isopropylidene-4,6-O-benzylidene glucopyranosyl radical. T/"C g a(a-H)/mT a(P-H)/mT a(y,-H)/mT a(y,-H)/mT - 2.003 1 1.706 2.893 0.207 0.040 2.003 I 1.393 3.445 0.156 0.084 (A) 30 (B) 17 planar arrangement of the singly occupied p-orbital at C-1 with both the p-type lone pair at the ring-oxygen and the p-C-OR bond, giving maximum electron delocalization.Further, the dihedral angle is as predicted by the P-H coupling constant. Finally, a fair W-arrangement of the p-orbital and the pseudo-axial y-C-H bond is more pronounced here than in the lt4B conformation. Interestingly, a twisted BT,5 boat conformation is very similar to the most stable but not so puckered conformation of cyclohexene. H OR (1) So far all hydroxy groups of the glucosyl radical were protected, thus eliminating the possibility of forming hydrogen bonds. Fig. 6 demonstrates the effect of a partial deprotection, showing free hydroxy groups at C-2 and C-3 of the pyranosyl radical. The P-H coupling is raised from ca.1.35 to 2.89 mT; simultaneously the y-couplings are lowered. Heteroatom-substituted six-membered cycloalkyl radicals which can assume an ideal chair conformation exhibit P-hydrogen couplings of ca. 4.0-4.8 mT. A value of 2.89 mT is significantly lower, nevertheless much larger than in the 0-protected forms. Thus, one can conclude that the radical exists in a flattened chair conformation which enables hydrogen bonding between the adjacent hydroxy groups. In addition, this observation demonstrates that hydrogen bonding is stronger than the energy gain through the quasi-anomeric effect in the slightly twisted BZs5 boat conformation. A second example (fig. 6) gives e.s.r. data for a radical which is blocked in a chair conformation due to its tricyclic structure and which, therefore, cannot take advantage of the quasi-anomeric effect.Its high P-hydrogen coupling supports the interpretation for a retained 4C, conformation.100 Con formational Analysis of Carboh ydrute-derived Radicals Fig. 7. ~t-~~C-Enriched tetra-acetylglucosyl radical. g = 2.003 13. (Splittings in mT.) Estimation of out-of-plane angle: a,(@) = ac(0) + 1190(2 tan2 8) gives 8 = 3.9*, i.e. ca. 1 % 2s character. 15 1.788 1.285 0.358 0.146 4.730 30 1.808 1.308 0.360 0.143 4.751 Compare : ': a('3C) = 4.14 mT; H,C-O-CH,: u('~C) = 4.72 mT. It is also no surprise that in aqueous solution, as is shown in (2), the unprotected glucosan-1 -ylT radical has a /?-hydrogen coupling similar to the partly protected derivative." Thus a delicate balance of the discussed effects finally decides the observed conformation.So far, the whole discussion rests on the presumption that the glucosyl radicals are of n-nature at the radical centre, In order to support this by experimental data, the [a-13C]-tetra-acetylglucosyl radical was generated (fig. 7). A value of 4.73 mT for the a-13C-enriched glucosyl radical is in good accordance with values known for open-chain a-oxyalkyl radicals or even pure alkyl radicals. If one adopts the approach of McKelvey et aI.,l9 one can estimate the 2s character of the semioccupied orbital from the equation given in fig. 7. An approximate s-character of 1% is obtained, pointing to a deviation from planarity of 3.9". Knowing the flatness of the potential-energy curve for defor- mation of radicals from planarity one can postulate that under the conditions of the experiment the radical behaves like an entirely n-type species. The positive temperature gradient of the a-13C hyperfine coupling is in accordance with a negative spin-density at the a-hydrogen atom.Besides glucosyl radicals we generated corresponding species derived from galactose and mannose derivatives (fig. 8). Particularly the C-1 mannosyl radical is of interest. Its e.s.r. spectrum clearly indicates the preservation of the 4C1 chair conformation. This is supported by the small /?-hydrogen coupling constant. This radical already possesses an arrangement of the p-C-OR bond which is coplanar with the SOMO and therefore does not need to interconvert into a B2,5 conformation in order to take advantage of the quasi-anomeric effect.In the galactosyl radical we recognize a molecule with a high /?-hydrogen coupling constant of 2.8 mT. This value together with a vanishing y-coupling at C-3 is consistent t This nomenclature is first introduced in ref. (17).R. Sustmann and H-G. Korth 101 D-mannose D-galactose Fig. 8. Galactosyl radical and mannosyl radical. (Splittings in mT.) 30 1.684 2.759 0.253 - (B) 66 1.712 2.597 0.243 0.03 (C) 20 1.852 0.353 0.307 < 0.03 (A) with a half-chair conformation. This and not an inverted conformation is adopted because the axial substituent at C-4 prohibits a complete conversion into the B2,5 conformation of the glucosyl radical. The substituents at C-3 and C-4 in such a conformation would be almost eclipsed, leading to an energetic situation which over- rides the energy gain of the quasi-anomeric effect.Carbohydrate radicals can also be prepared regiospecifically at carbon atoms C-2 through C-4.20 Bromine abstraction by trialkyltin radicals leads to e.s.r. spectra of sufficiently high signal-to-noise ratio. The 2-deoxy-~-~-glucosan-2-yl radical (3) is characterised by coupling parameters which are typical of a nearly *C, conformation. Two large 8-hyperfine couplings are in accordance with a pseudo-axial arrangement of the two P-CH bonds. The smaller of the two couplings disappears if the hydrogen at C-1 is replaced by deuterium (4). A calculated deuterium coupling constant of 0.35 mT is well matched by the observed value of 0.34mT. The smaller value for the hydrogen coupling at C-1 suggests a dihedral angle, which may indicate a slight conversion into the direction of a half-chair conformation.However, electronic effects due to the presence of neighbouring oxygen atoms can also reduce the coupling constant. The coupling patterns in the a-~-glucosan-2-yl derivatives (5) and (6) can be explained similarly. 0.085 H A c o ~ 1 0 7 A c OMe H H 3. u3 2.280 mT (3) 0.062 Ac AcO 0.340 mT H 2H S. 4w (4) 0. o(15 AcO%; A c H OMe ( 5 ) 3. w2 0.095 H A c O ~ Ac H OMe 3. m1 0. on mT 0.141 .T102 Conformational Analysis of Carbohydrate-derived Radicals A situation which corresponds closely to this is realized in the non-protected glucosan-2-yl radicals, observed by Gilbert et al.ll The P-~-glucosan-2-yl radical (7) shows two large /I-coupling constants.An assignment is possible on the basis of our result for the deuterated 2-deoxypyranosan-2-y1 radical. As expected, one of the large couplings is replaced by a smaller one in the a-form of the radical (8). The equatorial position of the hydrogen atom at C-1 is connected with an increased dihedral angle. 1.285 H o e o H H HO* HO H H H H OH (7) (8) 2.87s I T 2.880 2.280 .T Interestingly, only one doublet splitting is observed for the 2-deoxy-P-~-altro- pyranosan-2-yl radical (9). The 0.26 mT splitting is tentatively assigned to the hydrogen at C-1; its magnitude is consistent with a dihedral angle as predicted by a 4C1 chair conformation. The missing of the second P-H coupling at C-3 implies an almost orthogonal arrangement of this CH bond and the SOMO. In the bicyclic derivative (10) the 4C1 conformation is secured due to the higher rigidity of the system.E.s.r. data for the deoxypyranosan-3- and 4-yl radicals (11) and (12) are compared with corresponding values of radicals (13) and (14). In both situations two large P-hyperfine splittings point to a more or less undistorted 4C, conformation. Owing to the hydroxy groups at C-3 the P-coupling constants are reduced in magnitude in the glucosan-3- and 4-yl radicals. 2.925 2.281 H . . H mT OH 2.100 0.04' s.=* Ac H H OMe (12) 3. !m 0.087 mT 2.:H*H OH H 2.45 mTR . Sustmann and H-G. Korth Ac 0 Ac 0 103 AcO H H OAc H 0 A c R3Sn' 1 AcO Ac &ePh R3Sn' I AcO Ac Br I OAc Fig. 9. Rearrangement of tetra-0-acetylgalactosyl radical. Rearrangements of free radicals can only compete with other reactions like bimolecular self-reaction or disproportionation if they have sufficiently low activation energies2' or if a rigid matrix prohibits bimolecular processes.22 In the course of our studies on pyranosyl radicals we discovered (fig.9) a facile migration of an acetoxy group in the tetra-0-acetylgalactosyl radical to the 2-deoxy-a-tetra-acetylgalactosan-2-yl As is shown in fig. 9, this was proved by independent generation of the latter radical. What is the driving force for this reaction in which an a-alkoxyalkyl radical is trans- formed to a secondary carbon radical? This process goes along with the loss of the quasi- anomeric stabilisation in the half-chair conformation of tetra-acetylgalactosyl radical. One might reason that the driving force for this seemingly contrathermodynamic process is delivered by a lowering of the conformational energy of the product radical or results from a compensation effect. As an a-alkoxyalkyl radical possesses more electronic stabilisation than a secondary carbon radical, the destabilisation in going from the C-1 to the C-2 sugar radical might be outweighed by the gain in anomeric stabilisation, parallel arrangement of a ring-oxygen's lone pair with the P-C-OR bond, after formation of the 2-deoxy-a-tetra-acetylgalactosan-2-yl radical.Recently, denitration of secondary and tertiary nitro compounds by trialkyltin hydride has been shown to be a path of producing carbon-centred radicals which may undergo either reduction by trialkyltin h ~ d r i d e ~ ~ or CC-coupling in presence of 01efins.~~ In an application of this reaction to nitro sugars we were able to detect a paramagnetic intermediate (15).The hyperfine coupling parameter for nitrogen of 2.576 mT (in THF) is close to the value observed for nitro radical anions,26 which led us to the interpre- tation of this spectrum as a radical ion pair of 1-nitropyranosyl radical anion and tri-n-butyltin cation. This also seemed to be supported by the g value of 2.00543 and a small coupling of 0.30-0.34 mT of the tin isotopes l19Sn and l17Sn. Another example is shown in (16), which also displays a different interpretation. The covalently-bonded structure seems to be more reasonable than a radical ion pair on the basis of a comparison with work carried out by DaviesZ7 and Neumann.2s.They obtained similar e.s.r. spectroscopical results when they reacted aromatic and aliphatic nitro compounds with tin radicals. n RO (16) 'OAc104 Conformational Analysis of Carbohydrate-derived Radicals 'CN (b 1 Me3C- N-O-SnBu3 I ? Fig. 10. Rate constants and activation parameters for the decay of alkylstannyloxynitroxyl radicals in benzene solution. (a) 1.04 f 0.08 0.96 13.05k0.56 9.59 f 0.42 (b) 0.41 k 0.07 2.44 8.29f0.10 5.69 f 0.08 In order to obtain more information on the carbon radical-forming process from this paramagnetic intermediate we carried out kinetic measurements. 29 The reaction of the nitro sugar with tri-n-butyltin radicals was run under photolytic conditions at different temperatures using the rotating-sector method. The paramagnetic intermediate proved to be a fairly short-lived species.No carbon-centred radical could be detected even though the decay of the signal is interpreted as consequence of a homolytic cleavage of the carbon-nitrogen bond. The rate constants and the activation parameters are collected in fig. 10. The data follow a clean first-order rate law with high correlation coefficients. Even though this topic deviates from our original theme we would like to pursue it a little further. Guided by the results in the nitro sugar derivative, we undertook a kinetic study of this reaction for some aliphatic nitro compounds. The kinetic parameters for nitro-isobutane follow closely those of the sugar derivative, also being purely first-order. An investigation of secondary and primary nitro compounds reveals that the decay of the e.s.r.signals cannot be described by simple kinetic schemes. Second- and first-order processes are superimposed, the nature of which is not obvious. In all cases, prolonged irradiation leads to the build-up of dialkylnitroxyl radicals. As preparative reductions of nitro compounds with trialkyltin hydride are most successful for tertiary compounds and not possible for primary nitroalkanes one may infer other processes in the latter cases which do not lead to carbon-centred radicals. It is a pleasure to acknowledge the pleasant and intensive cooperation in this project with Prof. B. Giese, Technische Hochschule Darmstadt in whose laboratory the preparative work was carried out. This work was supported by Deutsche Forschungsgemeinschaft and Fonds der Chemischen Industrie.References 1 C. V. Sonntag, Adv. Carbohydr. Chem. Biochem., 1980, 37, 7 . 2 C. V. Sonntag and D. Schulte-Frohlinde, in Eflects of Zonising Radiation on DNA, ed. J. Huttermann, 3 J. F. Ward, J . Chem. Ed., 1981, 58, 135. 4 D. I. Davies and M. J. Parrott, Free Radicals in Organic Synthesis (Springer Verlag, Berlin, 1978); 5 B. Giese and J. Dupuis, Angew. Chem., 1983, 95, 633;'Angew. Chem., Znt. Ed. Engl., 1983, 22, 622; W. Kohnlein, R. Teoule and A. J. Bertinchamps (Springer Verlag, Berlin, 1978). D. J. Hart, Science, 1984, 223, 883.R. Sustmann and H-G. Korth 105 B. Giese and H. Groninger, Tetrahedron Lett., 1984, 25, 2743; J. Dupuis, B. Giese, J. Hartung, M. Leising, H. G. Korth and R. Sustmann, J.Am. Chem. SOC., 1985,107,4332; B. Giese and T. Witzel, Angew. Chem., 1986,98, 459. 6 R. M. Adlington, J. E. Baldwin, A. Basak and R. P. Kozyrod, J. Chem. SOC., Chem. Commun., 1983 944. 7 R. 0. C. Norman and R. J. Pritchett, J. Chem. SOC. B, 1967, 1329. 8 A. L. Buley, R. 0. C. Norman and R. J. Pritchett, J. Chem. SOC. B, 1966, 849. 9 M. Dizbaroglu, D. Henneberg, G. Schomburg and C. V. Sonntag, 2. Naturforsch., Teil B, 1975, 306, 416; M. N. Schuchmann and C. V. Sonntag, J. Chem. SOC., Perkin Trans. 2, 1977, 1958. 10 B. C. Gilbert, D. M. King and C. B. Thomas, J. Chem. SOC., Perkin Trans. 2, 1980, 1821. 1 1 B. C. Gilbert, D. M. King and C. B. Thomas, J. Chem. SOC., Perkin Trans. 2, 1981, 1186. 12 B. C. Gilbert, D. M. King and C. B. Thomas, J. Chem. SOC., Perkin Trans. 2, 1982, 169. 13 B. C. Gilbert, D. M. King and B. C. Thomas, J. Chem. SOC., Perkin Trans. 2, 1983,675. 14 F. Baumberger and A, Vasella, Helv. Chim. Acta, 1983, 66, 2210. 15 J. P. Praly, Tetrahedron Lett., 1983, 24, 3075. 16 B. Giese and J. Dupuis, Tetrahedron Lett., 1984, 25, 1349. 17 H. G. Korth, R. Sustmann, J. Dupuis and B. Giese, J. Chem. SOC., Perkin Trans. 2, 1986, 1453. 18 A. J. Kirby, The Anomeric Eflect and Related Stereoelectronic Eflects at Oxygen (Springer Verlag, Berlin, 19 R. D. McKelvey, T. Sugawara and H. Iwamura, Magn. Reson. Chem., 1985, 23, 330. 20 H. G. Korth, R. Sustmann, K. S. Groninger, T. Witzel and B. Giese, J. Chem. SOC., Perkin Trans. 2, 21 A. L. J. Beckwith and K. U. Ingold, in Rearrangements in Ground and Exited States, ed. P. de Mayo 22 R. Sustmann, D. Brandes, F. Lange and U. Nuchter, Chem. Ber., 1985, 118, 3500. 23 H. G. Korth, R. Sustmann and B. Giese, to be published. 24 D. D. Tanner, E. V. Bhckbmn and G. E. Diaz; J. Am. Chem. Soc., 1981, 103, 1557; N. Ono, H. Miyake, R. Tamura and A. Kaji, Tetrahedron Lett., 1981, 22, 1705; N. 0110, H. Miyake, H. Fuji and A. Kaji, Tetrahedron LRtt., 1983, 24, 3477. 25 J. Dupuis, B. Giese, J. Hartung, M. Leising, H. G. Korth and R. Sustmann, J. Am. Chem. SOC., 1985, 107,4332. 26 A. Berndt, in Landolt-Bornstein, New Series, ed. H. Fischer and K. H. Hellwege (Springer Verlag, Berlin, 1980), vol. 11/9d1, p. 376. 27 A. G. Davies and J. A. A. Hawari, J. Organomet. Chem., 1980,201,221. 28 K. Reuter and W. P. Neumann, Tetrahedron Lett., 1978, 19, 5235. 29 H. G. Korth, R. Sustmann and B. Giese, unpublished results. 1983). 1986, 1461. (Academic Press, New York, 1980). Paper 6/850; Received 1st May, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300095

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

.I. Chern. Soc., Faraday Trans. I , 1987, 83, 107-1 12 Electron Spin Resonance Spectroscopy as an Analytical Tool V. Axelsen and J. A. Pedersen" Department of Chemistry, University of Aarhus, 140 Langelandsgade, DK-8000 Aarhus, C, Denmark Monomethoxy- and dimethoxy-p-benzoquinones have been investigated in alkaline aqueous methanol and ethanol by electron spin resonance spec- troscopy. All compounds show solvent exchange of the methoxy group(s). 2,6-Dimethoxy-p-benzoquinone dissolved in aqueous ethanol exhibits spec- tra of three radicals, due to partial exchange of none, one and two methoxy groups. Computer simulations demonstrate that y-protons of the ethoxy group(s) give rise to splittings comparable to the linewidths. Monomethoxy- p-benzoquinone undergoes alkoxylation solely leading to the 2,5 analogue.Electron spin resonance (e.s.r.) spectroscopy has been relatively little applied as an analytical tool, in spite of its high selectivity and sensitivity. Identification of an unknown compound by way of its e.s.r. spectrum is often rendered difficult because radicals of compounds with differences remote from the radical centre may furnish nearly identical spectra, where others, with closely related structures may yield radicals the spectra of which appear rather different. A further nuisance is the frequent occurrence of composite spectra, when working with biological extracts or reaction mixtures. The application of computers for manipulating digitized spectra then becomes indispensable. Furthermore, data processing makes possible more accurate data analysis and extraction of hyperfine splittings in the mG range, i.e.of a magnitude comparable to the linewidths. We have utilized e.s.r. analytically in a study of a large number of naturally occurring quinones/quinols.'v These compounds are easily extracted by aqueous alcohol and converted to semiquinone radicals by increasing the pH and adding a reducing agent in the case of quinones. Many compounds may be studied when still contained in the crude extract. Quinones and quinols, however, frequently participate in various reactions in alkaline alcoholic solutions, e.g. hydroxylations, alkoxylations and decarboxylations, in addition to solvent exchange. The chance of such reactions must be taken into account, whenever one is analysing unknown spectra. In this paper we report the study of monomethoxy- and dimethoxy-p-benzoquinones in aqueous alcohols, with the aim of following solvent exchange and alkoxylation reactions.The structural study of suitable model compounds is of analytical interest and of interest for the understanding of how naturally occurring alkoxy-p-benzoquinones function, e.g. ubiquinone (coenzyme Q), whch participates in electron transport and contains a 2,3-dimethoxy-p-benzoquinone structure, or the allergen primin (2-methoxy- 6-pentyl-p-benzoquinone) which causes dermatitis. Experimental The p-benzosemiquinone radicals were generated in aqueous ethanol/methanol by reducing the corresponding p-benzoquinones with sodium dithionite, after adjustment 107108 E.S.R. as an Analytical Tool of the pH to the desired value with sodium hydroxide.The solvent composition was ca. 64% alcohol (v/v). 13C methanol was enriched to 90%. The spectra were recorded on a Bruker ER 200 spectrometer with a modulation frequency of 12.5 kHz and a modulation amplitude of 20 mG or less. To avoid saturation effects the microwave power was kept below 0.5 mW in all experiments. Spectra were stored digitally and transferred to a Vax 11/780 computer as 4k discreet signals for final data processing. The resolution was 2.5 mG per signal interval (scan width 10 G). Initially all spectra were simulated by means of a set of parameters obtained by the operator. ' Best-fit' parameters were then obtained by an iterative optimization procedure. By this procedure the splitting constants, offsets (g-factors), linewidths and relative intensities for each radical were adjusted by the computer and the sum of squares of the pointwise deviation between theoretical and experimental spectra calculated as a measure of fitness.The adjustments were continued with increments or decrements of each parameter until a minimum was reached in the sum of squares (best least-squares solution), For composite spectra different linewidths were usually obtained for each spectrum, cf. table 1. For each individual spectrum all the lines were simulated with the assumption of constant linewidth and an overall Lorentzian lineshape. Since semiquinone spectra usually exhibit alternating linewidth effects the assumption of constant linewidth may lead to minor discrepancies between the intensities of some of the lines of the observed spectra and the corresponding lines of the calculated ones.Results and Discussion Solvent Exchange When 2,6-dimethoxy-p-benzoquinone is dissolved in alkaline aqueous ethanol at a pH of ca. 13.3, a complete solvent exchange takes place in the course of minutes. The only e.s.r. spectrum observed is from 2,6-diethoxy-p-benzosemiquinone. At a pH of ca. 12.3, a partial exchange takes place leading to the simultaneous occurrence of three semi- quinone radicals, I, I1 and I11 in nearly equal amounts. Fig. 1 shows the superpositioned spectra of the radicals. A ' best-fit ' simulation obtained from parameter optimizations reveals the three spectra to be derived from 2,6-dimethoxy- (I), 2-ethoxy-6-methoxy- (11) and 2,6-diethoxy-p-benzosemiquinone (111).This is substantiated in the data of table 1, exhibiting a clear consistency among the three radicals. Thus the 8-methyl proton splitting (0.79 G) and the P-methylene proton splitting (0.95 G) are virtually unchanged when going from I to I1 and from I1 to 111, respectively. On the other hand, a, is reduced by 27-28 mG each time a methoxy group is exchanged. This reduction must occur via the oxygen at C( l), rather than at C(2), since no asymmetry appears for radical I1 (a, = a5). If we look at fig. 1 an interesting linewidth increase is observed for the six outermost lines of the spectra of I, I1 and 111. We have marked the lines 1, 2 and 3 and observe an increase in the order 1 < 2 < 3. One might expect some broadening to be introduced, when methoxy groups are exchanged with the more bulky ethoxy groups.However, by adding a small hyperfine splitting in the simulation from the methyl protons of the ethoxy group(s), the computer responds with a reduced linewidth (LwJ for I1 and 111, and reveals a final y-splitting of ca. 30 mG (see table 1). In addition one observes an improved simulation, visualized by a reduced figure for the sum of squares from the least-squares procedure applied (see Experimental). Notice, L,, increases by 13 mG for each ethoxy group incorporated. Fig. 2 shows the simulation of the three spectra of fig. 1 . In order to get an estimate of the relative concentration of the involved radicals the computer calculates the following relative intensity for radical i : intensity = h(i) N(i)V. Axelsen and J .A . Pedersen 109 3 G t I , Fig. 1. E.s.r. spectra of 2,6-dimethoxy- (I), 2-ethoxy-6-methoxy- (11), and 2,6-diethoxy-p- benzosemiquinone (111). Radicals obtained from 2,6-dimethoxy-p-benzoquinone by solvent exchange in aqueous ethanol. where h is the relative height of a single non-degenerate hyperfine component [h (1) is set arbitrarily equal to one], N the sum of the degeneracies of all the components (lines) and L, the linewidth, assumed to be constant for all the lines of the radical. The intensities obtained are shown in table 1. 2,5-Dimethoxy-p-benzoquinone undergoes solvent exchange as outlined for the 2,6-disubstituted quinone above. In concert with a marked linewidth decrease (1 57 to 82mG), we obtain by means of the computer optimization a hyperfine splitting of 50 mG from the six methyl protons of the ethoxy groups (table 1).Unfortunately, we have not been able to obtain a clear-cut spectrum of 2-ethoxy-5-methoxy-p- benzosemiquinone. 2,3-Dimethoxy-p-benzosemiquinone gives rise to a triplet with a splitting of 2.69 G from the two protons at C(5) and C(6). The lack of visual splittings from the methoxy group protons has been explained as steric hindrance, giving the 2,3 analogue a structure which does not correspond with the structures of the other disubstituted analogues3 The observed lines of the triplet (linewidths of 147 mG in aqueous ethanol), probably hide further splittings. Whether solvent exchange has taken place and these splittings are from four methylene protons, has not yet been determined from computer studies.Additions of splittings from four or from six protons in the simulations furnish linewidths of 107 and 105 mG and splittings of 46 and 39 mG in the two cases, respectively. 2-Methoxy-p-benzoquinone undergoes slow solvent exchange when dissolved in alkaline aqueous ethanol at a pH of ca. 12.3. The splitting constants obtained from the radicals of the methoxy (IV) and ethoxy (V) analogues are shown in table 1. A simulation of (V) yields an additional y-splitting of 27 mG from the methyl protons of the ethoxy group and an accompanying linewidth decrease (from 125 to 11 1 mG). The magnitude of this splitting might be incorrect (27 6 11 1). At higher pH, methoxy-p-benzoquinone is attacked by the solvent and the spectra of 2,5-dimethoxy-(VI) and 2,s-diethoxy- p-benzosemiquinone (VII) appear in methanol and ethanol, respectively.There is no indication that other dialkoxy-p-benzosemiquinones are formed, in line with the rule that nucleophilic attack is likely to take place at quinone positions with highest spin density in the corresponding semiquinone, i.e. at C(5).110 E.S.R. as an Analytical Tool Table 1. Splitting constants and linewidths (Gauss) from e.s.r. spectra of monoalkoxy-p- benzosemiquinone and 2,5- and 2,6-dialkoxy-p-benzosemiquinone 0- relative R R' a, = a5 a& ak a&. ak, L,: LW2 intensity I* CH, CH3 1.478 0.790 ~ 0.790 - 0.066 0.066 36% I1 CH, CH,CH, 1.451' 0.787 - 0.950 0.030 0.101 0.079 38% I11 CH,CH, CH,CH, 1.423 0.945 0.028 0.945 0.028 0.124 0.092 26% IV R=CH, 0.600 3.652 2.046 0.806 - 0.078 0.078 82% V R=CH,CH, 0.570 3.618 2.077 0.966 0.027 0.125 0.111 18% 6 a, = a, a{ = a( ah = at ~ - - VI R=CH, 0.172 1.033 0.052 - VII R=CH,CH, 0.213 1.179 0.050 0.157 0.082 - a L,,/LWp linewidth before/after addition of hyperfine splittings in the simulation from the y-protons of the ethyl groups.* I, I1 and I11 simultaneously obtained in aqueous ethanol, IV and V similarly obtained, VI and VII obtained in aqueous methanol and ethanol, respectively. ' The simulation reveals the figures 1.449 and 1.452 for the two splittings at position 3 and 5 . 13C Studies One of the drawbacks of quinone/quinol compounds in relation to e.s.r. studies is that most of their atoms (carbon and oxygen) form 'blind' spots on the spectroscopic mapping.Since the spectral information derives solely from the protons, it is desirable to extend studies of quinones to 13C and 1 7 0 enriched ones. Data of only a few enriched quinones have been reported, probably due to costly and time-consuming ~ynthesis.~ We have used 13C enriched methanol in solvent exchange in order to obtain additional information about the carbon atom of the methoxy groups. We observe splittings of 0.39V. Axelsen and J . A . Pedersen 111 I Fig. 2. Simulated spectra of I, I1 and I11 from data in table 1. and 0.37 G from the methoxy group carbon of 2,5- and 2,6-dimethoxy-p- benzosemiquinone, respectively. The 13C experiments are complicated, however, by the appearance of at least three spectra from any of the disubstituted p-benzoquinones, viz.those from p-benzosemiquinones with none, with one and with two 13C atoms. 2,3-Dimethoxy-p-benzoquinone is easily attacked at C(5) and C(6) by the solvents and a number of unknown reaction products results. The use of 13C enriched alcohols renders valuable information for the identification of these unknowns and should imply the possibility of obtaining a final proof of solvent exchange in the case of the 2,3 analogue. The complexity of the reactions, with superpositioned spectra and additional 13C splittings, makes extensive computer simulations indispensable and precludes immediate results. We are presently engaged in studies with 13C and 1 7 0 isotopes, the results of which will be published in a subsequent paper.Conclusion 13.s.r. spectroscopy is a convenient tool for the analytical study of alkoxybenzoquinones. Structural studies of several alkoxy exchanged derivatives can be made with ease from a few model compounds and extended to 13C and 1 7 0 studies by using isotopically enriched alcohols. Simulations of superpositioned spectra, by means of fast computers, give reliable data for all radicals, including y-hyperfine splittings of magnitude comparable to the linewidths. The digitized spectra must, in this connection, contain a sufficient number of data points (2.5 mG between points in the present study). It should be emphasized, however, that splitting constants of most semiquinones exhibit strong solvent and pH dependen~e.~ Care should be taken in comparing hyperfine splitting data not recorded from radicals exposed to absolutely identical conditions, e.g. aqueous solutions with different alcohols. In the present study the spectra of I, I1 and I11 were obtained with the radicals contained in one and the same sample as were the spectra of IV and V, thereby exposed to identical experimental conditions in either of the two cases. The authors thank Dr J. M. Bruce for samples of 2,5- and 2,6-dimethoxy-p-benzo- quinone.112 E.S.R. as an Analytical Tool References 1 J. A. Pedersen and B. Bllgaard, Biochem. Syst. Ecol., 1982, 10, 3. 2 L. P. K i s t and J. A. Pedersen, Biochem. Syst. Ecol., 1986, accepted for publication. 3 J. A. Pedersen and R. H. Thomson, J . Magn. Res. 1981, 43, 373. 4 J. A. Pedersen, Handbook of EPR Spectra from Quinones and Quinols (CRC Press, Boca Raton, Florida, 1985). Paper 6 / 1002; Received 22nd May, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300107

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1 , 1987, 83, 113-124 Addition-Elimination Paths in Electron-transfer Reactions between Radicals and Molecules Oxidation of Organic Molecules by the OH Radicalt Steen Steenken Max-Planck-Institut fur Strahlenchemie, 0-4330 Mulheim, Federal Republic of Germany Reactions between the 'OH radical and molecules Y that ultimately lead to electron transfer from Y to 'OH have been studied by in-situ radiolysis or photolysis, electron spin resonance and pulse radiolysis techniques with optical and conductance detection. These radical-molecule reactions pro- ceed in aqueous solution via the intermediate formation of covalently bound adducts HO-Y'. These radicals are able to undergo heterolysis, which may proceed by spontaneous or by catalysed paths. The heterolysis results in a one-electron oxidation of Y, and the overall reaction thus consists in a one-electron transfer from the molecule to 'OH.(1) (11) HO'+Y +HO-Y' 4 HO-+Y'+. In the addition step (I), a reducing radical is formed by reaction of the oxidizing 'OH with Y, which is usually neither an oxidant nor a reductant. In the heterolysis step (11), however, the reducing HOY' is converted into the oxidizing Y'+. This phenomenon, termed redox inversion, is the consequence of the change in oxidation state of Y by two units in going from HOY. to Y'+. Examples of redox processes of this kind are given from the class of substituted benzenes and of N-heterocyclics, and structure-reactivity relations governing the heterolysis of HO-Y' are discussed. The Generalized Reaction Scheme The *OH radical (in the following abbreviated as OH) is a particularly important radical owing to its high reactivity with organic and inorganic molecules.It is produced in aqueous systems by ionizing radiation, by photolysis or transition-metal-ion-catalysed decomposition of hydroperoxides and in other reactions, possibly including enzymati- cally catalysed ones. It is known that OH is a strong electrophile, as judged, e.g., by the Hammett p of -0.5 for reaction with substituted benzenes.2* Its reduction potential in aqueous solution has been determined*^ recently with a high degree of certainty to be 2.72-2.74 V us. NHE at pH 0 [the electrode reaction is described by the reverse of reaction (A)], from which a value of 2.32 V can be calculated for the reduction potential of OH at pH 7.That this number indicates strong oxidizing power is seen on comparing it with some well known oxidants such as Fe(CN):- ( E = 0.36 V), IrCli- (0.87 V) or T12+ (2.2 V). In fact, in aqueous solution the OH radical is the strongest of all thermody- namically stable one-electron oxidants, since in the presence of more powerful oxidizing This is based on Part 1 of the lecture given at radicals via addition-elimination, will be published Leeds. Part 2, which refers to reduction of molecules by separately. 113114 Oxidation of Organic Molecules by the OH Radical systems (including electrochemical oxidation) OH is the ' sink ' for oxidizing equivalents uia the reaction sequence (A) or (B): OH--e- - 'OH. (B) With respect to oxidation of solutes, H20'+ is of no significance since its deprotonation occurs within < 1 ps.However, in spite of the high oxidizing capacity, it appears that OH prefers not to react by (outer-sphere) electron transfer but by addition, cf. reaction (1) : ?--OH-+ y*+ 1 +H+ ~ 'OH+Y- HO-Yo k M + 4- HzO+-Y* L H,O+Y'+ (lb) - -H+ The adduct thus formed, HOY', may undergo heterolysis spontaneously [reaction (1 a)] with a rate constant khsp, or the heterolysis may be speeded up or even enforced by protonation of the OH group [reaction ( 1 b)], as a result of which the ability of that group to leave is considerably enhanced (H20 is a much better leaving group than OH-). A complementary way of looking at the catalytic effect of H+ on the rates of heterolysis of the -0-Y* bond is by considering that the electron-withdrawing power of OH is strongly increased by protonation, resulting in the bonding electron pair being pulled even more towards the oxygen.The prediction therefore is that always khH+ 2 khsp. From a redox potential point of view, an estimate can be made for the maximum rate acceleration by protonation. The difference in the potentials for reduction of OH at pH 7 and 0 (0.41 V) corresponds to 9.5 kcal mol-l.7 If this energy contributed fully to the stabilization of the transition state for dehydration, a rate enhancement by the factor 9 x lo6 would result. Reaction scheme (1) is an example of the more general scheme (2), which describes the one-electron oxidation or reduction of a molecule Y by a radical X' ('bonded' electron transfer or ' a d 4 ') : f x+ + kf Q ./ X'+Y - x - Y \x-+y*+ The mechanism involves the formation of X-Y' by covalent bond formation between X' and Y, with the second-order rate constant k,, followed by heterolysis of X-Y', with the first-order rate constant k,, to give either X+ and Y'- or X- and Y'+, depending on whether the electron pair connecting X and Y goes to X or to Y. The reaction is an example for 'bonded' (inner-sphere) electron transfer (e.t.). The kinetic condition for the detection of X Y ' is k, < kf[v {under the (usual) condition that [X'] + [Y]}. If kh > kfM, the reaction will appear to be 'non-bonded' (i.e. outer-sphere e.t.). Since kf is bimolecular (i.e. lo9 dm3 mo1-l s-l) and almost always [Y] < 10 mol dm-3, k, values 3 lo1* s-l are not experimentally accessible. It may be worth commenting further on the reaction type described in eqn (1) and (2).Although XY' is a radical and the overall reaction results in the transfer of a single t 1 cal = 4.182 J.S. Steenken 115 electron (s.e.t.), in the actual electron-transfer step an electron pair is transferred (heterolysis or electron pair transfer, e.p. t.), rather than a single electron. Concerning the direction in which the electron pair is transferred, there are obviously two possibilities that result in either a reduction or an oxidation of the molecule Y by the radical X'. Whether the electron pair goes to X or to Y will depend on the balance of electron- donating and attracting powers of X and Y, which can be described by, e.g., the difference in reduction potentials E(X+/X')-E(Y/Y'-) (for the case of reduction of Y) or P(X'/X-) - l?(Y'+/Y) (for oxidation of Y).The direction of e.p.t. should be determined by whether the double difference [P(X'/X-) - P(Y*+/Y)] - [ E ( X + / X ' ) - E(Y/Y*-)] is positive or negative, a positive value meaning oxidation of Y, a negative one indicating that reduction of Y is thermodynamically favoured. The qualitative conclusions that can be drawn from these considerations are well known. The more electrophilic a radical (i.e. the more positive its reduction potential) the more likely is it to oxidize a molecule; the more nucleophilic a molecule (i.e. the more negative its reduction potential), the more likely is it to be oxidized. The inverse is true for the conditions describing the reduction of a molecule by a radical.In practice the problem is the limited availability of redox potentials. A collection of one-electron redox potentials for organic radicals and molecules has recently been published; cf. ref. (6). In lieu of data relating to the thermodynamics in solution, gas-phase data (i.e. ionization potentials and electron affinities) can be used to estimate the direction and driving force of the electron-transfer reaction. Examples of Addition-Elimination (ad-el) with Electron-pair Transfer (e.p.t.) The Benzene System There are two modes of reaction of OH with organic substrates: addition to a double bond and H-atom abstraction, usually from a saturated carbon. Typically, addition is faster by 1-2 orders of magnitude.' In spite of the high rate constants for addition of OH to double bonds (109-1010 dm3 mol-l s-l), the attachment of OH to different positions of a molecule can be very specific, resulting in a high site selectivity.8? This type of behaviour is observed whenever there are clear differences in the electron densities at the various sites.The further fate of an OH adduct depends on the electron density of Y, the molecule to which OH is attached. The electron density (oxidizability, redox potential) can of course be systematically changed by substituents, e.g. in the benzene system: H H Q OH H QH H Q OH H QH 0- H QH116 via Oxidation of Organic Molecules by the OH Radical -H+ neutral . __* radical In the case of (unsubstituted) benzene, the radical cation, formed by reaction with SO;- lo or by photoionization,ll reacts with water to produce hydroxycyclohexadienyl radical (HCHD), which leads to phenol as one of the final products.12 Based on photoionization experiments with conductance detection, the lifetime of the radical cation is < 20 ns? The reverse reaction, i.e.H+-catalysed dehydration of hydroxy- cyclohexadienyl, seems to be inefficient even in 10 mol dm-, H,SO,, as judged by the non-detectibility by in-situ radiolysis e.s.r. of the benzene radical cation.ll Therefore, in the case of benzene, hydration of the radical cation is practically irreversible. If the electron density of the benzene system is increased by the introduction of a methyl substituent, the rate of H+-induced dehydration of the methylhydroxycyclo- hexadienyl radical(s) is increased and the rate of the reverse reaction, i.e.hydration of the radical cation, is probably decreased, resulting in the mutual conversion becoming quasireversible at ca. pH 2-3. The toluene radical cation is not only able to react by substitution (hydration), but also by elimination (deprotonation from the methyl group)?V lo, 13-15 Since this reaction is irreversible, benzyl is the radical seen at low pH.13-15 At pH 5-6 the lifetime of the toluene radical cation is < 30 ns, and its decay seems to proceed predominantly by hydration to produce isomeric methylhydroxycyclohexadienyl radicals that give rise to cresols. (The cresols were identified by h.p.1.c. with electro- chemical detection. 16) A further increase in the electron density of the benzene system, such as that resulting from the introduction of a methoxy substituent, leads to a pronounced increase in the stability of the radical cation with respect to hydration: If the radical cation is produced at pH 6-7 (by reaction of anisole with SO&-) hydration is not observed (k < lo3 s-l) and the decay is by bimolecular paths.Conversely, production of the methoxyhydroxycyclo- hexadienyl radical@)? at pH < 4 leads to the quantitative and irreversible formation of radical cation.l7~ l8 The radical cation can, however, be converted into the hydroxy- cyclohexadienyl radical by reaction with OH-,18 which is a stronger nucleophile than H20. As compared with benzene, the interconversion of the radical cation and HCHD is now irreversible in the opposite direction, i.e.in favour of the radical cation. When the electron density of the benzene is further increased by replacing OCH, by the even better electron donor OH, conversion? of the HCDH to give the one-electron oxidized species, the phenoxy radical,19 is even more facile and now also takes place via a spontaneous route with a rate constant of lo3 sV1.l9$ It is reasonable to assume that this dehydration reaction involves the elimination of OH- together with deprotonation of the developing radical cation. A contribution of deprotonation and hydration of the incipient proton in the transition state of the OH- elimination step will lower its energy considerably. This effect, rather than the (small) increase in electron density in going from OCH, to OH, may be decisive in making the dehydration in the phenol system faster than in the case of anisole, where the rate of OH- elimination is c lo3 s-l.17 As with anisole, with phenol there is a very efficient catalysis by H+ of dehydration of the HCHD.Sy19 The mechanism involves protonation of the OH group at the methylene carbon, as a result of which the electron deficiency at that position is greatly enhanced: the C-0 electron pair is thereby ‘pulled out’, leaving behind a one- elec tron-deficien t aromatic sys tem.The opposite to ‘pulling’ occurs on OH- catalysis of dehydration: deprotonation of t The different HCDH isomers have different reactivities, CJ ref. (9). $ This rate constant refers to the para-isomer, cf: ref. (9).S.Steenken 117 10 9 8 Y - $ 7 6 0.2 0 -0.2 -0.L o (R) Fig. 1. Correlation of the rate constants for reduction of Fe(CN)i- by R-Ph'-OH with the substituent constants a,(R). The values for OH and OCH, are from ref, (9), those for H and CH, are from ref. (1 1) and (20), respectively, and that for CO, is from ref. (45). The k-value for CH, probably refers to the p-methyl-HCHD. p = - 7.2. the phenolic OH group of the HCHD converts the good electron donor, -OH, into the excellent one, -0-. The electron density of the system is now sufficient to 'push out' the OH group at the methylene carbon, together with the C-0 electron pair. The rate constant for this process is 2 lo7 s-l.199 2o It may be appropriate at this point to suggest that the spontaneous dehydration of the HCHD of phenol proceeds by a push-pull mechanism, with water molecules serving as proton donors and acceptors, e.g.A similar mechanism is probably operative in the case of the OH adduct of aniline, which gives the anilino radical on dehydration.,' As a result of the fact that NH, is a better electron donor than OH, the rate constant for spontaneous dehydration (k = 1.5 x lo5 s-l) is larger than that (k = 1 x lo3 s-l) in the case of phenol. If NH, is replaced as a substituent by the even stronger N(CH,), group, the rate constant for dehydration (which yields the radical cation) increases to 6 x lo6 s-1.22 In fact, the rate constants for OH- elimination have been found to correlate with the ionization potentials of the parent m01ecules.~~ The result of the spontaneous or of the H+- or OH--induced dehydration is an aromatic system which has lost two electrons as compared to the HCHD-stage, or one electron relative to the parent.This means that the formal oxidation state of the HCHD is one unit less, and that the dehydrated system one unit larger than that of the parent. The prediction is therefore that HCHD species should be one-electron reductants and that radical cations should be oxidants, owing to their tendency for reconstruction of the stable parent molecule. This is in fact the case. For example, the (unsubstituted) hydroxycyclohexadienyl radical (formed by addition of OH to benzene) reduces118 Oxidation of Organic Molecules by the OH Radical Fe(CN)i- (E = 0.36 V us. NHE) with the rate constant 1.8 x lo7 dm3 mol-l s-l to give Fe(CN)t- and phenol.l2? 24 If the electron density is increased by substitution of the HCHD by CH30 or HO, the rate constant increases to (2.4-3.6) x lo9 dm3 m o t 1 s - ~ .~ CH,O- or HO-substituted HCHD species are in fact excellent electron donors, as is seen by their ability to reduce even weak oxidants such as methylated benzoq~inones~ (E = -0.24-0.10 V us. NHE).6 Fig. 1 shows that the dependence of the rate constants for reduction of Fe(CN),3- by some substituted HCHD species follows a Hammett relation. The strongly negative value for p (= - 7.2) is not in disagreement with an electron-transfer mechanism. The increase by two units in the oxidation state that results from dehydration of the HCHD leads to an inversion of the redox properties of the radical.For example, whereas, as mentioned above, the HCHD derivatives of methoxy- and hydroxy-benzenes are powerful one-electron reductants, the radical cations or phenoxyl radicals produced from them by dehydration are strong oxidants. [Phenoxyl radicals oxidize a~corbate,~~3 26 N,N,N’,N’-tetramethyl-p-phenylenediamine (TMPD),27 anilines or other p h e n ~ l s ; ~ ~ - ~ ~ methoxybenzene radical cations convert phenols into phenoxyl radicals with rate constants 108-109 dm3 mol-1 s - ~ ; ~ O they oxidize Fe(CN);- lS. 30 and I-.30] The dehydration-induced ‘ redox inversion ’ of the character of the radical is a general phenomenon which is the consequence of the change in oxidation state of the molecule corresponding to the loss of two electrons [cf.scheme ( 5 ) ] : +Red ( + e - ) I +Ox (-e-) I ‘Redox inversion’ is not restricted to aromatic radicals:? it is also observed with aliphatic systems, e.g. in the reaction X OH 0 I I II I I I -C-C’ + -C-C- + HX where the a-hydroxyalkyl radicals are ~ e d u c i n g ~ l - ~ ~ and the alkanon-2-yl radicals have oxidizing propertie~.~~-~~q 34 The p-elimination reaction (6) again involves heterolysis. Further examples (from heterocyclic systems) will be given below. Before moving on to heterocyclic systems, a comment may be made on the thermo- chemistry of the OH reaction with phenol at pH 7. Taking 1.3 V as the reduction potential for the phenoxyl radical at pH 7,289 29 the difference from that of OH (2.3 V) is 1 V, which translates into an exothermicity of 23 kcal mol-l for the one-electron t It is not restricted to radicals either, cf.e.g. the conversion by H+-induced dehydration of triphenyl- carbinols to give trityl cations.S . Steenken 119 oxidation of phenol by OH to give the phenoxyl radical and a molecule of water. In view of this large value it may at first sight appear surprising that OH does not react with phenol by electron transfer. However, in the transition state for the electron-transfer step combination of H+ and OH- will not have occurred [see reaction (4)], and if this is so the driving force is only 12 kcal mol-l (based on the reduction potential for the OH/OH- couple of 1.7 V us. NHE). Furthermore, in contrast to addition to a C=C double bond, electron transfer is unfavourable for entropic reasons since it requires a very drastic reorganization of the solvent shell on the way from the neutral reactants to the ionic electron-transfer products.As a result, addition wins the race and it remains with the solvent molecules to slowly (k z lo3 s-l) finish (by push-pull catalysis of the de- hydration) the job of the electron transfer from phenol to OH. In order to summarize the results presented so far, it can be stated that one-electron oxidation by OH of substituted benzenes proceeds by addition followed by elimination of OH- or of H20. The latter process is usually induced by H+, and it leads to radical cations which may react further, e.g. by deprotonation. The dehydration may also be reversible. Elimination of OH- is observed only if the organic substrates are very easily oxidized, as with phenolates or anilines.The observed strong dependence of the propensity for dehydration of HOY. on the electron density of Y is expected on the basis of the heterolytic nature of the dehydration reaction. N-Heterocyclic Systems Pyridines and Pyrimidines These systems can also be oxidized by OH to give one-electron-oxidized species provided the electron-donating power of Y in HOY’ is sufficiently high. With pyridine, the N-heterocyclic counterpart of benzene, there have to be two methoxygroups as sub- stituents to make the OH-adduct (predominantly that formed by addition to C-3) under- go an H+-induced dehydration:20 G:H2 - )@, + H20. (7) +H + L CH30 G i : H 3 - - H + CH,O OCH3 CH,O OCH, For comparison, with benzene one methoxygroup was sufficient to make dehydration by H+ observable by pulse radiolysis and e.s.r.l7? l8 With 4-pyridone, the N-heterocyclic analogue to phenol, an uncatalysed dehydration of the OH-adduct (to C-3) is not seen36 (in contrast to the situation with phenollg), and the OH- elimination from the ionized OH-adduct [reaction (S)] is much slower &:H 5 &:H - 1.8 x104s-’ k = 6 + OH- + H + pK = 10.0 (i.e.1 x 8 x lo4 s-1)36 than in the case of phenol ( 2 lo7 s-l). However, the H+-catalysed process (given by reaction (9) is seen:120 Oxidation of Organic Molecules by the OH Radical These differences in behaviour of the OH-adducts reflect the lower electron density of pyridine as compared to benzene systems, owing to the higher electronegativity of N as compared to C.In many cases dehydration of OH-adducts of heterocyclics is only observed if these contain strongly electron-donating substituents such as NH,, OH and, particularly, 0- conjugated with the 7t system. If OH is the substituent (or if there is a keto group that can be converted into OH by enolization) dehydration will be facilitated in basic solution owing to conversion by deprotonation of OH into the even stronger electron-donor 0-. With amino substituents the accelerating effect of these on dehydration is expected to be changed into the opposite as a result of their protonation (which occurs at pH < 7 with most N-heterocyclics), by which an electron-donating group is converted into an electron-wi thdrawing one. Examples of these mechanistic concepts are the reactions of OH with naturally occurring pyrimidines and purines.Concerning pyrimidines, OH has long been known to react by addition to the C-5/C-6 double bond. The adducts thus formed differ considerably with respect to their redox 38 the 6-yl radical (formed by OH addition to C-5) is a reductant [with respect to the oxidant tetranitromethane (TNM)], whereas the 5-yl radical (produced by OH addition to C-6) is an oxidant [relative to the reductant N,N,N’,N’-tetramethyl-p-phenylenediamine (TMPD)]. Owing to these differ- ences in redox behaviour, the isomeric radicals can be individually titrated, and by this method a mass balance accounting quantitatively for the fate of OH in its reactions with various substituted uracils3’ and cytosine~~~ was obtained. ox>!H oxidizing H H H reducing H Although these differences in redox properties of isomeric radicals are interesting in their own right, the changes in redox nature that these radicals may undergo are more relevant to the main theme of this discussion.Fig. 2 shows that in the system OH +cytosine the ratio of oxidizing to reducing radicals changes drastically from ca. pH 7 to ca. pH 12. The increase with pH in the oxidizing equivalents (titrated by TMPD) is mirror-imaged by the decrease in the reducing equivalents (titrated by TNM). Obviously the reducing radicals are converted into oxidizing ones by reaction with OH-. The explanati~n~~ involves OH- catalysed dehydration of the main radical, i.e. that produced by addition of OH to C-5 of cytosine, as shown below:S.Steen ken 121 I 6 10 1L PH Fig. 2. Dependence on pH of the yields of oxidizing (measured as TMPD-+; 0 ) and reducing radicals (measured as nitroform anion; A) produced on reaction of OH with cytosine. reducing ox id iz i ng Conversion of the amide function into the corresponding enolate is required to make the electron density of the system large enough to enable the C-0 heterolysis to take place. In support of the enolate as a necessary intermediate is the fact that (N-1)-alkylated pyrimidines (such as the nucleosides and nucleotides) do not show the dehydration r e a ~ t i o n . ~ ~ - ~ ~ The mechanism is also corroborated by the observation that electron- donating substituents like methyl at N-3, C-5 or C-6 increase the rates of dehydration of the C-5 OH adducts, whereas electron-withdrawing ones (at C-5 and C-6) decrease the rates.37* 38 The dehydrated OH adduct, which is identical with the product of a one-electron oxidation reaction followed by deprot~nation,~~ is oxidizingg1 (with respect to TMPD).38 The conversion of the reducing OH adduct into the oxidizing elimination product is another example for the phenomenon of ‘redox inversion’ by dehydration of OH adducts.Oxidation by OH of pyrimidines via addition/elimination is not restricted to the pyrimidines of biological importance such as the uracils and cytosines. This mechanism is observed also in the case of 4,6-dihydroxypyrimidines7 which may be considered isomers of the uracil system. As with the uracils and cytosines, the rates of dehydration of the neutral OH adducts (to C-5) are low (< lo3 s-l), and only after deprotonation does the electron density of the system become sufficient to make the C-0 heterolysis [eqn (12)] become observable by e.s.r. (see fig.3) and time-resolved optical detection. Owing to their electron-donating effect, methyl groups enhance considerably the rates ofdehydration of the OH-add~cts.~~ This again is similar to the situation with the uracils and cytosines.122 Oxidation of Organic Molecules by the OH Radical g=2.004 17 @ P Fig. 3. E.s.r. spectrum recorded on reaction of OH with 0.5 mmol dm-3 5-methyl-4,6-di- hydroxypyrimidine at pH 10 and ca. 5 "C. In the insets the field axis is expanded by a factor of 2.5. 0- 0- /!OH' Purines The reactions of the OH radical with purines have been studied for more than a generation using a variety of techniques (e.g.product analysis, pulse radiolysis, e.s.r. spectroscopy). However, so far the success of these techniques has been inconspicuous. This is largely due to the failure of liquid solution e.s.r. spectroscopy to provide information on the radicals formed. The reason for this is essentially the insufficient signal-to-noise ratio. In contrast, pulse radiolysis with optical detection does not suffer from bad signal-to-noise ratios, but the spectra are particularly uncharacteristic in that little structural and mechanistic information can be extracted from them. Conductance detection techniques have not yielded much information either, because at physiological pH the OH adducts of the naturally occurring purines do not give rise to conductance signals. However, using the redox scavenger technique it has recently been found that the reaction of OH radicals with guanosine and adenosine leads to roughly 1 : 1 mixtures of reducing and oxidizing radicals,43 although their identity could not be unambiguously determined.in our laboratory that replacement of the amino substituent in adenosine by the N,N-dimethylamino substituent leads to (a) better resolved absorption spectra (compared to adenosine) and (b) the observation that these absorption spectra undergo first-order changes with two different rate constants observable at different wavelengths. These changes are due to three radicals, two of which rearrange with the same rate constant. The two faster of the transformation reactions that the OH adducts undergo are It has beenS.Steenken 123 accompanied by an increase in conductance at pH 3 7. These conductance changes indicate that OH- is eliminated from the OH adducts. The conductance (and some of the optical) changes can be inhibited by 0,, but only in part. From a detailed comparison of the effect of 0, and other redox scavengers on the unimolecular changes detected optically and by conductance and from the analysis of the pH dependence of the transformation reactions it was concluded that there are two isomeric OH adducts responsible for the elimination of OH- and that they differ by their redox properties, whereas the rates and the activation parameters for OH- elimination are the same. The isomers were identified as A-4-OH and A-SOH, arising from addition of OH C-5 of the purine system, i.e.A-4-OH + 6H at C-4 and 'N' R A-8-OH The product of the elimination reactions A and B is the radical cation, oxidizing properties (relative to TMPD) in contrast to its precursors, which has which are reductants [relative to Fe(CN)i-], whereby A-5-OH is a much stronger reductant than A-4-OH. The dehydration reactions A and B are further examples of redox inversion. The third isomer formed, A-8-OH, undergoes opening of the imidazole ring, a proces that does not result in redox inversion.44 Summary It has been shown that OH-radical-induced oxidations of homo- and hetero-cyclic compounds with double bonds proceed by addition followed by elimination of OH- or H,O. Elimination of OH- occurs with rates observable by fast detection techniques (k 2 102-103 s-l) only if the electron density of the molecule from which OH- is split off is quite high, as with systems substituted by amino or (ionized) OH groups.The catalysis by OH- of the heterolysis of the OH-adducts involves deprotonation of the non-methylenic OH groups attached to the molecule (or the conversion of keto into enolate groups) followed by electron-pair donation from -0- to aid in the rupture of124 Oxidation of Organic Molecules by the OH Radical the methylenic C-0 bond. In other words, OH- catalysis results in an enhancement of the electron-pair-donating abilities of the molecule to which -OH is attached. In contrast, catalysis by H+ leads to a strong increase in the electron-pair-accepting properties of -OH, i.e.to an improvement of the leaving-group abilities. References 1 S. Steenken, to be published. 2 M. Anbar, D. Meyerstein and P. Neta, J . Phys. Chem., 1966, 70, 2660. 3 P. Neta and L. M. Dorfman, Adv. Chem. Ser., 1968, 81, 222. 4 H. A. Schwarz and R. W. Dodson, J. Phys. Chem., 1984,88, 3643. 5 U. K. Klaning, K. Sehested and J. Holcman, J . Phys. Chem., 1985,89, 760. 6 S. Steenken, Landolt-Bornstein, 1985, 13e, 147. 7 Farhataziz and A. B. Ross, Nut. Stand. Ref. Data Ser., Natl Bur. Stand., 1977, 59, 66. 8 (a) M. K. Eberhardt, J. Org. Chem., 1977, 42, 832; M. K. Eberhardt and M. I. Martinez, J . Phys. Chem., 1975, 79, 1917. (b) C. Walling and D. M. Camaioni, J . Am. Chem. Soc., 1975, 97, 1603. (c) C. Walling and R. A.Johnson, J. Am. Chem. SOC., 1975,97, 363. 9 (a) S. Steenken and N. V. Raghavan, J. Phys. Chem., 1977, 83, 3101. (b) N. V. Raghavan and S. Steenken, J. Am. Chem. SOC., 1980, 102, 3495. 10 P. Neta, V. Madhavan, H. Zemel and R. W. Fessenden, J . Am. Chem. SOC., 1977,99, 163; H. Zemel and R. W. Fessenden, J . Phys. Chem., 1978,82, 2670. 11 S. Steenken, unpublished results. 12 G. W. Klein and R. H. Schuler, Radiat. Phys. Chem., 1978, 11, 167; V. Madhavan and R. H. Schuler, 13 R. 0. C. Norman, P. M. Storey and P. R. West, J . Chem. SOC. B, 1970, 1087; 1099. 14 C. Walling, Acc. Chem. Res., 1975,8, 125; C. Walling, G. M. Taliawi and K. Amarnath, J. Am. Chem. 15 K. Sehested, J. Holcman and E. J. Hart, J. Phys. Chem., 1977,81, 1363; K. Sehested and J. Holcman, 16 S. Steenken, unpublished results. 17 P.O’Neill, S. Steenken and D. Schulte-Frohlinde, J. Phys. Chem., 1975, 79, 2773. 18 J. Holcman and K. Sehested, J . Phys. Chem., 1976, 80, 1642. 19 E. J. Land and M. Ebert, Trans. Faraday SOC., 1967, 63, 1 181. 20 S. Steenken, unpublished results. 21 H. Christensen, Znt. J . Radiat. Phys. Chem., 1972,4, 31 1 ; Ling Qin, N. R. Tripathi and R. H. Schuler, 22 J. Holcman and K. Sehested, J . Phys. Chem., 1977, 81, 1963. 23 J. Holcman and K. Sehested, Nukleonika, 1979, 24, 887. 24 0. Volkert and D. Schulte-Frohlinde, Tetrahedron Lett., 1968, 17, 21 51. 25 R. H. Schuler, Radiat. Res., 1977, 69, 417. 26 P. O’Neill, D. Schulte-Frohlinde and S. Steenken, Faraday Discuss. Chem. SOC., 1978, 63, 141. 27 S. Steenken, J . Phys. Chem., 1979,83, 595. 28 S. Steenken and P. Neta, J. Phys. Chem., 1979, 83, 1134. 29 S. Steenken and P. Neta, J. Phys. Chem., 1982, 86, 3661. 30 R. Schrank, Dissertation (Ruhr-Universitat Bochum, 1980). 31 K. M. Bansal, M. Gratzel, A. Henglein and E. Janata, J. Phys. Chem., 1973, 77, 16. 32 P. Neta, Adv. Phys. Org. Chem., 1976, 12, 2. 33 A. J. Swallow, Progr. React. Kinet., 1978, 9, 195. 34 A. Henglein, Electroanal. Chem., 1976, 9, 163. 35 S. Steenken, M. J. Davies and B. C. Gilbert, J. Chem. SOC., Perkin Trans. 2, 1986, 1003. 36 S. Steenken and P. ONeill, J . Phys. Chem., 1979, 83, 2407. 37 S. Fujita and S. Steenken, J. Am. Chem. SOC., 1981, 103, 2540. 38 D. K. Hazra and S. Steenken, J. Am. Chem. SOC., 1983, 105, 4380. 39 S. Steenken and V. Jagannadham, J. Am. Chem. SOC., 1985, 107, 6818, ref. (36). 40 K. M. Bansal and R. W. Fessenden, Radiat. Res., 1978, 75, 497. 41 K. M. Bansal and R. M. Sellers, in Fast Processes in Radiation Chemistry and Biology, ed. G. E. Adams, 42 H. M. Novais and S. Steenken, J. Phys. Chem., 1986, in press. 43 (a) P. ONeill, Radiat. Res., 1983, 96, 198; (6) P. ONeill, in Life Chemistry Reports Supplement, Oxidative Damage and Related Enzymes, ed. G . Rotilio, J. V. Bannister (Harwood Academic Publishers, London, 1984), p. 337; (c) P. O’Neill, P. W. Chapman, Znt. J. Radiat. Biol., 1985,47,71; ( d ) P. O’Neill, P. W. Chapman and D. G. Papworth, Life Chemistry Rep., 1985, 3, 62. 44 A. J. S. C. Vieira and S. Steenken, J. Am. Chem. SOC., 1986, submitted. 45 G. W. Klein, K. Bhatia, V. Madhavan and R. H. Schuler, J . Phys. Chem., 1975, 79, 1767. Radiat. Phys. Chem., 1980, 16, 139. SOC., 1984, 106, 7573. J . Phys. Chem., 1978, 82, 651 ; K. Sehested and J. Holcman, Nukleonika, 1979, 24, 941. 2. Naturforsch., Teil A, 1985, 40, 1026. E. M. Fielden and B. D. Michael (Wiley, New York, 1975), p. 259. Paper 6/ 1133; Received 5th June, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300113

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I , 1987, 83, 125-133 Radical Production evidenced by Dimer Analysis in y-Irradiated Amides in Aqueous Solutions and in the Solid State Anne-Catherine Dusaucy, Joelle De Doncker, Christine Couillard, Marc De Laet and Bernard Tilquin Catholic University of Louvain, Radiation Chemistry Laboratory (CHAW, Place Louis Pasteur, 1, B-1348 Louvain-la-Neuve, Belgium A study of the nature and distribution of radicals produced by the radiolysis of amides in the solid state has been made by two complementary methods : (i) final product analysis by capillary gas chromatography and (ii) the direct observation of radicals by electron spin resonance spectroscopy. The dimer isomer distribution for N-methylformamide (NMF), N-methylacetamide (NMA) and N,N-dimethylformamide (DMF) allows one to estimate the parent radical distribution by using a statistical rule for dimer formation.In the solid state the radiolytic formation of radicals is selective. Final product analysis by capillary gas chromatography and the direct study of radicals by pulse radiolysis are also complementary methods for studying the nature of radicals produced from the radiolysis of amides in aqueous solutions. The selective production of one parent radical is characteristic of dilute solutions. The purpose of this paper is to show how dimer analysis can be of some help in the identification of transient radicals. The direct observation of radicals produced by y-radiolysis of solid amides can be made by e.s.r. measurements and optical spectroscopy at low temperatures.'? For radicals trapped in irradiated solids, the e.s.r.spectrum, characterized by broad lines, represents the superposition of several spectra ; the interpretation of the individual spectrum in the solid state is uncertain and has to be confirmed by independent methods such as the analysis of the final prod~cts.~ Final products, formed by reactions between these radicals, are analysed by capillary gas chromatography. High-performance chromatography allows the separation of dimer isomers, so that the identification and distribution of the radicals can be deduced from qualitative and quantitative analysis, respectively. The first part of the paper deals with the analysis of dimer isomers produced by y-radiolysis of N-methylformamide (NMF), N,N-dimethylformamide (DMF) and N-methylacetamide (NMA) in the solid state at 77 K.The results for N,N-dimethylacetamide (DMA) are described in a previous paper.4 Although many papers have been published concerning pulse-radiolysis studies of organic radicals in aqueous sol~tions,~ very little work appears to have been done on free radicals produced from the radiolysis of simple amides in aqueous solutions.6. The technique of pulse radiolysis, coupled with absorption spectrophotometry, suggests that the radicals formed by abstraction of an H atom (parent radicals) from amides in aqueous solutions may exist in different isomeric forms, depending on the relative amount of water.6 Dimer isomers, as produced by the combination reaction of two parent radicals, preserve the structure of these two radicals.Thus final product analysis is useful to check the isomeric forms of the parent radicals. The second part of this paper provides an analysis of the dimer isomers produced by y-radiolysis of N-methylacetamide (NMA) and N,N-dimethylformamide (DMF) in aqueous solution. 125126 Radical Production in y-Irradiated Amides Experimental NMA, obtained from Merck, was purified by the method of Pucci et aZ.* and fractionally distilled. A g.1.c. analysis of the purified samples on a capillary column gave a final purity of 99.99%. NMF, DMF and DMA (Fluka) were fractionally distilled; their purity, determined by gas chromatography, was 99.5%. Before irradiation in the solid state, the amides were stored over a drying agent: molecular sieve 4 A for DMF and NMF, CaO for NMA.H,O-NMA or H,O-DMF mixtures were prepared by weighing with triply distilled water. Concentrated solutions (70-80% amide) and dilute solutions (25 % amide) were prepared. Samples were degassed in ampoules using freeze-pumpthaw cycles on a vacuum line, whereupon the ampoules were sealed. Irradiation of the solid samples quenched in liquid nitrogen was carried out at 77 K with 6oCo y-rays, at a dose rate of ca. 4 kGy h-l. The total doses absorbed were ca. 150 kGy. For irradiation of the aqueous solutions at 298 K the dose rate was ca. 3 kGy h-l and the total doses absorbed by the sample ranged from 20 to 140 kGy. The samples were bleached by visible light following irradiation. The final products of radiolysis were analysed by gas chromatography on an open tubular column (25 m x 0.32 mm) coated with Carbowax 20M.These analyses were carried out on a Varian Aerograph 3700 instrument with a flame ionization detector (FID) and splitter (1 /50) injector. Identification of the dimer isomers was made by g.c./m.s. coupling on a Finnigan Mat 4500 mass spectrometer. All mass spectra are available on request. Several dimer isomers were synthesized : N,N-dimethyloxamide from NMF, N,N-diformyl-N,W- dimethylethylenediamine and N-formyl-N-methylaminoacetyldimethylamide from DMF, and N,N-diacetylethylenediamine from NMA. Their structures were determined by lH n.m.r. spectroscopy at different temperatures (Varian XL 200). These dimer isomers were used as internal standards in the g.c./m.s. procedure.In this work, response factors of the FID were not determined; however, these factors must be similar for all the dimer isomers of one compound. Results Solid State NMF The following radicals correspond to the parent radicals : 'CONHCH, (I), HCON'CH, (11) and HCONHCH; (111). The chromatogram of fig. 1 is obtained when NMF is irradiated at 77 K: no peaks correspond to dimer isomers. However, dimer 11-11, 11-111 and 111-111 isomers are detected from irradiated liquid NMF. The radical produced by NMF photolysis in aqueous solutions is assigned to the parent radical (III).9-11 DMF The parent radicals are HCON(CH,)CH; (I) and 'CON(CH,), (11). Fig. 2 shows the chromatogram of solid irradiated DMF. Three dimer isomers are detected with the following distribution: dimer,-, 70.6% ; dimer,-,, 23.5% and dimer,,-,, 5.9%. An analysis of the dimer isomers suggests the production and trapping of the two parent radicals; the e.s.r.spectrum1 of solid DMF irradiated at 77 K can be interpreted as the spectrum of parent radical I.A-C. Dusaucy, J . De Doncker, C . Couillard, M . De Laet and B. Tilquin 127 L 4 Fig. 1. Capillary gas chromatogram of solid irradiated N-methylformamide (160 kGy). Fig. 2. Capillary gas chromatogram of solid irradiated NMA The e.s.r. spectrum of NMA y-irradiated at 77 K 1 N,N-dimethylformamide ( 1 60 kGy). c becomes, on heating (100 K), the spectrum attributed to CH,CONCH,CH, (I) by Pukhal'skaya et aL2 The other parent radicals, CH,CON'CH, (11) and 'CH,CONHCH, (111), are not detected by e.s.r. meas- urements at 100 K.According to the final product analysis (fig. 3), radiolysis of solid NMA at 77 K leads to the formation of three dimer isomers with the following distribution: dimer,-, 83.2% ; dimer,-,,, 9.9% ; dimerIII-~I, 6.9%. Dimer isomers produced by the parent radical I1 are not detected. Aqueous Solutions Fig. 4 shows a typical chromatogram of irradiated diluted NMA aqueous solution. Only one dimer isomer is detected by g.c./m.s. procedure. From the three possible parent radicals : CH,CONHCH; (I), CH,CON*CH, (11) and 'CH,CONHCH, (111), only the first leads to dimer formation. Fig. 5 shows a typical chromatographic trace of irradiated concentrated NMA aqueous solution. Five dimer isomers are detected by the g.c./m.s. procedure; the most abundant is the dimer formed by the combination of two parent radicals of type I.Two 5 FAR 1128 (CH3CONHCH2+2 t Dimer t Radical Production in y-Irradiated Amides - Fig. 3. Capillary gas chromatogram of solid irradiated Dimer r-methylacetamic: (1 10 kGy). J - I L Fig. 4. Capillary gas chromatogram of diluted NMA aqueous solution (26 wt %) irradiated by y-rays at a dose of 38 kGy. Fig. 5. Capillary gas chromatogram of NMA-H,O mixture (60-80 wt %) irradiated by y-rays at a dose of 21 kGy.A-C. Dusaucy, J . De Doncker, C. Couillard, M. De Laet and B. Tilquin 129 L Fig. 6. Capillary gas chromatogram of a DMF-H,O (50 wt %) mixture irradiated by prays at a dose of 80 kGy. dinier isomers are eluted in the NMA peak and the three others after this peak (the dark peaks in fig. 5). The analysis of final products of DMF irrdiated in aqueous solutions reveals three dinier isomers (fig.6). The dimer ((CH,),NCO), is the less abundant (< 2% relative area); the dimer (HCON(CH,)CH,), yields more than 85% of the dimer isomers. Results are similar in both concentrated and dilute aqueous solutions (1 1-73 wt % for DMF). In the radiolysis of DMF solutions, two parent radicals are possible: HCON(CH,)CH; (I) and 'CON(CH,), (11). Discussion Solid State With regard to the dimer product formation following radiolysis at 77 K, previous works indicate immediate and delayed processes. l2 Thus the combination of trapped radicals is only one part of the product formation. Even for these radical processes, dispro- portionation or conversion reactions may compete with those of radical combination.Therefore, the correlation between results from e.s.r. measurements and chromato- graphic data is not immediate. However, radical processes always contribute to the formation of dimer isomers,13 so that final product analysis is very useful in displaying the trapped radicals, especially for radicals in very low relative concentrations.12 The shape of the e.s.r. spectra in the solid phase is due to interactions which may be fairly complicated, and line-broadening may delete the hyperfine structure. Moreover, for a large number of compounds containing nitrogen, even for small disturbances of electronic state, sensitivity is apparently a common phenomenon.14 It is therefore useful to use the analysis of products formed by radical combination in order to improve and complete the e.s.r.results. 5-2130 Radical Production in y-Irradiated Amides Mechanism of Formation of the Dimer Isomers We may assume that the mechanism of formation of the dimer isomers obeys a statistical rule represented by the following diagram : / AB AB. A x A - A A B-B-BB The percentage distribution of dimer isomers is often satisfactorily described by the equation : where % is the percentage of each isomer. Eqn (1) is verified for a wide range of doses in solid aliphatic hydrocarbons,12 experimental values ranging from 1.3 to 7.6. Deviation from the value of 4 is explained by the portion of radicals which react by disproportionation reactions. This varies on going from primary to tertiary alkyl radicals. In this work, the following values are calculated for eqn (1): = 1.3 (% dimerI-II)2 (23.5)2 DMF: (% dimer,_,) (% dimer,I-II)= (70.6) (5.9) (% dimer,-,) (% dimerIII-III)= (83.2) (6.9) = 0.2.(% dimerI-III)2 (9.9)2 NMA : The very low value for the radicals produced in NMA does not correspond to a statistical recombination reaction. This result suggests that some of the dimer isomers are produced by an ionic mechanism during radiolysis (immediate processes). The very low proportion of the cross-dimer (1-111) is also indicative that most dimer,_, is produced by immediate non-radical processes. Spec$city of Radical Production in Solid Radiolysis Radiolysis of DMA at 77 K produced only one dimer i~omer,~ (CH,CON(CH,)CH,),. This specificity in the formation of one parent radical (CH,CON(CH,)CH;) is due to a secondary reaction at 140 K: the abstraction of an H atom by a fragment radical on a solvent molecule: (2) E.s.r.measurements at 130-140 K showed an irreversible conversion of the signal, in addition to a decrease in the total radical concentration. Reaction (2) is in competition with radical recombination reaction^.^ As far as final product analysis is concerned, the specificity in the production of one parent radical in the solid state, observed in DMA at 77 K,4 is unique. "(CH,), + CH,CON(CH,), --+ HN(CH,), + CH,CON(CH,)CH;. Selectivity of Radical Production in Solid Radiolysis E.s.r. and g.c. results are in good agreement for the selectivity of radical production in the solid radiolysis of arnides. Not all the possible parent radicals (R') are produced. However, according to the number of dimer isomers detected, g.c.results do not agree with the specificity of the radical production proposed by e.s.r. measurements. In DMF and NMA samples, more than one parent radical are produced and lead to dimerA-C. Dusaucy, J. De Doncker, C. Couillard, M. De Laet and B. Tilquin 131 formation. In NMF samples, conversion reactions of trapped radicals may explain the absence of dimer isomers; another explanation is that trapped radicals are not parent radicals. Kinetic studies are needed to answer this question. Note that preliminary studies of the radiolysis of amides in the liquid state do not show such a selectivity in the formation of parent radicals. For example, g.c. analysis of liquid irradiated NMA presents five dimer isomers. Radical Distribution Our approach is pragmatic; corrections (immediate non-radical processes etc.) are not taken into account in the estimation of the radical distribution.With the following equations the distribution of radicals may be calculated when the results verify statistical rules : % [A'] + % [B'] = 100 (4) In these equations, as the kinetics of radical combination is limited by diffusion, the rate-constant values are assumed to be similar (kAA z k,, x kBB). As radical combi- nation in the solid state is not totally statistical, a mean value for the radical distribution must be calculated (extreme values are eliminated). For DMF radicals, the estimated radical distribution of parent radical I ranges from 82 to 89%.Aqueous Solutions The spectrum obtained by pulse radiolysis may belong to several transient species; the individual absorption spectrum for each radical is not definitely known.6 Product analysis is used to confirm the assignment of radicals. Diluted Soh t ion The radiolysis of water is known to lead to the formation of reactive species (OH', eLq, etc.) which are available to interact with any solute present in water.5 Hydroxyl radicals abstract hydrogen atoms from carbon-hydrogen bonds; hydrogen atoms also abstract hydrogen atoms, but the rate constants are generally ~maller.~ In a diluted aqueous solution of NMA (fig. 4) only one dimer isomer is formed in detectable quantities [dimer,-, or (CH,CONHCH,),]. The other dimer isomers that could be formed by the combination of parent radicals I and I1 were not detected, even by the g.c./m.s.procedure using chemical ionization in the MIS mode. Dilute aqueous solutions of DMF (fig. 6) have a distinctive radiolytic behaviour: all the dimer isomers which can be produced by radiolysis are detected. The measured distribution of the dimer isomers is as follows: dimer,-, 85.4%, dimer,-,, 13.6% and dimer,,-,, ca. 1 % . Of the parent radicals 92% are radical I isomers, assuming a statistical radical combination reaction. Specific (NMA) or selective (DMF) formation of one type of radical isomer may be due to the production of radicals by secondary mechanisms. The analysis of final products is in good agreement with pulse radiolysis for the assignment of the radicals in irradiated NMA aqueous According to Hayon et uZ.,' the abstraction reaction (6) is very selective: (6) CH,CONHCH, +OH' -, CH,CONHCH; + H,O.132 Radical Product ion in y - Irradiated Am ides Analysis of final products suggests that this reaction is specific with NMA and selective with DMF.The most important feature of the radiolysis of amides in aqueous solution is the formation of other products: the hatched peak in fig. 6 is not formed after the radiolysis of pure DMF. Some experiments with NMF in aqueous solution showed that the main effect of the radiolysis is the production of compounds other than dimer isomers. It is necessary to introduce N,O into the samples before irradiation in order to detect the dimer isomers. In this case the amount of OH' is doubled: eiq + N,O + OH- +OH' + N,.(7) Another characteristic of these preliminary experiments in NMF is the low yield of cross-dimer (21.2% of the dimer isomers); in the pure phase the cross-dimer yields 56% of the dimer isomers. Concentrated Solutions For the concentrated aqueous solutions of NMA, the radiolysis of H,O plays a minor part. NMA is a highly structured medium, able to form hydrogen bonds, and it has a tendency to aggregate.15 The chromatogram in fig. 5 is the same after the irradiation of pure NMA or of H,O-NMA mixtures (H,O ---* 40 wt % ). The dimer,-, isomer is still the main heavy product of the radiolysis. Nevertheless, four other isomers are detected by chemical ionization in the g.c./m.s. coupling. Corresponding to these dimer isomers there must be three parent radicals.This result is in agreement with the assignment of radicals by pulse radiolysis. Absorption spectra obtained after a pulse of electrons (Febetron) were measured between 50 ns and 20 ,us6 for aqueous solutions of NMA. From the pulse radiolysis the distribution of radicals is the following: parent radicals I 40%, I1 lo%, I11 50%. If the combination of radicals submits to a statistical rule, this proportion of radical I is undervalued according to the relative yields of the dimer isomers. Pulse measurements during a broader interval of time (+ ms) could reveal slow radical mechanisms. In the radiolysis of neat liquid compounds, the selectivity in radical production may be excluded by the high energy of molecular excitation (15-25 eV). Moreover, radicals are produced in ' spurs ', and radical-radical combination reactions are favoured by the very high concentration of radicals in the spurs.Changes of radical identity: CH,CON'CH, + CH,CONHCH, ---* CH,CONHCH, + CH,CONHCH; (8) 'CH,CONHCH, + CH,CONHCH, -+ CH,CONHCH, + CH,CONHCH; (9) are in competition with radical recombination only after radical diffusion in the bulk. At point some of the parent radicals have already recombined, and at least 20% of the dimer isomers yield are formed in the spurs. The quality of dimer isomers does not change with the dilution of DMF in water before irradiation ; the distribution varies slightly. Moreover, the cross-dimer is more abundant (13.6%) in dilute solution than in concentrated solution (9.8%). There is no spur effect for reactive species from DMF in concentrated solution.Conclusions E.s.r. measurements suggest a specificity in the production of radicals in solid radiolysis. G.c. results complete the e.s.r. measurements by allowing the identification of radicals produced in relative low concentration ; final product analysis indicates a selectivity inA-C. Dusaucy, J . De Doncker, C. Couillard, M. De Laet and B. Tilquin 133 the formation of the parent radicals. Their distribution depends, for one thing, on the nature of the amide. In order to study radiation damage in the liquid phase, pulse radiolysis and product analysis are complementary methods. The former gives information on reaction inter- mediates, while the latter takes into account the overall damage, and, as in this study, completes the pulse-radio1 ysis results.The assistance of Dr A. M. Koulkes-Pujo in preliminary experiments is gratefully acknowledged. Financial support was obtained from the F.N.R.S. which is gratefully acknowledged. References 1 G. V. Pukhal’skaya, A. G. Kotov and S. Ya. Pshezhetskii, Khim. Vys. Energ., 1969, 3, 340. 2 G. V. Pukhal’skaya, A. G. Kotov and S. Ya. Pshezhetskii, BioJizika, 1972, 17, 756. 3 (a) B. Tilquin, Th. Baudson, P. Claes, A. Lund and 0. Claesson, J. Phys. Chem., 1982, 86, 3324. (b) B. Tilquin, C. Gourdin-Serveniere, T. Miyazaki and K. Fueki, Bull, Chem. SOC. Jpn, 1984,57,2029. 4 B. Tilquin, B. Massaut and P. Claes, Radiat. Phys. Chem., 1982, 19, 283. 5 A. J. Swallow, in The Study of Fast Processes and Transient Species by Electron Pulse Radiolysis, ed. 6 A. M. Koulkes-Pujo, L. Gilles, J. C. Halle, T. H. Tran-Thi and J. Sutton, 2nd Int. Symp. Organic Free 7 E. Hayon, I. Ibata, N. N. Lichtin and N. Simic, J. Am. Chem. Soc., 1970,92, 3898. 8 E. Pucci, J. Vedel and B. Tremillon, J. Electroanal. Chem., 1969, 22, 253. 9 R. Livingstone and H. Zeldes, J. Phys. Chem., 1967,47, 4173. J. H. Baxendale and F. Busi (D. Reidel, Dordrecht, 1982). Radicals (CNRS, Aix-en-Provence, 1977). 10 T. Yonezawa, I. Noda and T. Kawamura, Bull. Chem. SOC. Jpn, 1968,41, 766. 11 S. Rustgi and P. Riesz, Int. J. Radiat. Biol., 1978, 33, 325. 12 B. Tilquin, Dissertation AES (University of Louvain, 1985), ed. B. Tilquin (Bibliotheque des Sciences, 13 B. Tilquin, P. Tilman and P. Claes, Radiat. Phys. Chem., 1980, 16, 321. 14 S. Ya. Pshezhetskii, A. G. Kotov, V. K. Milinchuk, V. A. Roginskii and V. I. Tupikov, EPR of Free 15 A. M. Koulkes-Pujo, L. Gilles, J. C. Halle and J. Sutton, Radiat. Phys. Chem., 1977, 10, 73 and Louvain-la-Neuve, Belgium). Radicals in Radiation Chemistry (J. Wiley, New York, 1974). references cited. Paper 6/ 1142; Received 6th June, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300125

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1987,83, 135-139 An Electron Spin Resonance Study of the Radicals formed on Ultraviolet Irradiation of the Photoallergens Fentichlor and Bithionol Jonathan N. Delahanty, Jeffrey C. Evans* and Christopher C. Rowlands Chemistry Department, University College Card& P.O. Box 78, Card13 CFl 1 XL Martin D. Barratt Environmental Safety Laboratory, Unilever Research, Colworth House, Sharnbrook, Bedford Radicals formed on U.V. irradiation of aqueous solutions (pH 8) of fentichlor and bithionol have been detected by e.s.r. spectroscopy. The radicals have been identified as substituted benzosemiquinone radicals and the proton hyperfine coupling constants have been assigned on the basis of methyl substitutions. One of us has shown previously1 that the fungicide and germicide fentichlor (1) [bis(2-hydroxy-5-chlorophenyl)sulphide] together with the related compound bithionol (2) [bis(2-hydroxy-3,5-dichlorophenyl)sulphide] under the influence of u.v.light can bind C l C l C l Cl (1 1 (2) to soluble proteins, it being suggested that such interactions occur as a result of the loss of one or more chlorine atoms. If such a loss did occur then free radicals might be formed, and hence the use of e.s.r. spectroscopy would be a valuable tool in elucidating their structure. In this study we have examined the radicals formed on irradiating alkaline solutions of both fentichlor and bithionol and propose a mechanism for their formation. Experiment a1 Materials 4-Chlorophenol, Aldrich Chemical Co., was used as supplied.Fentichlor was synthesised from 4-chlorophenol by the method of Dunning2 and recrystallised from toluene, m.p. 174 "C; lit. 173 oC.2 The commercially available bithionol (Sigma) was recrystallised from p-xylene. Procedure Aqueous samples were prepared from stock solutions of fentichlor and bithionol in methanol (0.01 g in 3 cm3) and adding 3 drops of this stock solution to distilled water 135136 E.S.R. Study of U. V.-irradiated Photoallergens (a) (b) ( C) 1 G U Fig. 1. E.s.r. spectra obtained on irradiating alkaline solutions of fentichlor (PH 8.0). (a) Initial spectrum, (b) after further irradiation, (c) after prolonged irradiation. (3 cm3). The solutions were made alkaline with 0.1 mol dm-3 sodium hydroxide and the pH adjusted as desired. E.s.r. measurements were made with a Varian E3 spectrometer.Sample irradiation was carried out using either a Bausch and Lomb 150 W xenon lamp or a medium-pressure Hg lamp in the spectrometer cavity. No differences were observed in the radicals produced by either method. Results and Discussion Fen tic hlor Irradiation of fentichlor in alkaline solution (pH 8) gives rise initially to an eight-line spectrum [fig. 1 (a)] (primary radical). Further irradiation showed the appearance of a new species (secondary radical). Fig. l(b) shows the e.s.r. spectrum of both radicals. Prolonged irradiation gives the complex spectrum shown in fig. 1 (c) (primary, secondary and tertiary radicals). Identification of the Radicals Primary Radical The initial eight-line spectrum can be analysed in terms of a triplet of doublets arising from three inequivalent protons with the following hyperfine coupling constants (h.f.c.) : aH = 2.9, 2.0 and 1.25 G.The h.f.c. were assigned to their respective positions by reference to the result obtained from irradiating the 3,3’-dimethyl- and 4,4’-dimethyl- substituted fentichlor analogue~.~ Grabonski4 has shown that photolysis of halogenated phenols in aqueous alkaline solutions results in the replacement of a halogen with an OH group to yield the corresponding dihydroxybenzenes. Continued irradiation would then in the presence of oxygen lead to the formation of the benzosemiquinone radical. This we believe to be the case for the formation of the fentichlor primary radical (scheme 1). Indeed, irradiation under anaerobic conditions did not lead to the formation of radicals.The triplet of doublets obtained suggests that there is no delocalisation of the electron throughout the molecule. This is probably due to the tetrahedral character of the sulphur atom in the bridge.J. N. Delahanty, J. C. Evans, C. C. Rowlands and M. D. Barratt 137 OH OH CL CI 0 OH @ST$ 0 Cl /OH- 0- OH H QS$ 0 CI HO &s+ C l 0 1’ radical ihV 0- 0 0 3’ radical .1 OH OH 0- OH OH Cl 0 Cl 2 O radical Scheme 1. Secondary Radical The spectrum [fig. l(b)] obtained on further irradiation showed the presence of new features which can be shown to be a 1 : 2: 1 triplet (a, = 0.90 G) by subtraction of the eight-line spectrum of the primary radical. The mechanism for the formation of this radical species is given in scheme 1 and is in agreement with the work of Peder~en.~ Evidence to support this mechanism is found in a study of the 4,4’-dimethyl-~ubstituted analogue., In this case no secondary radical is obtained, as OH- attack at this position is now blocked by the CH, group ; however, the 3,3’-dimethyl-~ubstituted fentichlor does give a secondary radical.Tertiary Radical On prolonged irradiation the spectrum shown in fig. 1 (c) is obtained. In addition to the primary and secondary radical spectra we now have a spectrum which can be interpreted138 E.S.R. Study of U. V.-irradiated Photoallergens Table 1. Assignment of hyperfine coupling cons tan ts species 3 4 6 0- OH 2.0 2.9 1.25 0- OH 0.9 HO 0 4H = 2.40 as a 1 : 4 : 6 : 4 : 1 quintet with a coupling of aH = 2.4 G.This radical is formed by cleavage of the C-S bond followed by hydrogen abstraction from the water, which on further oxidation gives the benzosemiquinone radical. The spectrum obtained is identical to that produced by the oxidation of hydroquinone to the benzosemiquinone radical (a, = 2.4 G). If the photolysis is carried out in deuterium oxide and NaOD then the spectrum observed is that from three equivalent protons and a deuteron. The hyperfine coupling constants are given in table 1. Bithionol When bithionol was irradiated under the same conditions as fentichlor, the initial e.s.r. spectrum consisted of 18 lines [fig. 2(a)]. On further irradiation the spectrum shown in fig. 2(b) was obtained. This is very similar to that obtained from the fentichlor system [cf.fig. 1 (b)]. Continued irradiation now leads to the same sequence as seen previously for fentichlor. We cannot interpret the initial spectrum [fig. 2(a)]. However, since the e.s.r. spectrum changes with time to that of fig. 2(b) the molecule must rearrange losing the chlorine (position 5 ) as before and also exchanging the chlorine (position 3) for a proton from the solvent. Thus the system has become the same as the fentichlor radical system (the eight-line spectrum from the primary species superimposed on which is a triplet from the secondary radical). If experiments are carried out in deuterium oxide and NaOD the initial 18-line spectrum does not change to the fentichlor primary and secondary radical system thus supporting the fact that a proton is abstracted from the solvent .J.N . Delahanty, J. C. Evans, C. C . Rowlands and M. D. Barratt 139 ( a ) 1.25 G - Fig. 2. E.s.r. spectra obtained after irradiation of alkaline solutions of Bithional (pH 8.0). (a) Initial spectrum, (b) after further irradiation. Conclusion These results show that irradiation of fentichlor (pH 8) leads to the formation of at least three radicals, the primary being a benzosemiquinone radical which undergoes further oxidation and nucleophilic attack by OH- to give rise to the secondary radical which is a 4-hydroxybenzosemiquinone-type system. To obtain the 1 : 2 : 1 triplet the two remaining protons must be nearly equivalent, even though they are in different environments. The tertiary radical is obtained by homolytic cleavage of the C-S bond and subsequent hydrogen abstraction from the solvent. The bithionol system is at first much more complex, producing initially a radical not yet understood but eventually following the fentichlor mechanism. One of us J.N.D. thanks the S.E.R.C. for a CASE studentship with Unilever. References 1 D. Rickwood and M. D. Barratt, Chem. Biol. Interact., 1984, 52, 213. 2 F. Dunning, B. Dunning and E. W. Drake, J. Am. Chem. SOC., 1931,53, 3466. 3 J. N. Delahanty, J. C. Evans, C. C. Rowlands and M. D. Barratt, unpublished work. 4 Z . R. Grabonski, J. Phys. Chem., 1961,27, 239. 5 J. A. Pedersen, J . Chem. SOC., Perkin Trans. 2, 1973, 424. Paper 6/1001; Received 22nd May, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300135

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1 , 1987, 83, 141-149 Generation of Radicals from Antioxidant-type Molecules by Polyunsaturated Lipids Pierre Lambelet, Francine Ducret, Fransoise Saucy, Marie-Claude Savoy and Jurg Loliger Nestec Ltd., Nestlk Research Department, Avenue Nestlk 55, CH-1800 Vevey, Switzerland A method for the formation of radicals from antioxidant-type compounds which makes use of the radicals generated during the autoxidation of polyunsaturated lipids is described. This simple system allows observation of radicals by e.s.r. spectroscopy as well as the study of radical exchange reactions in conditions similar to those encountered during lipid autoxida- tion and biological peroxidation reactions. Advantages and limitations are discussed. Applications presented include studies of the fate of antioxidants and the role of oxygen during lipid autoxidation, investigations of radical exchange reactions occurring in living cells, and of the role of oxidative reactions in the antipsoriatic drug action of anthralin.A growing interest in the understanding of radical reactions in the living cell is apparent in the recent 1iterature.l Within the range of this research we address here the function and mechanism of lipid a-utoxidation in relation to synthetic and biological radical scavengers. A wide variety of different reactions are currently used for generating radicals from biological radical scavengers to elucidate spectroscopically their structures and reactivities. These include one-electron oxidation of the parent molecule by various chemical agents;2-6 chemical' or photo-8 reduction of a precursor; reaction with peroxy radicals generated in situ by thermalg or U.V.irradiation-induced1° decomposition of a peroxide or by pulsed U.V. irradiation decomposition of a ketone;ll U.V. irradiation;12 pulse radiolysis;13 electron transfer with another free radical ;14 enzymatic oxidation;15 electrochemical oxidation or reduction. l6 Formation of radicals by reacting with polyunsaturated lipids has already been reported. Karel et a1.l'~ l8 studied free-radical transfer from oxidised methyl linoleate to amino acids and proteins in dry model systems. Yamamoto et aI.l9 studied the oxidation of oils and fats by the assay of radicals resulting from the reaction of antioxidants with lipids. Bascetta et a1.20 used a similar system for studying the inhibition of fatty acid oxidation by Vitamins E and C in a model membrane.We intend to show here that radical reactions of autoxidising lipids can be studied in a complex biological matrix under conditions which create fewer artefacts than those already described.lg9 2o The principles and scope of this method as well as its advantages and limitations are outlined here. Materials and Methods Materials Chicken fat liquid fraction (CFLF) was prepared from deodourized chicken fat by dry fractionation at 23 "C using a procedure described elsewhere.21 All other. compounds listed in the present study are of commercial origin except for ascorbyl palmitate, which 141142 Formation of Radicals from Antioxidant Molecules by Lipids was kindly supplied by Hoffmann La Roche (Basel, Switzerland), anthralin, which was a gift from B.Shroot (Centre International de Recherches Dermatologiques, Sophia Antipolis, France), 2,6-di-t-butylhydroquinone7 which was prepared by reduction of the corresponding quinone,22 and methyl-3,4,5-0-trideuterogallate, which was prepared by H-D exchange in BunOD. They were used without further purification. Methods Peroxide values were determined by a colorimetric method using the iron(II)/iron(m) thiocyanate complex according to Lips et al.23 Reaction leading to the Formation of Radicals The CFLF was oxidized by heating, at up to 150 "C for 3 h. After cooling, the chemical species from which a radical was to be formed was added at a concentration of ca.0.1-1 % (w/w). The mixture was allowed to react at room temperature, and spon- taneously formed radicals were studied by e.s.r. spectroscopy. For this purpose, aliquots of the reaction mixture were periodically transferred into a quartz tube. E.s.r. spectra were run on a Varian E-109 Century series mark I11 spectrometer (X band) using 100 kHz magnetic-field modulation. The microwave frequency was determined using a Hewlett Packard 5342A frequency counter and the magnetic field intensity with a Varian E-500-2 n.m.r. gaussmeter calibrated with perylene cation radical in H,SO, as g-factor ~eference.~, Relative radical concentrations were determined by double integration of the e.s.r. spectra using a Varian E-935-3 data acquisition system. Results Many antioxidant-type molecules have been oxidized using peroxidizing lipids : substi- tuted phenols or hydroquinones, gallates, tocopherols, Vitamin C and quinolines.Fig. 1 shows an example of a strong, stable e.s.r. spectrum observed on reacting 1% 2,6-di-t-butyl-4-methylphenol (BHT) with peroxidizing CFLF, which could be analysed in terms of a quartet of triplets. The e.s.r. parameters associated with this spectrum are given in table 1. The less substituted 2,6-di-t-butylpheno17 on mixing with peroxidizing CFLF, gave at the beginning of the reaction with CFLF (within one day) only a weak doublet of triplets as seen in fig. 2(a). Further, the e.s.r. spectrum changed with time to give an eight-line e.s.r. spectrum about one week later [fig. 2(b)]. Other diversely substituted phenols and hydroquinones gave stable e.s.r. spectra by reaction with peroxidizing lipid, as summarized in table 1.1 % of gallate esters (methyl-, n-propyl-, n-octyl-, n-dodecyl-) dissolved in peroxidizing CFLF gave rise to single broad-line e.s.r. spectra [fig. 3(a)]. The hyperfine structure of these spectra could be resolved by deuterating the hydroxy groups, as illustrated by the e.s.r. spectrum from methyl 3,4,5-0-trideuterogallate [fig. 3 (b)] which was composed of a triplet with a proton hyperfine coupling constant equalled to 2.0 G. Reacting 1% of a tocopherol isomer (dl-a-, dl-B-, dl-y- or dl-6-) with peroxidizing CFLF led to the observation of e.s.r. spectra already discussed elsewhere, which could be unambiguously attributed to the corresponding tocopheroxy radicals.25 Reaction of ascorbic acid (0.5% dissolved in 25% n-butanol as cosolvent) and its liposoluble derivative, ascorbylpalmitate (1 % ) showed in each case two e.s.r. lines with proton hyperfine coupling constants equal to 2.0 and 1.8 G for the acid and the ester, respectively, corresponding to the well known ascorbyl Oxidation of quinolines in CFLF led to the observation of strong e.s.r. signals (table 2). For instance, reacting 1 % 1,2-dihydr0-2,2,4-trimethyl-6-ethoxyquinoline (Ethoxy- quin) gave rise to a triplet (1 : 1 : 1 relative intensities) of doublets (fig. 4).P . Lambelet, F. Ducret, F. Saucy, M-C. Savoy and J. Loliger 143 Fig. 1. E.s.r. spectrum recorded on reacting 2,6-di-t-butyl-4-methylphenol (1 %) with peroxidizing CFLF. Table 1.E.s.r. parameters of radicals formed by reaction of various substituted phenols (1 %) with peroxidizing CFLF OH hyperfine coupling constants/G - t - but yl t-butyl H 2.0 9.8 2.0045a 1.7 11.2 - 2.0045 - 2.0044 t-butyl t-butyl CH20H 1.7 12.4 t- butyl t-butyl OH 0.6 1.4 - 2.0045 t-butyl t -bu t yl OCH, 0.9 1.6 - 2.0047 t-butyl H OH < 1 1.9 5.5 2.0047 t- butyl t-butyl CH3 a Radical not stable in the present conditions. Discussion Choice of Lipid ,4 number of lipids were tested as substrates. Pure methyl esters of various polyunsat- urated fatty acids such as arachidonic and linolenic acid were used, as were triolein and natural lipids including soya oil or CFLF. CFLF was chosen as the lipid substrate for the following reasons: it permits the recording of e.s.r.spectra at room temperature, it is highly unsaturated and therefore highly reactive towards oxygen and, if it contains tocopherols at all, these are below the detection level.1 44 F 1 (b? 1 Formation of Radicals from Antioxidant Molecules by Lipids I H - 4 G - I 2.5 G (b? H 2.5 G _t I Fig. 3. E.s.r. spectra recorded on reacting (a) methyl gallate (1 %) and (b) methyl-3,4,5-O-trideuterogallate (1 % ) with peroxidizing CFLF. Table 2. E.s.r. parameters of radicals formed by reaction of quinolines (1%) with peroxidizing CFLF hyperfine coupling constants/G quinoline aN aH g-factor 1 ,2-dihydro-2,2,4-trimethyl-6-ethoxyquinoline 10.9 3.8 2.0053 1,2,3,4-tetrahydro-2,2,4-trimethyl-6-ethoxyquinoline 11.5 3.3 2.0052P . Lambelet, F. Ducret, F. Saucy, M-C. Savoy and J.Loliger 145 Fig. 4. E.s.r. spectrum recorded on reacting 1,2-dihydro-2,2,4-tetrarnethyl-6-ethoxyquinoline (ethoxyquin) (1 % ) with peroxidizing CFLF. Type of Radicals Observed A number of different radical types are present in an oxidizing lipid, namely hydroxy, alkoxy and peroxy radicals of fatty-acid chains (various different isomers are formed according to the structure of the fatty acid). E.s.r. detection of these radicals has already been reported at low temperat~re,~’~ 28 at room temperature, but only for a very short time after formation,29* 30 or using the spin-trap 32 However, these radicals are very unstable and can only be detected under the special conditions ~ ~ e d . ~ ~ - ~ Recording e.s.r. spectra of peroxidized polyunsaturated lipids in the conditions used for this study (room temperature) showed that no radical species could be detected.However, after adding a radical scavenger different signals could be seen. It can be speculated that under our conditions, the concentration of the lipid radicals (hydroxy-, peroxy and alkoxy radicals) was too low for e.s.r. detection, but that adding a radical scavenger led to detectable e.s.r. signals owing to the oxidation of the scavenger by these lipid radicals. Antioxidant radicals generated by reaction with autoxidizing lipids have similar e.s.r. parameters as the corresponding radicals obtained by other commonly used oxidation techniques, for example by U.V. irradiation in the presence of t-butylperoxide in an aprotic solvent. Therefore species observed during the reaction of antioxidant type molecules with autoxidizing lipids are certainly neutral species formed by the abstraction of a phenolic (phenoxy radical) or hydroxylic (alkoxy radical) hydrogen atom.Exceptions could be the radicals derived from Vitamin C or its derivative, ascorbyl palmitate, which might be anions.26 Likewise the species derived from the quinolines are probably nitroxide radicals resulting from the abstraction of the aminic hydrogen atom and the subsequent reaction of this radical with oxygen.6 In peroxidizing lipids the antioxidant radicals normally decomposed to form dia- magnetic products only. In some rare instances, however, the decomposition of these radicals results in the formation of secondary radicals which can also be detected by e.s.r.spectroscopy. In this way the e.s.r. signals observed at the beginning of the reaction of the 2,6-di-t-butylphenol with lipid [fig. 2 (a)] can be assigned33 to the 2,6-di-t-butylphenoxy radical (primary species). The subsequent e.s.r. signals recorded during this reaction are to be attributed to a mixture of the primary radical and a secondary-species which is most probably derived from the 3,3’,5,5’-tetra-t-b~ty1-4,4’-biphenyldiol.~~ Influence of Experimental Parameters on the Formation of Radicals from Antioxidants It is generally agreed that in a system such as the one considered here (mixture of lipid and antioxidant type molecules) the radicals from antioxidants are formed by reaction of these antioxidants with the peroxy radicals or with other radicals produced by decomposition of the peroxy radicals present in the lipid.Peroxy or other related radicals are not easily determined, although peroxides are. We therefore tried to relate relative146 Formation of Radicals from Antioxidant Molecules by Lipids Y Fig. 5. Comparison of e.s.r. spectra of dl-a-tocopheroxy radicals generated from the parent tocopherol by (a) oxidation with PbO, in toluene under vacuum (taken with permission from K. Mukai et al., Chern. Phys. Lipids, 1981, 29, 129) and (b) reaction with peroxidizing CFLF. 0 50 100 150 200 2 50 fat oxidation level/mequiv 02kg-' Fig. 6. Relative radicals concentration recorded on reacting 0.5% (w/w) BHT (A) or dl-a-tocopherol (e) with CFLF at different levels of oxidation. concentrations of radicals derived from the antioxidant with the amount of peroxides initially present in the lipid.For this purpose we chose two antioxidants, one known for its relatively high reactivity with peroxy radicals (dl-a-tocopherol) and the other (BHT) known for its relatively low reactivity with peroxy radicals.35 These two antioxidants were mixed at 0.5 % w/w with CFLF oxidized at different peroxide values ranging from ca. 3 to 250 mequiv 0, kg-l. The relative radical concentrations were measured ca. 20 min after the start of the reaction, and as seen in fig. 6, the amount of radicals formed with dl-a-tocopherol is much higher than with BHT. This reflects the difference in reactivity towards peroxy radicals of the two antioxidants. Moreover, while the relative concentration of dl-a-tocopheroxy radicals is independent of the peroxide value of the lipid, the relative concentration of 2,6-di-t-butyl-4-methylphenoxy radicals increasesP.Lambelet, F. Ducret, F. Saucy, M-C. Savoy and J . Loliger 147 Table 3. E.s.r. parameters of the dl-a-tocopheroxy radical hyperfine coupling constants/G oxidizing agents ads-CH3) ad7-CH3) a,(8-CH3) aH(4-CH,) g-factor ref. ~~ ~ superoxide ion 5.88 4.78 0.74 0.74 2.0046 5 4 lead dioxide 5.98 4.57 0.94 1.47 2,2.-diphenyl-l- 6.07 4.55 0.98 1.52 2.0046 14 4.46 0.84 1.47 2.0046 - - picrylhydrazyl oxidizing CFLF 5.77 slightly with the peroxide value of the fat. No direct correlation can therefore be found between the peroxide value of a fat and the amount of radicals generated by this fat during the reaction with an antioxidant-type molecule.However, sufficiently high concentrations of radicals for e.s.r. detection are formed in the presence of a lipid with concentrations as low as 3 mequiv 0, kg-l and probably even lower. The influence of oxygen on the formation of radicals from antioxidants during reaction with oxidizing lipid was on the other hand investigated. Radical reactions between dl-a-tocopherol and peroxidizing lipid (CFLF or ethyl linolenate) have been studied under controlled oxygen supply. Reactions were carried out both in a sealed and in an open e.s.r. quartz tube. As reported the concentration of dl-a-tocopheroxy radicals rapidly decreased to below their detection level as long as no oxygen was supplied to the reaction mixture. As soon as oxygen was supplied again, the radical concentration rapidly increased and fell once more when the oxygen supply was stopped by resealing the tube.In comparison, when the same reaction was carried out under constant oxygen supply, only a slow decrease in the radicals concentration as a function of time was observed. These results clearly show that the amount of oxygen available for the lipid oxidation is a mediator for this reaction and will influence the amount of radicals formed from antioxidant type molecules. Potential and Limitations of Generating Radicals by Peroxidizing Lipids The driving force of this study was the possibility of investigating radical reactions in lipids under a great variety of different conditions. Many of the vital biological processes are related to radical exchange reactions which are supposed to take place in the lipid bilayer of membranes.The fact of being able to study radical exchange reactions between polyunsaturated oxidizing lipids and antioxidant-type molecules opens up a wide field of applications on the implications of radical scavengers in, for example, cancer induction and 38 This approach to radical reactions allows their observation in their ‘lipid environment ’ with less perturbation than that encountered with most radical generating schemes usually employed for this type of investigation. This ‘ in situ’ generation of radicals in autoxidizing lipids compares very favourably with most of the known radical generation systems used for studying antioxidant radicals. As an example, fig.5(a) shows the e.s.r. spectrum of dl-a-tocopheroxy radical generated by lead dioxide oxidation in toluene under va~uum,~ and fig. 5(b) shows the corresponding spectrum in a peroxidizing lipid. There is no doubt as to the excellent resolution which could be obtained in the degassed toluene solution but the hyperfine coupling constants and g-values in table 3 show that all essential data for the characterisation of the e.s.r. spectrum of the radical from dl-a-tocopherol can be extracted from the lower-resolution spectrum in fig. 5(b). It could be shown that most of the e.s.r. spectra of antioxidant-type molecules obtained in peroxidizing lipids were148 Formation of Radicals from Antioxidant Molecules by Lipids unambiguously attributable to the already known39 spectra determined under optimal e.s.r.conditions. This in situ formation of the radicals of antioxidants does not of course allow the precise determination of rate constants for hydrogen-atom transfer reactions, so well performed by the azo-bis(isobutyronitri1e) initiator systems often used by Burton and Ing01d.~~ However, by comparing relative radical concentrations formed for different antioxidants under identical conditions, their relative reactivities could still be determined. 25 Conclusions The possibility of directly observing radical reactions in various lipid containing systems has already found applications in as different fields as food chemistry, dermatology and biochemistry. The different types of behaviour observed for 2,6-di-t-butylphenol and other hindered phenols during their reaction with lipids thus provide useful information on the fate of antioxidants in a lipid medium.It should therefore be of great benefit to relate these results with findings on the biological activities of these chemicals, for example with respect to cancer induction or inhibition and 38 It has been shown on the other hand25 that the four tocopherol isomers in the course of their reaction with lipids form different concentrations of tocopheroxy radicals : dZ-a-tocopheroxy radicals were formed in relatively high concentration but were rather rapidly destroyed; dl-6-tocopheroxy radicals were formed in rather low concentrations but were observable for a longer period of time; dZ$- and dl-y-tocopherols behaved in an intermediate manner. These observations by e.s.r.spectroscopy can well be related to the antioxidant properties of the four tocopherol isomers. Investigations of radical reactions of anthralin in peroxidizing lipids represent an approach for studying the fate in viuo of this antipsoriatic drug after topical application on skin. Previously described indicated the formation of the 1,8-dihydroxy- 9-anthron-10-yl radical only when formed by U.V. irradiation of an anthralin solution. However, reacting this drug with oxidizing lipids leads to the observation of a succession of radical species.41 At the beginning of the reaction the 1,8-dihydroxy-9-anthron- 10-yl is formed. As the reaction proceeds further, secondary radicals are observed. Although the detailed structures of these species are still unknown, it could be shown that they derive from the dimer of anthralin. These results contribute to the understanding of the mechanisms of antipsoriatic action of anthralin as well as of its major side effect, skin irritation. Indeed they allow us to say that not only the 1,8-dihydroxy-9-anthron-lO-y1 radical but also the secondary species can be resppnsible for the two physiological effects of anthralin, both of which are believed to be of a radical A number of authors have discussed mechanisms of Vitamin E and Vitamin C interactions and their importance as chain-breaking antioxidants in the in uiuo autoxi- dation of polyunsaturated lipids of cellular membranes.T a ~ p e 1 ~ ~ first proposed a regeneration scheme whereby the primary radical formed from Vitamin E can be intercepted by Vitamin C, which itself becomes oxidized by a hydrogen-atom transfer reaction.By generating radicals using pulse radiolysis in an heterogenous system, Packer et al.44 showed that a hydrogen-atom transfer reaction takes place between ascorbic acid and tocopherol radicals. Using the generation of radicals by peroxidizing lipids we have observed radical exchange reactions between Vitamin E and ascorbyl palmitate in peroxidizing lipids.45 These radical exchange reactions between Vitamin E and Vitamin C could also be observed in an heterogenous water/lipid system.46 The results so obtained give additional support to the ' regeneration scheme' of Tappel already mentioned. We are grateful to M. Huynh-Ba for preparing the 2,6-di-t-butyl-4-hydroxyphenol, Ian Horman for reviewing the manuscript and Cathy Isom for typing it.P .Lambelet, F. Ducret, F. Saucy, M-C. Savoy and J. Loliger 149 References 1 B. Halliwell and J. M. C. Gutteridge, Free Radicals in Biology and Medicine (Clarendon Press, Oxford, 2 W. T. Dixon and D. Murphy, J. Chem. Soc., Faraday Trans. 2, 1976,72, 1221. 3 J. C. Westfahl, Rubber Chem. Technol., 1973,46, 1134. 4 K. Mukai, N. Tsuzuki, K. Ishizu, S. Ouchi and K. Fukuzawa, Chem. Phys. Lipids, 1981, 29, 129. 5 T. Ozawa, A. Hanaki, S. Matsumoto and M. Matsuo, Biochim. Biophys. Acta, 1978, 531, 72. 6 J. S. Lin and H. S. Olcott, J. Agric. Food Chem., 1975, 23, 798. 7 M. R. Das, H. D. Connor, D. S. Leniart and J. H. Freed, J. Am. Chem. Soc., 1970, 92, 2258. 8 T. Foster, A.J. Elliot, B. B. Adeleke and J. K. S . Wan, Can. J. Chem., 1978, 56, 869. 9 J. C. Westfahl, C. J. Carman and R. W. Layer, Rubber Chem. Technol., 1972, 45,402. 1985). 10 K. Loth and F. Graf, Helv. Chim. Acta, 1981, 64, 1910. 1 1 J. A. Howard, Rev. Chem. intermed., 1984, 5, 1 . 12 K. Loth, F. Graf and H. H. Gunthard, Chem. Phys., 1976, 13, 95. 13 B. H. J. Bielski and A. 0. Allen, J. Am. Chem. SOC., 1970, 92, 3793. 14 M. Matsuo and S. Matsumoto, Lipids, 1983, 18, 81. 15 I. Yamazaki, H. S. Mason and L. Piette, J. Biol. Chem., 1960, 235, 2444. 16 J. Oakes and M. C. R. Symons, Trans. Faraday Soc., 1968,64, 2579. 17 M. Karel, K. Schaich and R. B. Roy, J. Agric. Food Chem., 1975, 23, 159. 18 K. M. Schaich and M. Karel, Lipids, 1976, 11, 392. 19 H. Hasegawa, Y. Yamamoto and A.Iizuka, Japan Patent application 106,665/79 August 1979. 20 E. Bascetta, F. D. Gunstone and J. C. Walton, Chem. Phys. Lipids, 1983, 33, 207. 21 A. Dieffenbacher, Swiss Patent application 5467-83, October 1983. 22 C. J. Worrel and R. L. McLean, US patent 3,415,850, 10 December 1968. 23 A. Lips, R. A. Chapman and W. P. McFarlane, Oil and Soap, 1943,20, 240. 24 J. E. Wertz and J. R. Bolton, Electron Spin Resonance. Theoretical and Applications (McGraw-Hill, 25 P. Lambelet and J. Loliger, Chem. Phys. Lipids, 1984, 35, 185. 26 G. P. Laroff, R. W. Fessenden and R. H. Schuler, J. Am. Chem. SOC., 1972,94,9062. 27 M. Haydar and D. Hadziyev, J. Am. Oil Chem. Soc., 1973,50, 171. 28 C. U. Deffner, H. Luck and R. Kohn, Z. Lebensm. Unters. Forschung, 1964, 125, 281. 29 E.Bascetta, F. D. Gunstone and J. C. Walton, J. Chem. Soc., Perkin Trans. 2, 1983, 603. 30 E. Bascetta, F. D. Gunstone and J. C. Walton, J. Chem. SOC., Perkin Trans. 2, 1984, 401. 31 T. Yamada, E. Niki, S. Yokoi, J. Tsuchiya, Y. Yamamoto and Y. Kamiya, Chem. Phys. Lipids, 1984, 32 J. C. Evans, K. R. N. Rao, S. K. Jackson, C. C. Rowlands and M. D. Barratt, J. High Resolution 33 J. Pannell, Chem. ind., 1962, 1797. 34 P. Pelikan, A. Tkac, L. Omelka and A. Stasko, Org. Magn. Reson., 1982, 20, 205. 35 G. W. Burton, T. Doba, E. J. Gabe, L. Hughes. F. L. Lee, L. Prasad and K. U. Ingold, J. Am. Chem. 36 J. Loliger and P. Lambelet, work in preparation. 37 D. C. McBrien and T. F. Slater, Free radicals, Lipid Peroxidation and Cancer (Academic Press, London, 1982). 38 H. J. Armbrecht, J. M. Prendergast and R. M. Coe, Nutritional intervention in the Aging Process (Springer, New York, 1984). 39 W. Uber and €3. B. Stegmann, in Landolt-Bornstein: Numerical Data and Functional Relationships in Science and Technology, ed. H. Fischer and K-H. Hellwege (Springer, New York, 1979), group 11, vol. 9, part C2, p. 29. New York, 1972), p. 33. 36, 189. Chromatogr. and Chromatogr. Commun., 1985, 8, 829. Soc., 1985, 107, 7053. 40 A. G. Davies, J. A. A. Hawari and M. Whitefield, Tetrahedron Lett, 1983, 24, 4465. 41 F. Ducret, P. Lambelet, J. Loliger and M-C. Savoy, J. Free Radicals in Biology and Medicine, 1985, 1, 42 R. E. Ashton, P. Andre, N. J. Lowe and M. Whitefield, J. Am. Acad. Dermatol., 1983, 9, 173. 43 A. L. Tappel, Geriatrics, 1968, 23, 97. 44 J. E. Packer, T. F. Slater and R. L. Willson, Nature (London), 1979, 278, 737. 45 P. Lambelet, F. Saucy and J. Loliger, Experientia, 1985, 41, 1384. 46 J. Loliger, P. Lambelet, M-C. Savoy and F. Saucy, Seventeenth International Symposium on 301. Radicals, Granby, Colorado, August 18-23 (1985). Paper 6/1005; Received 22nd May, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300141

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. I, 1987, 83, 151-159 Biomolecular Dynamics and Electron Spin Resonance Spectra of Copper Complexes of Antitumour Agents in Solution R. Basosi," L. Trabalzini and R. Pogni Department of Chemistry, University of Siena, Pian dei Mantellini, 44, 53100 Siena, Italy W. E. Antholine National Biomedical ESR Center, Milwaukee, Wisconsin 53226, U S . A . For the purpose of developing new antitumour agents which are more efficacious and have less generalized toxicity than existing ones, the free- radical generation and metal complexation of well known anticancer agents have been studied. Copper(~r) ion complexes are readily formed with several members of a class of hydroxyurea derivatives which are known to be effective ribonucleotide reductase inhibitors. E.s.r.measurements and u.v.- visible titration illustrate weak binding for 3,4-dihydroxybenzohydroxamic acid and tight binding in complex formation for gallohydroxamic acid and 2,3,4-trihydroxybenzohydroxamic acid. These data were used in a prelimi- nary investigation of cytotoxicity, and the results are consistent with single phase cell cycle killing. Until recently, the majority of chemotherapeutic agents has been designed by organic or natural-products chemists as toxic precursors or antimetabolites of specific cellular reactions. Currently, studies of metal-bound antitumour agents,'* as well as the existence of several other such low-molecular-weight suggest a more prominent role for the inorganic chemist. The first such complex to be successfully tested in humans was cis-dichlorodiamine platinum(~r),' in which renewed interest has been shown recently for various gynaecological cancers previously considered untreatable. Bleomycins-l0 and one class of thiosemicarba~ones~~ have been shown to be metal-bound in their active form.Thus, metal reactions within cells seems to be a well established phenomenon. Some of these reactions with Ehrlich cellsl17 l2 have been explored in recent papers13 using a multifrequency e.s.r. approach combined with computer simulation of the spectra. In addition to those heavy metals generally considered cytotoxic, essential metals such as Cu and Fe seem to show considerable promise. In theory, the lipid solubility of various complexes of these metals should enable transmembrane passage, after which these complexes could act either selectively as the metal complexes or additively as cytotoxic metabolites.One would then require good stability of these complexes enroute to the active site. Since the original discovery of the enhancement of antitumour activity with metal- complex formation, increased effort has been made to find new and better compounds and to understand their mechanism of action. Using the above-mentioned compounds as models, other drugs have been sought which exhibit similar structural patterns, including possible metal binding sites and a similar potential for enhancement of activity and biological usefulness. Unlike bleomycin which is a large, structurally complex molecule, hydroxyurea is quite simple. Its hydroxamic acid derivatives, however, exhibit some of the same potential 151152 Copper Complexes of Antitumour Agents binding characteristics as the rifamycins, namely an aromatic ring structure with attached amino and hydroxy groups.Hydroxyurea is a known chemotherapeutic agent whose primary mode of action is inhibition of the reaction which converts ribonucleotides to deoxyribonucleotides, and thus is involved in the regulation of DNA synthesis and cellular replication.14-16 Moore has attributed the mode of action of hydroxyurea and its derivatives to their ability to chelate transition meta1s.l' Another proposed mechanism is that the hydroxamic acids interfere with a free radical generated during ribonucleotide reductase action, since these compounds appear readily to form free radicals themselves. Thus, the objective of this study is to explore both the free-radical formation as well as the metal interactions and possible complexation associated with the hydroxamic acid derivatives, compounds which share similar structural components and hopefully, similar behavioural characteristics.In addition, a preliminary exploration of differential cytotoxicity as well as clinical impact is tested. Experimental 3,4-Di hydroxybenzo h ydroxamic acid (3,4-DH B A), gallohydroxamic acid (3,4,5,- trihydroxybenzohydroxamic acid, GHA) and 2,3,4-trihydroxybenzohydroxamic acid (2,3,4-THBA) were obtained courtesy of H. L. Elford, Medical College of Virginia. All the other chemicals were from J. T. Baker Chemical Co., and the solvents were of reagent grade from Aldrich Chemical Co. Because of their inherent propensity to undergo spontaneous oxidation-reduction reactions in aqueous solution, all materials were prepared fresh at the onset of each experiment, with the exception of CuI1 and KCl buffer.Stock 0.2 mol dm-3 KH2P04 buffer at pH 7.8 was used as solvent/buffer for studies of the hydroxamic acid derivatives. The mol dm-3 CuI' solution was prepared by dissolving 0.05 g Cu(NO,), in 25.0 cm3 distilled water. This solution was used in multiple experiments and stored at room temperature. mol dm-3 solutions of 3,4-DHBA, GHA and 2,3,4-THBA were prepared by dissolving appropriate amounts of drug in a 0.1 mol dm-3 KCl solution at pH 7. Solutions were further diluted for study with 0.2 mol dm-3 KH,PO, buffer solution and stored for a maximum of 1 week at room temperature before new solutions were prepared.Copper(I1) nitrate was added via micropipette to sequential samples of 3,4-DHBA in KCl at pH 7 in 1 :2, 1 : 1,2: 1, 3: 1 and 4: 1 concentrations of Cu: 3,4-DHBA. Following complexation, the final pH was adjusted to 7.0 by addition of KOH. A similar sequence was followed for addition of copper(r1) nitrate to GHA and 2,3,4-THBA. Likewise, iron(I1) chloride and iron(m) nitrate were added to 3,4-DHBA, GHA and 2,3,4-THBA following this same method. Once these complexes were generated, various methods of structural and functional analysis were used. A Beckman Acta V spectrophotometer was used for basic charac- terization of 3,4-DHBA, GHA and 2,3,4-THBA in solutions titrated to pH both with nitric acid and KOH over the visible and ultraviolet spectrum.Room-temperature and 77 K e.s.r. spectra were obtained for the CuI1 titration of each drug under study, as well as free-radical generation at pH 2 12 in the absence of appreciable metal ligand. The stabilities of the metal complex in minimal essential media (MEM), fresh dog plasma and ascites fluid were determined by measuring the stability of the observed Cu-drug complex e.s.r. spectra as a function of time. Computer simulation of the data using parameters which gave the best fit for the observed spectra was carried out to resolve complicated hyperfine patterns. Finally, once the basic characterization was completed, preliminary studies of the cytotoxic effect were carried out using Cu-3,4-DHBA with Cu alone, 3,4-DHBA alone 8.5 xR .Basosi, L. Trabalzini, R. Pogni and W. E. Antholine OH 153 OH ( c 1 Fig. 1. Structure of (a) 3,4-dihydroxybenzohydroxamic acid (3,4-DHBA), (b) gallohydroxamic acid (GHA) and (c) 2,3,4-trihydroxybenzohydroxamic acid (2,3,4-THBA). and MEM as controls. The drugs were incubated in a Napco double-CO, incubator with CHO (Chinese hamster ovary) cells, and surviving colonies were counted with a Nikon LKE phase-contrast microscope. This was accomplished with assistance from Dr L. Hopwood in the radiation-biology cell-culture laboratory (Medical College of Wisconsin). The e.s.r. studies were carried out using a Varian E-9 spectrometer with 100 kHz field modulation operating at 9.1 GHz. The magnetic-field scan was calibrated using a Radiopan MJ-1 10R gaussmeter, and the temperature was monitored by a Fluke model 2100 digital thermocouple.Results and Discussion Oxygen is a prevalent donor atom contained in several known classes of antitumour agents, including the mycins and many hydroxyurea derivatives. These compounds thus tend to be good chelating agents, principally bidentate, and are often oxidizable or reduceable in order to form more stable metal complexes. Their stereochemistry seems to be directed towards forming five- or six-membered metal ligand rings. The structures of the compounds studied in this work are shown in fig. 1 . Free-radical Generation Free radicals have been observed to be generated from the anthracycline antibiotics in Z J ~ V O , ~ ~ and this may be related to their mechanism of action.In the absence of metal ion, addition of OH- to hydroxyurea derivatives leads to characteristic free-radical formation, as illustrated by their e.s.r. spectra in fig. 2. Low-power, room-temperature spectra were obtained, and signal averaging was required for clear resolution of the patterns. Free radicals of compounds characterized by a six-membered aromatic-ring hydroxamic acid with either two or three adjacent hydroxy substitutions on the ring were generated. Both trihydroxy derivatives, 2,3,4-THBA [fig. 2 (a)] and GHA [fig. 2 (b)] show considerably more free-radical formation at high pH than the dihydroxy compound 3,4-DHBA [fig. 2 (c)]. Power saturation experiments ruled out contamination by metal cations (data not shown) for all the solutions.A scheme for the 2,3,4-THBA free-radical e.s.r. spectrum at high pH is reported on the right of fig. 2(a); similar schemes are reported in fig. 2(b) and 2(c) for GHA and 3,4-DH BA, respectively. The diagram in fig. 2(a) is consistent with a proton splitting of 5.2 G, a second proton splitting of 1.27 G and a further splitting of 0.26 G whose interpretation is made difficult because of the overlapping of hyperfine lines. The spectrum observed for GHA and shown in fig. 2 (b) is consistent with a splitting of 1.86 G, due to two equivalent protons154 Copper Complexes of Antitumour Agents 0.26 G 1331 1. 06G d%l?h 1.20 G 1 2 2 1 4 1 2 2 1 Gh 0.58G 1 2 1 Fig. 2. Room-temperature e.s.r. spectra (left) and related schemes (right) of the free radicals formed by addition of NaOH to 4 x mol dm-3 solutions of (a) 2,3,4-THBA, (6) GHA and ( c ) 3,4-DHBA (sensitivity increased 10 x ); pH 12.Microwave frequency 9.1 GHz. Fig. 3. Low-temperature e.s.r. spectra of (a) Cu-3,4-DHBA, (b) Cu-GHA and (c) Cu-2,3,4-THBA at pH 7. T = 77 K. 5 mW power. Microwave frequency 9.1 GHz. In (b) and (c) the sensitivity is increased 100 x .R. Basosi, L. Trabalzini, R. Pogni and W. E. Antholine 155 in ortho position to the carbonyl group in the ring. A second triplet splitting with relative intensity 1 : 2: 1 can be attributed to the protons of two equivalent OH groups in the molecule. The scheme for the 3,4-DHBA radical is most consistent with a triplet splitting of 0.58 G, which can be ascribed to two equivalent protons on the ring. However, only small amounts of free radical are seen in this last case [fig. 2(c)].Metal-complex Formation [f copper(I1) ion is added to the hydroxyurea compounds, complexation is unaffected on cycling the pH. Since the same amount of copper(I1) complex is formed at pH 7 and 12, complexation seems to be independent of radical formation prior to introduction of the intended chelating agent. Characteristic e.s.r. data at 77 K are shown in fig. 3. Half-field transitions are not found, and hence Cu-Cu dimer formation is ruled out.20 E.s.r. titrations of ligand 'with copper(r1) ion show maximum complex formation at a ratio of two hydroxyurea ligands to each copper ion in each case. The e.s.r. spectrum for Cu-3,4-DHBA is well characterized showing gll = 2.25, gl = 2.06, Ail = 190 G and .4, = 31 G, and the complex has been observed to be stable in aqueous solution at pH 7 for 24 h.These values are consistent with copper(I1) binding to two aromatic hydroxy groups, and the following structure and reaction sequence is proposed. C-NHOH 0 I- cuz+ +*Hoy& HO C-NHOH s[ HOHN-C ao\Loa 0'1 ' 0 II II II 0 0 Nevertheless, U.V. titration shows that copper is not tightly bound to the molecule. 3,4-DHBA was titrated with CuII at pH 7.6 and the resultant series of spectra are shown in fig. 4(a). Two characteristic peaks at 3 13 and 273 nm are evident. The amount of CuII added is plotted us. absorbance at 3 13 nm in fig. 4(b). The resulting graph illustrates weak binding in the complex formed with an apparent titration constant at 80 mmol dm-3 CuII or a Cu: 3,4-DHBA ratio of 1 : 2.If CuII is added to 3,4-DHBA in a 1 : 2 ratio at pH 5.0, no complexation is observed. If the pH is subsequently increased to ca. 6, complexation can be demonstrated spectroscopically. As the pH is again lowered, the complex formed is stable until pH 5.5, where dissociation occurs [(fig. 4 (c)]. Although u.v.-visible spectra at physiological pH reveal tight binding in complexes formed with a ratio of 1 : 2 for Cu: 2,3,4-THBA and Cu: GHA, the e.s.r. spectra for those compounds are less well characterized [fig. 3(b) and (3c)l. As expected from the high stability of the free radicals generated from 2,3,4-THBA and GHA, in both these cases copper e.s.r. signal is very weak in connection with the limited concentration of the species available for metal complexation.These data are consistent with the tentative picture of dynamic mechanisms and structures shown below. GHA [CU(GHA),]~- + 4H+ -no e.p.r-1.8 1.5 0 5 f 1.2 s 0,s 0 . E 0.2 225 250 275 300 325 wavelength/nm 0.2 2.0 1.5 m " 0 T 1 .o I I I I I I I I I I ~~~ ~ 100 200 [Cu*+]/mmol dm-3 2.4 2.1 1.8 1.5 a, e 4 1.2 0.S 0.E 0.2 250 275 300 325 350 400 wavelength/nm Fig. 4. (a) U.v.-visible titration of 3,4-DHBA with CuII at pH 7.6; (b) graph resulting from plotting CuI1 concentration us. absorbance at 313 nm (vertical line shows 1 : 2 Cu: 3,4-DHBA); (c) u.v.-visible titration of Cu-3,4-DHBA in a 1 : 2 ratio at various pH values: (i) 5.5, (ii) 6.5, (iii) 10; (iv) 3,4-DHBA alone.R. Basosi, L. Trabalzini, R . Pogni and W.E. Antholine 157 I I I I 1 10 20 30 40 50 13,4-DHBA]/pg ~ r n - ~ Fig. 5. Cytotoxicity effect of Cu(DHBA), (0) compared with that of DHBA alone (A) and Cu2+ alone ( s e e ) in CHO cells us. concentration of the agent. Toxicity Studies To postulate the Cu-drug complex as the active cytotoxicity agent in vim, one would require its stability to dissociative competition from various potential ligands in the physiological system. Otherwise, any toxicity observed could be due to the actions of individual components alone. The stability of Cu-3,4-DHBA in various model physio- logic systems was examined. Stability in fresh dog plasma, ascites fluid and minimal essential media (MEM) was studied by incubating these fluids with appropriate quantities of Cu-3,4-DHBA to bring the drug solution to 1.5 x mol dm-3 at 37 "C.The copper(r~) complex of 3,4-DHBA7 once formed, is stable in plasma and MEM. A 4 x mol dm-3 solution of Cu-3,4-DHBA was incubated in 10% MEM at 37 "C for 5 h, and the intensity and shape of the e.s.r. spectrum were unchanged. Because of this stability, it can be postulated that the binding of CuII to 3,4-DHBA is tighter than the bindings of CuII to plasma histidine at 37 "C in the system used. Therefore, Cu-3,4-DHBA was chosen as the ideal candidate for preliminary cytotoxici ty studies with CHO cells. To study the antitumour effect of these metal-drug complexes, the complex was added in varying concentrations to CHO cells, and surviving colonies were counted after incubation. The survival data for both ligand and Cu-ligand complex are shown in fig.5. One observes enhanced cytotoxicity in the presence of complexation, and the characteristic shape of the graph illustrates single phase killing during the cell cycle. The enhanced cytotoxic effect shown for the copper(I1) complex differs strikingly from the effect expected for copper(I1) ion alone. Elford et aZ.18 demonstrated the similarity between hydroxyurea and 3,4-DHBA in inhibiting ribonucleotide reductase, although the IDso? for 3,4-DHBA was 30 pmol dmP3 and that for hydroxyurea was 500 pmol dm-3. Thus, our newly synthesized drugs may be important in that they could circumvent the major drawback of hydroxyurea in human chemotherapy, i.e. the requirement of large doses to achieve maximum efficacy. Further trials, however, must be carried out to observe in vitru activity and observe any potentially dangerous side-effects.t The concentration of drug required to inhibit the observed activity of the enzyme by 50%.158 Copper Complexes of Antitumour Agents Conclusions This study attempted to illustrate metal complexation of known antitumour agents, the hydroxyureas, using three analogous benzohydroxamic acids, 3,4-DHBA, GHA and 2,3,4-THBA. The CuI1 complexes of these drugs were well characterized using u.v.-visible spectrop ho tome try and e. s. r . spectroscopy . It was discovered during basic characterization of these drugs that, although similar in structure, they showed significant differences in behaviour. Benzohydroxamic acid binding to CuII follows a seemingly pH-independent mechanism.Furthermore, no evidence of dimer formation is exhibited, However, each derivative studied, although differing only in the number of hydroxy groups, exhibited a unique complexation reaction or series of reactions. 3,4-DHBA7 the only dihydroxy compound tested, was best characterized by u.v.-visible and e.s.r. spectroscopy forming a 1 : 2 (Cu: 3,4-DHBA) complex which is stable in biological media. The two trihydroxy derivatives appear to undergo a more complex series of reactions and are less well characterized by u.v.-visible and e.s.r. spectroscopy. Since free-radical formation is a postulated mechanism of the in vivo action of these compounds, this study also attempted to determine if free radicals could indeed be formed, and to characterize those that did appear.The dihydroxybenzohydroxamic acid derivative 3,4-DHBA formed only a weak free radical, but both trihydroxybenzohy- droxamic acid derivatives gave strong free-radical signals. Those compounds forming large quantities of free radical are poorly characterized by e.s.r. with regard to copper(I1) complexation, while those forming only weak free-radical signals form well characterized copper(I1) complexes. This may indeed reflect the mechanism of complexation, since it is known that CuII binding to a free radical can result in dipole coupling and no detectable e. s.r. spectrum. 21 Once it has been shown that copper complexes are indeed formed and they have been structurally characterized, studies of in vitro and in vivo cytotoxic activity can begin.Because of its stability in biological media and its distinctive and easily followed e.s.r. spectrum, Cu-3,4-DHBA was used in our preliminary study. Enhanced cytotoxicity was indeed shown towards CHO cells for the cupric complex. Before any definitive results can be known, future repetitive trials are necessary in tumour-cell culture followed by animal model studies with both transplanted solid tumours and leukaemic mice. How- ever, our study indicates that the major difficulty with hydroxyurea, i.e. that huge quantities of drug need be given systemically, may indeed be circumvented by using well chosen derivatives. Should systemic toxicity permit, copper complexes may prove to be such derivatives. This work was supported by NIH grant GM-35472-01 (U.S.A.) and the C.N.R.(Italy). References 1 J. C. Dabrowiak, F. T. Greenway, W. E. Longo, M. Van Husan and S. T. Crooke, Biochim. Biophys. 2 J. R. White, Biochim. Biophys. Res. Commun., 1977, 77, 387. 3 W. E. Antholine, J. M. Knight, H. Whelan and D. H. Petering, Mol. Pharmacol., 1977, 13, 89. 4 W. E. Antholine, J. M. Knight and D. H. Petering, Med. Chem., 1976, 19(2), 339. 5 W. E. Antholine, J. M. Knight and D. H. Petering, Znorg. Chem., 1977, 16, 569. 6 K. C. Agrawal, B. A. Booth, E. C. Mooreand A. C. Sartorelli, Proc. Am. Assoc. Cancer Res., 1974,15, 7 B. Rosenburg and L. Von Camp, Cancer Res., 1970, 30, 1799. 8 Y. Iitaka, H. Nakamura, T. Nakatani, Y. Moraoka, A. Fujii, T. Takita and H. Umezawa, J. Antibiotics, 9 T. Takita, Y. Muroka, T. Nakalani, A. Fujii, Y. Iitaka and H. Umezawa, J . Antibiofics, 1978, 31(10), Acta, in press. 289. 1978, 31(10), 1970. 1073.R . Basosi, L. Trabalzini, R. Pogni and W. E. Antholine 159 10 W. E. Antholine, G. Riedy, J. S. Hyde, R. Basosi and D. H. Petering, J. Biomol. Struct. Dynam., 1984, 11 D. H. Petering, Inorganic and Nutritional Aspects of Cancer, ed. G. N. Scrauzer (Plenum Press, New 12 W. E. Antholine, S. Lyman, D. H. Petering and L. Pickart, Biological and Inorganic Copper Chemistry, 13 W. E. Antholine, R. Basosi, J. S. Hyde, S. Lyman and D. H. Petering, Znorg. Chem., 1984,23, 3543. 14 H. Elford, Adv. Enzyme Regul., 1972, 10, 19. 15 A. Larson and P. Reichard, Proc. Nucleic Acid Res. Mol. Biol., 1967, 7 , 303. 16 M. K. Turner, R. Abrams and I. Lieberman, J. Biol. Chem., 1968, 243, 3725. 17 E. C. Moore, Cancer Res., 1969, 29, 291. 18 H. L. Elford, G. L. Wanyler and B. Van’t Riet, Cancer Res., 1979, 39, 844. 19 S. Sato, M. Iwaizumi, K. Handa and Y. Tamura, Gazzetta, 1977, 68, 603. 20 J. F. Boas, J. R. Pilbrow, C. R. Hartzell and T. D. Smith, J. Chem. SOC. A , 1969, 572. 21 Y. S. Hyde and T. Sarna, J. Chem. Phys., 1978,68,4439. 2(2), 469. York, 1977), p. 179. ed. K. D. Karlin and Y. Zubieta (Adenine Press, New York, 1985), vol. 1, p. 125. Paper 6/ 1 13 1 ; Received 5th June, 1986 6 FAR 1

ISSN:0300-9599    

DOI:10.1039/F19878300151

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1987, 83, 161-166 The Role of Solvent Reorganization Dynamics in Homogeneous Electron Self-exchange Reactions Gunter Grampp,* Wolfgang Harrer and Waltber Jaenicke Institute of Physical and Theoretical Chemistry, University of Erlangen, Egerlandstrasse 3, 0-852 Erlangen, Federal Republic of Germany The homogeneous electron self-exchange reaction of N,N,N',N'-tetramethyl- p-phenylenediamine (TMPPD/TMPPD'+ ) shows the classical solvent de- pendence according to Marcus' theory. The rate constants are a linear func- tion of y = (l/n2 - 1 / ~ ) (n = refractive index, E = dielectric constant of the solvent). For 0.27 d y < 0.53, log k,, = const. x y. However, the reactions of tetracyanoquinonedimethane (TCNQ/TCNQ'- ) and tetracyanoethylene (TCNE/TCNE'-) are dominated by the solvent dynamics, including the longitudinal relaxation time zL of the solvent.This is expressed by log (kexzLy-1/2) = const. x y, for 0.05 < y < 0.53. The role of solvent reorgan- ization dynamics is discussed together with the decision equation of Ovchinnikova. Theory Usually the rate constants of electron-transfer reactions are related to the activation energies AG*. In his landmark Marcus pointed out that the rate constants of homogeneous electron-transfer reactions k e x Q+Q" e Q'* +Q k,, = Ar, exp [-(AGE + AG,*,)/RT]. are given by F;or homogeneous electron transfer, the standard free reaction enthalpy AGO is zero, and no Coulomb work term between the two reactants in eqn (1) has to be considered AGE and AG& are a quarter of the inner- and oiter-sphere reorganization energies, respectively.The inner-sphere contribution to reach the transition state is expressed by (WR, = 0). AGfs is a function of the force constantsf, and the difference in bond length Aqj of both reactants, summarized over all bonds. In the classical theory the solvent dependence of k,, is expressed by extending the Born m~del.l-~ Within the dielectric continuum treatment AG& depends on the molecular radius F and the reaction distance d: (4) where the solvent parameter y, the Pekar factor from polaron theory, is given by y = ( l/n2 - 1 / E ) (n = refractive index and E = static dielectric constant of the solvent). Marcus expressed the pre-exponential factor Ar, by a simple gas-phase collision factor [Ar, = 2 = ( ~ F ) ~ N , (16dW/M)1'2], whereas in Sutin's43 'encounter pre- equilibrium model ', the activation process follows the formation of the precursor complex.The precursor equilibrium constant A can be expressed by A = 4nNLPSr; the AG,*, = e: NL/( 16 mo)( 1 / ~ - l/d)y 161 6-2162 Soluen t Reorganization Dynamics nuclear frequency factor r, for surmounting the free-energy barrier is usually given by a mean value of solvent (r,) and reactant frequencies6* ’ (Ti): r, = [(rf AG,*, + r: AG&)/(AGg + AG&)]i. ( 5 ) The most important conclusion of recent theoretica18-12 studies is the fact that the solvent influences the rate of electron-transfer reactions also via the pre-exponential factor. Eqn ( 5 ) describes the limiting case of rapid dielectric relaxation of the solvent.It has been recognized that the effective value of ro can be closely related to the longitudinal solvent relaxation time zL (the relaxation time of ‘constant charge’).13 For slow relaxing solvents ro is a function of z i l : ro = 7~ (AG&/4nRT)i. (6) The quantity zL is related to the Debye relaxation time zD by ZL = zD = 3 V M ~ E ~ / E R T (7) where zL depends on the static and high-frequency dielectric constants E and P, and zD on the viscosity q and on the molar volume VM. It should be noted that is usually approximated by n2. [ E ~ z n2. differs from n2 only by vibrational relaxation effects. For details see ref. (1 3).] Under these circumstances r, becomes a function of zzl: rn = z,l (AG,*,/~zRT)~ (8) (9) and the pre-exponential factor in the rate expression becomes solvent-dependent :Is k,, = AT, (AG&I4nRT)i exp[ - (AG:s + AG,*,)/RT].[AG:s/(AG:s + AG&)]i ri exp (- AG&/RT) > z, l. The condition for this limiting case is given by an inequality derived by Ovchinnikova:l4? l5 (10) The equilibrium constant A is not influenced by the solvent, because the Coulomb work terms are zero. Therefore, for substances fulfilling eqn (1 0), the solvent dependence of the rate constants is given by log (k,, ~ ~ 7 - 4 ) = const. x y log k,, = -const’. x y. (1 1) according to eqn (8). However, the ‘classical’ solvent dependence according to Marcus leads to (12) This paper reports on the different solvent behaviour of homogeneous electron- exchange reactions of N,N,N’,N’-tetramethyl-p-phenylenediamine (TMPPD), tetracy- anoquinonedimethane (TCNQ) and tetracyanoethylene (TCNE).The measurements were carried out using a wide y-range of solvents (0.27 < y -c 0.53 for TMPPD and 0.05 < y < 0.53 for TCNQ and TCNE). Experimental The rate constants were determined by analysing e.s.r. linewidth effects.17 The method of evaluation and experimental details with reference to substance and purification of the solvents are described in ref. (1 7) and (1 8). The substances used and the solvents are listed in table 1. Results and Discussion Table 1 lists the measured rate constants corrected for diffusion, according to 1 lkex = 1 lkobs - 1 I b i f f -G. Grampp, W. Harrer and W. Jaenicke 163 9.0- Table 1. Homogeneous electron-exchange rates at T = 298 K k C HCi3 DME DMSO DMF CH3N0#P$N 1 TCNQ TCNE dioxane 0.05 methyoxybenzene 0.20 bromobenzene 0.22 CHCl, 0.27 THF 0.37 CH2C12 DME 0.38 0.39 DMSO 0.44 DMF 0.46 propylene carbonate 0.48 nitromethane 0.50 CH,CN 0.53 10.0 6.1 4.86 2.2 0.8 0.9 0.86 2.4 1.3 2.6 0.2 - 8.80 - 8.97 9.84 9.27 - 9.31 - 9.52 9.46 - 9.29 8.83 9.21 - - 8.81 9.13 - 9.07 9.67 - - - - 1.55 -2.01 - 2.10 -2.57 -2.31 - - - 2.62 - -2.61 - - 2.89 8.57 8.46 9.33 9.60 8.78 9.60 8.62 - 1.82 - 2.40 - 1.65 - 1.77 - 3.10 - 2.23 - 3.24 - 8.60 9.40 - - - 2.82 - - 3.16 a Data from Digest of Literature on Dielectrics (Natl Acad. Sci.Natl Res. Council, Washington, D .C., 1 9 57- 1 979). Y Fig. 1. The rate constant of the homogeneous electron self-exchange reaction TMPPD/TMPPD'+ as a function of y [see eqn (1 2)].kdiff is obtained from the Smoluchovski equation kdiff = 8RT/3q. The AG,*, values for TMPPD/TMPPD'+ radical cation and TCNQ/TCNQ'- radical anion were obtained using crystallographic bond-length datala, l9 and the force constants from Raman data. For TCNE, Huckel bond-order relations were used, similar to ref. (17). For the TMPPD/TMPPD'+ couple the left side of eqn (10) is smaller than zt1 for all solvents used. Therefore the rate constants should be directly proportional to y, according to eqn (12). Fig. 1 displays this relation for T = 293 K. The straight line164 - 1.5- n -N I+ -2.0 - Solvent Reorganization Dynamics -1.5 n .-P( '* -2.0 - cl X c $ -25- - i 0.1 0.2 0.3 0.4 0.5 0.6 Y Fig. 2. The solvent dependence of the electron self-exchange reaction TCNQ/TCNQ'- [see eqn (1 1)l.\ I 011 d.2 Oy3 0!4 Oj5 016 b Y Fig. 3. The solvent dependence of the electron self-exchange reaction TCNE/TCNE'- [see eqn (1 113. obtained indicates the classical behaviour in the sense of Marcus' theory. Classical solvent behaviour is also reported for the bis(biphenyl)Cr0/+,20 cobalt clathochelatesO/+ 21 and tris(hexafluor~acetonate)Ru~+/~+ couples.22 In contrast, the rate constants of the TCNQ/TCNQ*- and TCNE/TCNE*- couples increase with decreasing zL of the solvent. According to eqn (8) the rate constants of these systems are expressed by eqn (1 1). Fig. 2 and 3 show these relations. Unlike these systems ferrocene/ferri~inium~~ and hydrazineO/+ 24 show quantitative deviations of eqn (12). In the electrochemical heterogeneous case, the influence of solvent reorganization dynamics according to eqn (9) is evident for phen~thiazine,~~G.Grampp, W. Harrer and W. Jaenicke 165 p-phenylenediamine26 and nitr~mesitylene.~~ Recently Bard and coworkers28 found a dependence of the heterogeneous rates of Fe(CN);-i4- and ferrocene/ferricinium electron transfer in H20 and DMSO, respectively, on viscosity, changed by adding certain amounts of sucrose. khet q-l and khet x yl-0.7 were reported. Performing the same experiments for the homogeneous TCNQ/TCNQ'- reaction, we did not find any remarkable change in k,, with sucrose con~entration.~~ A specific solvation of the TCNQ/TCNQ'- reactant pair in DMSO can be assumed and therefore the increasing viscosity influences only kdiff in eqn (1 3). Recently2s it was assumed that in heterogeneous and homogeneous electron-transfer reactions the activated complex may have different structures and that this may be the reason for different solvent behaviour.Perhaps an additional rotational orientation of the reactants in the homogeneous case can influence r030931 since ro is proportional Further examinations of solvent-dependent reactivities are necessary, both in homo- to z;&. geneous solution and at electrode surfaces. Summary This paper shows the different solvent dependence of homogeneous electron self-exchange reactions for TMPPD/TMPPD*+, TCNQ/TCNQ'- and TCNE/TCNE'-. This difference originates in the solvent dependence of the pre-exponential factor in the rate expression. The nuclear frequency factor r, becomes a function of z ~ l , if it is controlled by the solvent reorientation.Note added in pruofi Recently Marcus and S ~ m i ~ ~ have published important papers on the solvent dynamics of electron-transfer reactions. References 1 R. A. Marcus, J. Chem. Phys., 1956, 24, 966; 979. 2 R. A. Marcus, Annu. Rw. Phys. Chem., 1964, 15, 155, 3 R. A. Marcus, Int. J. Chem. Kine?., 1981, 13, 865. 4 N. Sutin, Progr. Inorg. Chem., 1983, 30, 441. 5 M. D. Newton and N. Sutin, Annu. Rev. Phys. Chem., 1984,35,457. 6 N. B. Slater, Theory of Unimofecular Reactions (Cornell University Press, Ithaca, New York, 1959). 7 M. D. Newton, Int. J. Quant. Chem. Symp., 1980, 14, 363. 8 L. D. Zusman, Chem. Phys., 1980,49, 295. 9 G. van der Zwan and J. T. Hynes, J. Chem. Phys., 1982,76,2993. 10 I. V. Alexandrov, Chem.Phys., 1980,51,449. 1 1 D. F. Calef and P. G. Wolynes, J. Phys. Chem., 1983,87, 3387. 12 T. Gennett, D. F. Miller and M. J. Weaver, J. Phys. Chem., 1985,8!4, 2787. 13 H. Frohlich, Theory of Dielectrics (Oxford University Press, London, 1949). 14 M. Ya. Ovchinnikova, Russ. Theor. Exp. Chem., 1981, 17, 507. I5 A. M. Kuznetzov, Elektrokhimiya, 1984, 20, 1226. 16 M. J. Weaver and T. Gennett, Chem. Phys. Lett., 1985, 113,213. 17 G. Grampp and W. Jaenicke, Ber. Bunsenges. Phys. Chem., 1984,88, 325; 335. 18 G. Grampp and W. Jaenicke, J. Chem. SOC., Faraday Trans. 2, 1985,81, 1035. 19 W. Harrer, G. Grampp and W. Jaenicke, Chem. Phys. Lett., 1984,112, 263. 20 T. T. T. Li, M. J. Weaver and C. H. Brubaker, J. Am. Chem. SOC., 1982,104,2381. 21 D. Borchardt and S. Wherland, Inorg. Chem., 1984,23, 2537. 22 M-S. Chan and A. C. Wahl, J . Phys. Chem., 1982,80, 126. 23 E. S. Yang, M-S. Chan and A. C. Wahl, J. Phys. Chem., 1980,84, 3094. 24 S. F. Nelson and S . C. Blackstock, J . Am. Chem. Soc., 1985, 107, 7189. 25 M. Opallo and A. Kapturkiewicz, Electrochim. Acta, 1985,30, 1301. 26 M. Opallo, J . Chem. SOC., Faraday Trans. 1, 1986,82, 339. 27 A. Kapturkiewicz and M. Opallo, J. Electroanal. Chem., 1985, 185, 15. 28 X. Zhang, J. Leddy and A. J. Bard, J . Am. Chem. SOC., 1985,107, 3719. 29 W. Harrer, G. Grampp and W. Jaenicke, J . Electroanal. Chem., in press.166 Solvent Reorganization Dynamics 30 W. Jaenicke, G. Grampp and W. Harrer, Proc. VIIth Symp. Electrochem. React. Nonaqueous and Mixed 31 P. Siders, R. J. Cave and R. A. Marcus, J . Chem. Phys., 1984,81, 5613. 32 H. Sumi and R. A. Marcus, J . Chem. Phys., 1986, 83, 4272; 4894; R. A. Marcus and H. Sumi, Solvents, Poznan, Poland 1986. J. Electroanal. Chem., 1986, 204, 59. Paper 61854; Received 1st May, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300161

出版商:RSC    

年代:1987

数据来源: RSC