摘要:

J.. Chem. SOC., Faraday Trans. I, 1987, 83, 167-176 Linewidth Alternation, as a Result of Intramolecular Cation Migration, in the Electron Spin Resonance Spectrum of the 1,4-Naphthoquinone Radical Anion Nicholas J. Flint and Brian J. Tabner" The Department of Chemistry, The University of Lancaster, Lancaster LA1 4YA The radical anion of 1,4-naphthoquinone has been prepared by electro- chemical reduction in dimethylformamide and by alkali-metal reduction in 1,2-dimethoxyethane. The e.s.r. spectrum of the radical anion in dimethyl- formamide can be readily interpreted in terms of three pairs of equivalent protons. However, in 1 ,Zdimethoxyethane, association of the alkali-metal counter-ion with one of the carbonyl groups removes the equivalence within each of these pairs. Linewidth alternation is observed in the e.s.r.spectrum when either Na+, K+ or Cs+ is the counter-ion as a result of intramolecular cation migration between equivalent sites adjacent to the two carbonyl groups. The rate of migration varies in the sequence Cs+ > K+ > Na+ > Li+, indicating that Li+ interacts more strongly with the radical anion than (say) Cs+. The disruption of the spin distribution within the radical anion, which is also greatest when Li+ is the counter-ion, supports this conclusion. A detailed study of an e.s.r. spectrum over a range of temperatures can give valuable information on processes compatible with the timescale of the e.s.r. experiment. Particularly interesting are those processes which result in linewidth variations as a consequence of cation migration, conformational interconversion and restricted rotation. One of the first examples in which intramolecular cation migration was postulated to account for linewidth variation was the pyracene radical anion when formed by alkali-metal red~ction.l-~ This particular system proved to be more informative than many others, as two quite distinct cation migrations have been observed under different conditions.One of these processes involves migration of the cation from one side of the molecular plane to the other, whilst the other process involves oscillation of the cation between two equivalent sites on the same side of the molecular plane. Cation migration from one side of the molecular plane to the other has also been observed in the radical anions of acenapht hene, 7 9 9,lO-di hydroan thracene,6 tetra- t-bu tylnap hthalene, benzocyclobutenes and 5,12-dihydrotetra~ene.~ In addition the influence of specific solvating agents, such as crown ethers, on the rates of these processes has also been in~estigated.l~-~~ Other early examples of intramolecular cation migration between two equivalent sites were reported for the radical anions of pyrazine,14-16 benzoquinone17 and various methyl-substituted benzoquinones.1s-20 Again, the addition of crown ethers influences the rate of cation migration in some of these ~ystems.~l-~~ Several nice illustrations of conformational interconversions were also among the early reports of spectra exhibiting linewidth variations.One of the first of these reports concerned the 1,2,3-trihydropyrenyl in which the axial and equatorial 8-protons are not equivalent at low temperature. Linewidth alternation was observed as the temperature was increased as a result of an increase in the rate of interconversion.Similar interconversions were observed shortly afterwards in the radical anions of 1,2,3,6,7,8 -hexahydr~pyrene~~ 3 27 and 4,5,9,10- te trah ydr op yrene. 27 Restricted rotation can also lead to linewidth variation in an e.s.r. spectrum. One of 167168 1,4-Naphthoquinone E.S.R. Spectrum the first examples of linewidth alternation to be recognised resulted from restricted rotation of the hydroxy group in the duroquinol radical cation.28 More recently, restricted rotation has been observed in the benz~phenone~~ radical anion and in some thienyl- and furanyl-methanone radical In the majority of e.s.r.spectra the width of each line in the spectrum is the same, and consequently information on the types of process described above cannot be obtained. Conformational interconversion and restricted rotation are, of course, not possible in many systems. However, even in systems where cation migration, conforma- tional interconversion and restricted rotation are possible in principle, the rates of these processes often occur outside the timescale of the e.s.r. experiment. When the lifetime of the species is too great for linewidth variation to be observed, useful information can still be obtained. For example, equilibria between different types of ion pair, or between ion pairs and free ions, are often observed, as are distinct spectra from more than one isomer.A detailed study of such spectra over a temperature range can therefore give information on these equilibria. If the lifetime of the different species is too small for linewidth variation to be observed then a time-averaged spectrum usually results. Even in systems where none of the processes described above is possible, useful information can still be obtained. For example, the sign of each hyperfine splitting constant is usually indicated by the sign of the corresponding da/dT value.33 In this paper we report the results of our investigation of the e.s.r. spectrum of the 1,4-naphthoquinone radical anion which should, at least in principle, exhibit an alternating linewidth effect due to intramolecular cation migration between equivalent sites adjacent to the two carbonyl groups.In previous studies of this radical anion, which include its electrochemical preparation in dimethylf0rrnamide~~1 35 and its preparation by potassium reduction in hexamethylph~sphoramide,~~ linewidth alternation was not reported. Our first experiments in tetrahydrofuran and 1,2-dimethoxyethane revealed that, in these two solvents, the spectra were largely obscured by a broad single line resulting from precipitation of the radical anion. However, we were able to overcome this problem by preparing the sample in 1,2-dimethoxyethane at low temperature (ca. - 70 "C) and recording a reasonable number of spectra below room temperature before precipitation became too extensive. We report below the results of our investigation of linewidth alternation in the e.s.r.spectrum of this radical anion (with Li+, Na+, K+ or Cs+ as counter-ion) resulting from intramolecular cation migration. Experimental All vacuum operations were performed using standard high-vacuum techniques. Materials 1,2-Dimethoxyethane (Aldrich) was purified as described previo~sly~~ and dried employ- ing sodium/potassium alloy. Dimethylformamide (Lancaster Synthesis) was distilled under reduced pressure and stored over a 4 A molecular sieve before use. 1,4- Naphthoquinone (B.D.H.) was purified by soxhlet extraction and repeated recrystal- lisation from (40/60) petroleum ether. Lithium, sodium and potassium were all washed in light petroleum before use. Procedure Solutions of 1,4-naphthoquinone in 1,2-dimethoxyethane were prepared by the standard methods described previ~usly.~~ The concentration of 1,4-naphthoquinone was in the range (3-4) x mol dm-3 and the radical anion was prepared from these solutions, by alkali-metal reduction, as de~cribed.~' Metal films were used in the experimentsN .J . Flint and B. J . Tabner 169 involving sodium or potassium and small pieces of metal in experiments involving lithium. The metal film used in experiments involving caesium was prepared by the careful heating of a mixture of caesium chloride and calcium turnings. The concentration of the parent ketone in dimethylformamide was in the range (3-5) x mol dmP3, with tetra-n-butylammonium iodide (B.D.H.) as supporting electrolyte (0.1 mol dm-3). The radical anion was prepared by electrochemical reduction from this solution as described previo~sly.~~ Spectroscopic Measurements All e.s.r. spectra were recorded on a Varian E3 spectrometer, the magnetic field sweep of which was calibrated with Fremy’s The temperature of the samples was controlled by means of a Varian E4557 variable-temperature unit.Computer Simulation of E.S.R. Spectra Computer simulations of e.s.r. spectra were obtained using a Data General Nova 1220 computer on line to the spectrometer or on a Vax 11/785 computer linked to a Calcomp plotter. The program used in the simulation of spectra exhibiting an alternating linewidth effect employed the modified Bloch equations appropriate to a two-jump process in which three pairs of two non-equivalent protons interchange their splitting constants.The hyperfine splitting constants quoted from these simulations are considered accurate to within & 1 pT when z > ca. 0.25 x s, but +2 pT when z < ca. 0.25 x 1 0-6 s. Spectra not exhibiting linewidth alternation were computer-simulated, employing a conventional spectrum simulation program. The hyperfine splitting constants obtained from these latter simulations are also considered accurate to 1 pT. Molecular-orbital Calculations Molecular-orbital calculations were undertaken using a Vax 1 1 /785 computer. The influence of an alkali-metal counter-ion upon the spin distribution in the radical anion was simulated by applying a perturbation to the value of a, for the oxygen atom associated with the carbonyl group at position 1.Results and Discussion Electrochemical Reduction in Dimethylformamide ’The e.s.r. spectrum of the 1,4-naphthoquinone radical anion, prepared by electrochemical reduction in dimethylformamide, has been published previo~sly,~~ and our results agree closely with those in this previous report. Spectra were recorded over the temperature range 235-276 K [see fig. 1 (a)] and are readily interpreted in terms of a(2H) 327, 60 and 31 pT [see fig. 1 (b)]. The values of these hyperfine splitting constants are virtually independent of temperature over the range covered and, as expected for a radical anion prepared by electrochemical reduction, there is no evidence for linewidth alternation. It is clear that the largest splitting constant (327 pT) should be assigned to the two equivalent protons at the 2 and 3 positions.We anticipated that the splitting constant with a value of 60pT would be associated with the protons at the 6 and 7 positions, with the smallest splitting constant (31 pT) being associated with protons at the 5 and 8 positions. In order to confirm these expectations we have undertaken a series of Huckel molecular-orbital calculations with a, = a, + 1.4 pCc and pco = 1.7 pee. These values are fairly typical of those normally employed to represent a carbonyl g r ~ u p . ~ O - ~ ~ In addition the McLachlan parameter, A, was taken to be 1.2,43 and the value of lQl to be170 1,4-Naphthoquinone E.S.R. Spectrum Fig. 1. The experimental e.s.r. spectrum of the 1,4-naphthoquinone radical anion, in dimethyl- formamide, at 235 K (a) together with its computer reconstruction (b).2.2 mT. The predicted splitting constants obtained from these calculations are sum- marked in table 1 together with the corresponding experimental values. Alkali-metal Reduction in 1,2-Dimethoxyethane The radical anion of 1,4-naphthoquinone was prepared in 1,2-dimethoxyethane by reduction with lithium, sodium, potassium and caesium. In each preparation a pale green solution of the radical anion was readily formed. The lowest temperature at which spectra were recorded in each system was dependent upon spectrum intensity and resolution. This was in the range 193-218 K when either Na+, K+ or Cs+ was the counter-ion. The highest temperature at which spectra were recorded was dependent upon precipitation of the radical anion as the temperature was raised.In some systems it was still possible to obtain useful data from spectra even when precipitation was present, as the resulting broad single line did not obscure the high- and low-field groups of lines. When lithium is the counter-ion, strong association of the counter-ion with one of the carbonyl groups is observed and the spectrum can be interpreted in terms of coupling to six non-equivalent protons. For each of the other alkali-metal counter-ions linewidth alternation due to intramolecular cation migration between sites adjacent to the two equivalent carbonyl groups is observed, the rate of the process depending upon the counter-ion.N . J . Flint and B. J . Tabner 171 Table 1. A summary of the 'slow' exchange experimental (best-fit computer simulation) and theoretical (McLachlan molecular-orbital calculation) hyperfine splitting constants (in pT) for the 1,4-naphthoquinone radical anion prepared by electrochemical reduction in dimethylformamide and by alkali-metal reduction in 1,2-dimethoxyethane system T/K 42-H) 43-H) 45-H) 48-H) 46-H) 47-H) experimental calculated experimental, Li+ calculateda experimental, Na+ calculatedb experimental, K+ experimental, Cs+ calculated" electrochemical reduction 235 327 31 336 30 - alkali-metal reduction 260 192 462 16 72 - 202 452 18 72 239 24 1 41 5 13 49 - 238 420 21 61 235 262 387 14 46 237 270 382 16 43 268 394 24 51 - 97 93 81 86 85 82 80 1 60 66 29 40 36 47 44 41 53 Calculated with an a.value of: a ac+2.05 Bee, carbonyl oxygen at position 1 and a, + 1.4 pCc for the carbonyl oxygen atom at position 4.a,+ 1.85 BCc and " a,+ 1.70 pCc for the Table 2. Hyperfine splitting constants (in pT) for the 1,4- naphthoquinone radical anion in 1,2-dimethoxyethane with Li+ as the counter-ion, obtained by best-fit computer simulation T/K 42-H) 43-H) 45-H) 46-H) 47-H) 48-H) Z / ~ S 260 192 462 16 97 29 72 > 5 268 186 469 13 99 24 73 > 5 284 184 472 5 99 19 72 > 5 294 180 479 5 100 19 76 > 5 309 178 480 5 100 19 76 > 5 Lithium When the 1,4-naphthoquinone radical anion was prepared by reduction with lithium metal the resulting spectra could be interpreted in terms of coupling to six non-equivalent protons. It was possible to obtain useful data from the spectra recorded over the temperature range 260-309 K and a computer simulation of each of these spectra was obtained.These simulations were obtained using the alternating linewidth program described above. However, identical simulations were obtained employing a constant- linewidth program, indicating that the lifetime (z) of the Li+ ion adjacent to one particular carbonyl group was too great for linewidth alternation to be observed. Computer simulations reveal that variations in the height and width of various lines in the spectrum can be observed when the species lifetime is <ca. 5 x 10+ s. We can therefore only suggest a lower limit for z in this system. The various best-fit parameters obtained from the spectrum simulations at each temperature are summarised in table 2. The hyperfine splitting constants have been provisionally assigned by comparing their average values with those observed for the radical anion prepared by electrolytic reduction.For example, the a values (at 260 K) of 192 and 462 pT average 327 pT and are clearly associated with the protons at the 2 and 3 positions (cf. table 1). Similarly172 1,4-Naphthoquinone E.S.R. Spectrum Table 3. Hyperfine splitting constants (in pT) for the 1,4- naphthoquinone radical anion in 172-dimethoxyethane with Na+ as the counter-ion, obtained by best-fit computer simulation, together with appropriate values of z ~~ ~ T/K 42-H) 43-H) 45-H) 46-H) 47-H) 48-H) ~ / 1 0 - ~ s 218 240 416 13 81 36 49 85 229 242 414 13 81 36 49 65 239 241 415 13 81 36 49 55 the values of 16 and 72 pT average at 45 pT and the values of 29 and 97 pT average at 63pT.Again, by comparison with the values in table 1 these would appear to be associated with the 5 and 8 positions and with the 6 and 7 positions, respectively. The remaining problem is the assignment of the individual values within each pair to the appropriate protons. As an aid to this assignment we have undertaken further molecular- orbital calculations employing an auxiliary parameter associated with one a, to allow for the influence of the alkali-metal ion on that particular carbonyl group. Calculations on the benzoquinone radical anion indicate that an increase of ca. 0.4 to Pcc in the a, value (i.e. cc, = ac + 1.8 Pcc at position 1) reproduces the experimental hyperfine splitting constants in this system quite nicely.In the 1,4-naphthoquinone radical anion system, however, it is clear that the disruption of the spin distribution varies with the alkali-metal ion (being greatest for Li+). In this particular system therefore a value of a, = a, + 2.05 Pcc gives closest agreement to the experimental splitting constants, and the individual splitting constants within each pair have been assigned on the basis of these calculations (see table 1). sodium Unfortunately it was only possible to obtain spectra of reasonable quality at three temperatures when Na+ was counter-ion. However, it is clear that in order to computer- simulate these spectra the presence of an alternating linewidth effect has to be assumed. The various best-fit parameters for these simulations are summarised in table 3, and the experimental splitting constants have been assigned to particular protons employing the same procedure as described above for lithium (see table 1).Despite the relatively few number of data points we have estimated E, for intramolecular cation migration to be 9 kJ mol-l, although there may well be a significant error on this value. Potassium It was possible to record spectra of the 1,4-naphthoquinone radical anion over the temperature range 193-276 K when K+ was the counter-ion. This system therefore provides the most extensive results of all four alkali-metal counter-ions. Fortunately, linewidth alternation is present in these spectra over almost the entire temperature range, with the spectrum at 193 K featuring virtual ‘slow’ exchange and the spectrum at 276 K featuring virtual ‘fast’ exchange.A selection of three spectra (at 193, 235 and 268 K) are illustrated in fig. 2[(a), (b) and (c), respectively] together with their computer simulations [(d), (e) and d f ) , respectively]. The various best-fit parameters required for the simulation of each spectrum are summarised in table 4. The experimental hyperfine splitting constants have been assigned to particular protons employing exactly the same procedure as before (see table 1). We have also obtained a value for the activation energy for intramolecular cation migration in thisN . J . Flint and B. J . Tubner 173 268 K 103 K tn /-- -- Fig. 2. The experimental e.s.r. spectra of the 1,4-naphthoquinone radical anion, in 1,2- dimethoxyethane with K+ as the counter-ion, at 193 (a), 235 (b) and 268 K (c), together with their computer simulations [(d), (e) and cf), respectively; parameters as given in table 41.Table 4. Hyperfine splitting constants (in pT) for the 1,4- naphthoquinone radical anion in 1,2-dirnethoxyethane with K+ as the counter-ion, obtained by best-fit computer simulation, to- gether with appropriate values of T T/K 42-H) 43-H) 45-H) 46-H) 47-H) 48-H) ~ / 1 0 - ~ s 193 263 384 15 80 42 49 300 197 262 387 15 80 43 45 200 207 262 387 15 80 43 45 80 219 260 389 16 85 43 43 42 229 262 387 16 85 44 46 30 235 262 387 14 85 44 46 20 246 262 385 14 85 44 46 12.5 257 262 385 14 85 44 46 7 261 262 387 14 85 44 46 6 268 262 387 14 85 44 46 5 276 262 387 14 88 44 46 4 system of 22.7 f 1 kJ mo1-l. The appropriate Arrhenius plot (of In z us.I / T ) is presented in fig. 3. Caesium Again precipitation of the radical anion restricted the temperature range (208-237 K) over which quality spectra were available for interpretation. However, it is again clear from these spectra that, although they are close to the ‘fast’ exchange limit, linewidth alternation is present. We have again computer-simulated these spectra allowing for174 -13 -14 -15 h c - -16 - 1 7 -18 1,4-Naphthoquinone E.S.R. Spectrum 4.0 4.5 5.0 10' KIT Fig. 3. Arrhenius plot for intramolecular cation migration in the 1,4-naphthoquinone radical anion, in 1,2-dimethoxyethane, with K+ as the counter-ion. Table 5. Hyperfine splitting constants (in pT) for the 1,4- naphthoquinone radical anion in 1,2-dimethoxyethane with Cs+ as the counter-ion, obtained by best-fit computer simulation, together with appropriate values of z T/K 42-H) 43-H) 45-H) 46-H) 47-H) 48-H) z/lO-' s 208 270 382 16 82 41 43 10 219 270 383 16 82 41 43 4 229 270 382 16 82 41 43 2 237 270 382 16 82 41 43 1 linewidth alternation, and the best-fit parameters required for each spectrum are summarised in table 5.Although only four data points are available we have estimated E, for intramolecular cation exchange to be 32 kJ mol-l. The experimental hyperfine splitting constants have again been assigned following the procedure described above. A comparison of the data summarised in table 1 indicates that the perturbation effects of the K+ and Cs+ ions are very similar and the same value of a, for the oxygen atom associated with the carbonyl group at position 1 seems appropriate for both of these coun ter-ions.Hyperfine Splitting Constants Table 1 shows that the nature of the alkali-metal counter-ion influences the degree of perturbation of the spin distribution in the 1,4-naphthoquinone radical anion, This can be clearly illustrated by noting the difference between 43-H) and 42-H) for each counter-ion. This difference changes with counter-ion in the sequence 270, 174, 125 and 112 pT for Li+, Na+, K+ and Cs+, respectively, while the average value of a(3-H) and a(2-H) remains virtually constant (at ca. 327pT). As expected the perturbation is greatest for Li+ and decreases with the charge density of the alkali-metal counter-ion.N . J. Flint and B.J . Tabner 175 Table 6. Activation parameters for intramolecular cation migration in the 1,4-naphthoquinone radical anion in 1,2-dimethoxyethane at 238 K ~ AG + Ea AH# A S counter-ion z/ lo-* s /kJ mol-1 /kJ mol-1 /kJ mol-l /J mol-l K-’ - - - - Li+ > 500 Na+ 55 29.3 9.0 7.1 -93.7 cs+ 1 21.4 32.0 30.1 36.5 K+ 17.5 27.1 22.7 20.8 -26.5 This perturbation has been reflected in molecular-orbital calculations by applying an auxiliary parameter to the carbonyl oxygen atom associated with the counter-ion. The best-fit values of this auxiliary parameter are 0.65 pee, 0.45 bCc and 0.30 BCc for Li+, Na+ and K+ (and Cs+), respectively, and the calculated values for the hyperfine splitting constants obtained using these values are collected together in table 1. Since the value of a(3-H)-a(2-H) is so dependent upon counter-ion we assume that we have a ‘close’ ion pair in each system.The value of a(3-H)-a(2-H) is also markedly temperature- dependent when Li+ is the counter-ion (see table 2), and there is some evidence for temperature dependence when K+ is the counter-ion (see table 4), but the temperature range covered is too small to detect a significant trend when either Na+ or Cs+ is the counter-ion (see tables 3 and 5). In the case of the Li+ system (and to some extent the K+ system also) the difference between 4 3 - H ) and 42-H) increases with temperature. These two effects would again be expected in a ‘close’ ion-pair situation with the separation between the radical anion and the counter-ion decreasing as temperature increases, the effect being more pronounced as the size of the counter-ion decreases.Activation Parameters The perturbation of the spin distribution in the 1,4-naphthoquinone radical anion increases with a decrease in the radius of the alkali-metal counter-ion. One would therefore expect Li+ to be more strongly associated with a particular carbonyl group than (say) Cs+, thus leading to a greater lifetime, at a particular site, for this counter-ion. A summary of the various activation parameters is presented in table 6, and clearly shows this to be the case with the lifetime of the cation at a particular site decreasing in the sequence Li+ > Na+ > K+ > Cs+. A positive AS# value might be anticipated in moving from the initial state to the final state, as a greater freedom would be associated with the latter.That is, we would anticipate that the counter-ion would be more loosely associated with the radical anion in the transition state, by contrast with its strong association with the carbonyl group in the initial state. A positive value of AS# is indeed observed when Cs+ is the counter-ion, but not when either Na+ or K+ is the counter-ion. We believe this reflects the change in the ‘solvation’ environment of the counter-ion. Li+ is strongly associated with the carbonyl group, with some solvent molecules also presumably involved in the ‘ solvation’ of the ion. However, in the transition state the counter-ion would be solely ‘solvated’ by solvent molecules, thus involving more solvent molecules in the transition-state ‘ solvation’ environment.This change in ‘ solvation’ of the counter-ion, which would be more pronounced for Li+ than for (say) Cs+, would lead to a marked negative contribution to AS#. The values of AS# for intramolecular cation migration in these systems, therefore, indicate the importance of counter-ion solvation in determining the overall value of AG#.176 1,4-Naphthoquinone E.S.R. Spectrum References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 E. de Boer and E. L. Mackor, J. Am, Chem. SOC., 1964, 86, 1513 and E. de Boer, Recl. Trav. Chim. Pays-Bas, 1965, 84, 609. M. Iwaizumi, M. Suzuki, T. Isobe and H. Azumi, Bull. Chem. SOC. Jpn, 1967,40, 2754. A. H. Reddoch, Chem. Phys.Lett., 1971, 10, 108. A. M. Hermann, A. Rembaum and W. R. Carper, J. Phys. Chem., 1967, 71,2661. A. H. Reddoch, J . Magn. Reson., 1974, 15, 75. M. Iwaizumi and J. R. Bolton, J. Magn. Reson., 1970, 2, 278. I. B. Goldberg and H. R. Crowe, J. Phys. Chem., 1976, 80, 2603. R. D. Rieke, S. E. Bales, P. M. Hudnall and C. F. Meares, J. Am. Chem. SOC., 1971, 93, 697. M. Iwaizumi and T. Isobe, Bull. Chem. SOC. Jpn, 1970,43, 3689. E. J. Rothwell and B. J. Tabner, J. Chem. SOC., Perkin Trans. 2, 1981, 1384. B. J. Tabner and T. Walker, J. Chem. SOC., Perkin Trans. 2, 1981, 1508. A. J. Burke and B. J. Tabner, J. Chem. SOC., Perkin Trans. 2, 1983, 205. N. J. Flint and B. J. Tabner, J. Chem. SOC., Perkin Trans. 2, 1986, in press. N. M. Atherton and A. E. Goggins, Mol. Phys., 1964, 8, 99.N. M. Atherton and A. E. Goggins, Trans. Faraday SOC., 1965, 61, 1399. N. M. Atherton and A. E. Goggins, Trans. Faraday SOC., 1966, 62, 1702. D. H. Chen, E. Warhurst and A. M. Wilde, Trans. Faraday SOC., 1967, 63, 2561. T. E. Gough and M. C. R. Symons, Trans. Faraday SOC., 1966, 62, 269. P. S. Gill and T. E. Gough, Can. J. Chem., 1967, 45, 21 12. P. S. Gill and T. E. Gough, Trans. Faraday SOC., 1968, 64, 1997. M. P. Eastman, G. V. Bruno, C. A. McGuyer, A. R. Gutierrez and J. M. Shannon, J. Phys. Chem., 1979,83, 2523. S . Konishi, S. Niizuma and H. Kokubun, Chem. Phys. Lett., 1980, 71, 164. N. J. Flint and B. J. Tabner, J. Chem. SOC., Perkin Trans. 2, 1984, 569. G. F. Pedulli and A. Alberti, Chem. Phys. Lett., 1977, 48, 72. F. Gerson, E. Heilbronner, H. A. Reddoch, D. H. Paskovich and N. C. Das, Helv. Chim. Acta, 1967, 50, 813. E. de Boer and A. P. Praat, Mol. Phys., 1964,8, 291. M. Iwaizumi and T. Isobe, Bull. Chem. SOC. Jpn, 1965,38, 1547. J. R. Bolton and A. Carrington, Mol. Phys., 1962, 5, 161. T. Takeshita and N. Hirota, J. Chem. Phys., 1969, 51, 2146. M. Guerra, G. F. Pedulli, M. Tiecco and G. Martelli, J. Chem. SOC., Perkin Trans. 2, 1974, 562. C. J. Leach and B. J. Tabner, J . Chem. SOC., Perkin Trans. 2, 1986, 337. C. J. Leach and B. J. Tabner, J. Chem. SOC., Perkin Trans. 2, in press. P. D. Sullivan and E. M. Menger, Adv. Magn. Reson., 1977, 9, 1. L. H. Piette, M. Okamura, G. P. Rabold, R. T. Ogata, R. E. Moore and P. J. Scheuer, J. Phys. Chem., 1967, 71, 29. J. M. Fritsch, S. V. Tatwawadi and R. N. Adams, J. Phys. Chem., 1967, 71, 338. G. R. Stevenson, A. E. Alegria and A. M. Block, J. Am. Chem. SOC., 1975, 97, 4859. D. Casson and B. J. Tabner, J. Chem. SOC. B, 1969, 572. B. J. Tabner and J. R. Zdysiewicz, J. Chem. SOC., Perkin Trans. 2, 1973, 81 1. R. J. Faber and G. K. Fraenkel, J. Chem. Phys., 1967,47,2462. P. H. Rieger and G. K. Fraenkel, J. Chem. Phys., 1962, 37, 281 1. A. J. L. Sevenster and B. J. Tabner, J. Chem. SOC., Perkin Trans. 2, 1981, 1148. C. J. Leach and B. J. Tabner, J. Chem. SOC., Perkin Trans. 2, 1985, 653. A. D. McLachlan, Mol. Phys., 1960, 3, 233. Paper 61851 ; Received 1st May, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300167

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1987,83, 177-190 Oxygen Diffusion-Concentration in Phospholipidic Model Membranes An Electron Spin Resonance-Saturation Study A. Vachon,? C. Lecomte and F. Berleur Commissariat a I’ Energie A tomique, IRDI/DESICP/D PCISCM CEN Saclay, 91191 Gif-sur- Yvette, France V. Roman* and M. Fatome Centre de Recherche du Service de Santk des Armkes, Division de Radiobiologie et Radioprotection, 92141 Clamart, France P. Braquet Institut Henri Beaufour, 17 Avenue Descartes, 92350 Le Plessis-Robinson, France Fully hydrated liposomes of dipalmitoyl-phosphatidylcholine have been labelled with 5 (or 7, 10, 12, 16)-doxy1 stearic acid at pH 6 and 8, and studied by the continuous-wave e.s.r.-saturation technique. The e.s.r. spectral magnitude depends on the hyperfrequency power P and on both and T, relaxation times.Saturation, i.e. the non-linearity of the spectral magnitude plotted us. dP can be quantified by a P+ parameter (power at which the signal is half as great as it would be without saturation). If we assume & is weakly modified by spin exchange between the paramagnetic spin probe and oxygen in its triplet state, Pi is inversely proportional to and becomes a sensitive parameter to appreciate the oxygen transport (oxygen diffusion- concentration product) inside the bilayers. According to the DPPC bilayer phase-transition diagrams, Pi (oxygen diffusion-concentration) is related to the thermodynamic state of the membrane. This technique provides further information on a particular property of a radioprotective agent, cystearnine, which seems to inhibit spin-triplet exchange and hence maximizes (minimizes PL).Since radioprotective agents are known to act by scavenging radiation-inkuced free radicals and by inhibiting oxygen-dependent free radical processes, such a result may contribute to the elucidation of radioprotecting mechanisms. The presence of molecular oxygen (0,) in biological molecules is a critical feature of a number of biochemical processes; for example: oxygen is consumed in mitochondria1 oxidative phosphorylations ;l oxygen-derived radical species are created during photosynthesis ;2 peroxide and hydroperoxide radicals might explain the membrane mechanism of photosensitizing drugs (p~rphyrins);~~ anthracyclins, a series of anti- tumour drugs, seem to present a cardiotoxicity through oxygen-activated spe~ies.~7 Among the environmental factors liable to modify the cellular and tissular radiotox- icity, oxygen has been the most studied.In irradiated cultured cells, the enhancement of radiotoxicity observed upon oxygenation is a basic phenomenon in radiobiology, but its interest can easily be extended to radioprotection and radiotherapy. The presence of various molecules such as thiols, histidine, methionine or cysteamine in the culture me- dium reduces this oxygen-dependent radiosensitivity and, thus, they can be considered as radioprotective agents. The in viuo action of the hydrophilic molecule cysteamine t Also at Institut Henri Beaufour, 17 Avenue Descartes, 92350 Le Plessis-Robinson, France. 177178 Phospholipidic Model Membranes @-mercapto-ethylamine), a standard among radioprotective sulphur-containing mol- ecules, has been attributed to a variety of mechanisms among which trapping radiation- induced and oxygen-dependent radicals was pointed A tentative explanation of part of this radioprotective effect is a specific interaction with lipidic membranesg> lo and a prevention of membrane peroxidation processesll by a mobilization of membrane- dissolved oxygen.l29 l3 Our purpose was to set up an appropriate methodology to study the oxygen-dependent effects of drugs on model membranes.Many experiments dealing with molecular oxygen are listed in the e.s.r. 1iterature:lT 14-16 Windrem and Plachy,17 and Subczinski and Hydel4? l8 have particularly studied the oxygen diffusion-concentration product in lipid bilayers by monitoring the intrinsic linewidth and the magnitude of the spectrum central line us.incident power P, of amphiphilic spin probes incorporated in lipidic model membranes. This methodology was developed in our experiments on fully hydrated liposomes of dipalmitoyl-phosphatidylcholine. The results might contribute to an attempt to elucidate the radioprotecting mechanism of cysteamine in biological molecules. Experiment a1 Reagents and Sample Preparation An aqueous liposome suspension (100 mg ~ m - ~ ) was prepared by vortexing in a 20 mmol dm-3 phosphate buffer m-a-dipalmitoylphosphatidylcholine (Sigma). The bulk pH was controlled after liposome preparation with an Ingold microelectrode and adjusted at pH 8 and pH 6 if necessary in order to experiment outside the pK, region of the stearic acid spin probe which is around 6.8 and 7.4 in DPPC bilayers, respectively, in gel phase and in liquid-crystal Indeed, temperature-dependent variations of these spin-probe spectra depend on the protonated or unprotonated form.Cysteamine hydrochloride was purchased from Merck. Solutions were freshly pre- pared before each experiment to prevent irreversible oxidation of cysteamine into its inactive disulphide dimer, cystamine (fig. 1). The cysteamine/DPPC molar ratios were 1 : l and 1:5. For runs without oxygen, the gas was removed from the liposomal preparation by bubbling argon through the suspension for 5 h before e.s.r. measurements were taken. The quartz cell was then kept under an argon atmosphere while the spectra were recorded.The spin label stearic acids were purchased from Aldrich and Interchim Chemicals. For e.s.r. experiments, 0.1 cm3 of liposomal suspension was supplemented with 5 x cm3 of a 10 mmol dm-3 stock solution of spin label in DMSO. The mixture was then vortexed above the main transition temperature to allow probe penetration. E .S . R. Measurements The spectra were recorded on a Varian E-109 spectrometer equipped with a Varian temperature controller. An additional temperature probe was introduced into the quartz cell, to take into account the heating of the sample, which becomes noticeable when the incident power exceeds 40 mW. For each spin label, the magnitude (h,) and the hyperfine splitting All were plotted us.the incident power, P, for every temperature. To simplify, in a Lorentzian lineshape, these parameters would have the following form in the first-derivative spectrum (fig. 2): fi Tl h, = k (1 +yT, T2P)$A. Vachon et al. HS- C H ~ - CH,- NH; . C I - 179 - + H ~ N - c H ~ - c H ~ - s - s - c H ~ - c H ~ - N H ~ + . 2 c i ~ r ) C,,H,,-C-0 -CH2 II O I C,, H3,- C-0 - CH ! 1 0- I It CH2-O-P-O-(CH2)2-N+(CH3)3 OH I 0' 0 2 ;"i (5 1 Fig. 1. Chemical formulae. (1) Cysteamine hydrochloride, (2) cystamine dihydrochloride, (3) DPPC, (4) tempo1 and (5) 12-doxylstearic acid (12-NS). where yTl T, P is the saturation term. The half-saturation power, defined as the power at which the signal is half as great as it would be without saturation (fig. 3), becomes: Pi is then representative of the variations of the relaxation times Tl and T,.The paramagnetic species 0, is detected by the variation of the 4 values assuming that Tl is modified by spin-exchange interaction with 0, inside the bilayer. According to Subczinski and Hyde:13 ( 3 ) 1 1 2 - (with 0,) = - (without 0 , ) + 3 p w e Tl Tl where p and we are, respectively, the efficiency and the frequency of the collisions. The frequency, we, is proportional to the product (Do + DN) No, where Do and D , are the180 Fig. Phospholipidic Model Membranes I-,2All \ .----- - /--- '-/ - 0.5 mT \ / '\ ,' 2. E.s.r. spectrum of DPPC probed with 5-NS spin label at pH 8, 21 "C and incident power of 2 mW. (-) Oxygen-containing liposomes, (- - - -) deoxygenated liposomes. linear weak strong zone saturation saturation -4 L / I I I 1 1 1 2 5 20 Pl/Z incident power, PImW Fig.3. Theoretical amplitude of the firstderivative e.s.r. signal us. 4P.A . Vachon et al. 181 o m , I I I I 1 2 5 20 40 80 120 200 300 400 Fig. 4. Magnitude of the central line tis. dP for oxygen-saturated liposomes of DPPC probed with 5-NS spin label. Temperatures (“C) shown on the figure. incident power, P/mW 0 ’ I 1 1 I I 2 5 40 120 200 Fig. 5. Magnitude of the central line us. z / P for deoxygenated lioposomes of DPPC probed with 5-NS spin label. Temperatures (“C) shown on the figure. incident power, P/mW5-NS 0-0 15 30 45 60 T/"C 7-NS 10-NS ' 0'0 15 30 45 60 15 30 45 60 T/'C 200 200 100 ..J- -_ 100 0 15 30 45 60 15 30 45 60 T/'C T/'C P+/mW 400 " 0 15 30 45 60 T/'C 12-NS 100 2- - - - -- 15 I 1 30 45 I 60 I 0 / Tl'C A- \- _ - _ - - - - l i 3'0 45 60 T/"C 4 /J ---\ ',/ - - 15 30 ' 45 ' 60 ' 0 7 k x - G T o T/"C T/'C Fig. 6.Evolution of the Pi parameter us. temperature for 5-, 7-, lo-, 12- and 16-NS spin labels upon oxygenation or without. (-) Oxygen-containing liposomes; (- - - -) deoxygenated liposomes. (A) DPPC, pH 6; (B) DPPC, pH 8.A . Vachon et al. 183 diffusion coefficients for the oxygen and the nitroxide, respectively, and No is the number of oxygen molecules per unit volume, Neglecting D , we obtain: 1 1 - (with 0,) = - (without 0,) + ADo No. Tl TI We assume that 0, has but a weak influence on T2, then: (4) DON0 ( 5 ) 1 1 (with 0,) = - (without O,)+A ~ Tl T2 TI T2 T2 Do No Pi (with 0,) = Pi (without 0,) + A T . '2 This relation is valid for a single Lorentzian line; for a composite anisotropic spectrum, the magnitude of the central line is a function of the power and the probe mobility (correlated with the A s litting).The difference: [Pi (with 0,) -Pi (without O,)] at a given temperature wil still be representative of the oxygen diffusion+oncentration product Do No; this will be discussed in more detail below. For 12-doxy1 stearic acid spin label, there is a measurable amount of non-incorporated spin label, exhibiting an isotropic-like spectrum. The parameter Pl observed for the non-incorporated spin label was estimated by plotting the magnitude of the high-field line (rn = - l), and by comparing it to the Pi values of the hydrophilic radical TEMPOL in the medium. i ' ? Differential Scanning Calorimetry (D .S . C . ) D.s.c. experiments were carried out on a Perkin-Elmer DSC-4 scanner. Scans were performed between 5 and 60 "C at a heating rate of 10 K min-l. The peak characteristics are designated by their onset and maximum temperature. The heats involved are related to the peak area. Results Effect of 0, on the Magnitude of the Central Line Fig. 4 and fig. 5 present the set of curves obtained with 5-doxy1 stearate from the magnitude of the central line as a function of the incident power and at different temperatures, with or without oxygen. Note that the presence of oxygen markedly increases the maximum value of each curve with only a minor modification of the initial slope (hO/2/P); this fact is further quantified by the Pi parameter.Comparative Pi values are plotted us. temperature for the different (5-16) doxyl stearates, upon oxygenation or not and at both pH 6 and 8 (fig. 6). If main transitions are clearly displayed with 5 and 16-NS, problems occur with 7 to 12-NS which do not actually probe the bilayer regions as they would be expected to do, particularly in the unprotonated form. If we consider the doxyl group rather hydrophilic, the conjunction of a hydrophilic COT at one end of the stearic acid (C-18), and of a rather hydrophilic doxyl group located in the mid-part of the acyl chain, renders this molecule unliable to insert parallel between the phospholipidic hydrocarbon chains. This could be explained by the fact that the two hydrophobic carbon chains remaining free on both sides of the doxyl group are not long enough to confer a sufficient hydrophobic property to the whole spin probe.In these conditions the doxyl group might cross the lipid-water interface, within a wide location distribution. Spectra would then contain artifacts which must be taken into account.184 0 Phospholipidic Model Membranes 4 00 \ % qb 300 CI s '3 200 g a E: Y 5 2 100 I ' ro 15 30 45 60 T/'C Fig. 7. Half-saturation power, Pi, as a function of temperature for DPPC probed with 5-NS spin label at pH 8. (-O-) Oxygen-containing liposomes, (- -0- -) deoxygenated liposomes, (-a-) oxygenation plus cysteamine-DPPC (1 : l), (-A--) oxygenation plus cysteamine- DPPC (1 : 5). Fig. 8. 0.00 I I 1 10.00 20.00 30.00 40.00 50.00 T/"C Thermogram of fully hydrated DPPC (2.17 g): (a) maximum 37.9 "C, (b) maximum 43.6 "C.Scan rate 10 "C min-'.A . Vachon et al. 185 tempo 1 / h \ - 0.5 mT ii \ i.i Fig. 9. 5-NS spectrum in DPPC bilayers plus superimposed ternpol (5 x mmol dm-3) isotropic spectrum, at 25 "C and pH 8. Spectral Lineshapes obtained with or without Oxygen Spectra were recorded (fig. 2) under the same conditions, in the presence or absence of oxygen. Oxygen does not affect the measured values of the hyperfine splitting (2AIl) whatever the temperature, as compared with the oxygen-depleted suspension. Thus, we can assume that oxygen does not modify the phase-behaviour of the DPPC membranes. The Pi Parameter The Pi parameter was estimated graphically and is reported in fig. 7. We notice that the value of Pi is systematically enhanced in the presence of oxygen and exhibits two sharp maxima at 38 and 41 "C, which correlate, respectively, with the ' pre-transition' (T,) and the 'main transition' (T,,) temperatures of the phase-transition diagram, as obtained from d.s.c.(fig. 8). The peak observed at the main transition temperature seems like a discontinuity (critical temperature), as far as the values obtained in the 40-43 "C temperature range were essentially not reproducible. Saturation of the Hydrophilic Spin Label Ternpol The saturation phenomenon was studied with ternpol in the usual phosphate buffer in the presence or absence of 5-doxy1 stearic labelled membrane. The measurements were made on the high-field line of the ternpol spectrum (fig. 9). We obtained for Pi (02, tempol) a constant value of ca.45 mW, irrespective of phase transitions and of the presence of 5-doxy1 stearic acid in the membrane (fig. 10). Saturation of the Non-incorporated 12-NS Spin Label Below the ' pre-transition' temperature of DPPC bilayers (36 "C), 12-doxy1 stearic acid exhibits a composite spectrum which is the sum of two spectra: an 'immobilized' component, due to part of spin probes deeply incorporated in the membrane, and an isotropic spectrum attributed to non-incorporated probes (fig. 1 1). Pi measurements were undertaken on this isotropic component in order to determine the location of the probe relative to the lipid-water interface. As measured from the high-field isotropic line, the186 Ph osp holip idic Model Membranes 15 30 4'5 60 TI0C Fig.10. Plot of the half-saturation incident power, 4, as measured from the magnitude of the high-field line (m = - 1). (-*-) Ternpol in aqueous solution, (--.--) non-incorporated part of 12-NS label at pH 8 (see fig. 11). not incorporated A/\ c--r 0.5 mT Fig. 11. 12-NS spectrum in DPPC bilayers, at 14 "C and pH 8, in the presence of oxygen. result shows a continuous decrease of PI? us. temperature up to 36 "C. Beyond this value, the isotropic component is no longer observable in the experimental spectra. Effect of the Radioprotective Agent Cystearnine Since this is thought to interfere in vitro with oxygen radicals in biological membranes, its influence upon Pi was tested (fig. 7). For an equimolar ratio (cysteamine/DPPC = 1 : l), the membrane behaves as if it was deoxygenated: the Pi parameter presents an almost constant low value and does not undergo any sharp transition related to the temperature. Simultaneously, the 2A hyperfine splitting is not modified.For a lower ratio (cysteamine/DPPC = 1 : 5 ) , an intermediate Pi value is observed. Discussion Correlation between Pi and the Thermograms The Pi parameter depends on the averaging of the anisotropic component of the g and tensors, and on the local diffusion-concentration D,[O,] product.A . Vachon et al. 2- 187 n 0 15 30 4-5 60 T/" C Fig. 12. Plot of the slope (h,/z/P)k in the linear region as obtained from 5-NS spin label in DPPC at pH 8. (--m-) Oxygencontaining liposome, (--O- -) deoxygenated liposomes, (-O-) ratio of both slopes (with O,/without 02). If the line broadening is assumed to be isotropic, then relation (6) is valid and the variations of Pi represent the variations of the Do [O,] product.This product is maximal at the main transition temperature (in agreement with other reported data in the literature),22 suggesting that at the main transition the membrane is in a minimal organization state. Transport and diffusion properties are then enhanced. We can remark, and this is the rule in W E technics, that Tp (d.s.c.) and Tp (RPE) do not coincide: this phenomenon can be explained by the fact that Tp is a second-order transition.21 Effect of Oxygen on the Relaxation Times Tl and T, According to relation (l), the initial slope h,/dP is proportional to Ti and is a function of anisotropy. (h,/.\/P)i is presented in fig.12 as a function of the temperature in the presence or absence of oxygen; the proportionality coefficient is then a function of the number of spins and of the membrane organization. The transition observed at the intermediate phase, between 30 and 40 "C, may be explained by the variation in the angular distribution of nitroxidep,-axes inside the lipidic matrix, which is described by the anisotropy coefficient, and the motional narrowing of the lines. Until now we have assumed that Tl is decreased by Heisenberg spin-exchange interaction and that T, is not modified in the presence of oxygen for a given temperature. The results presented in fig. 12 are consistent with a T, decrease of ca. 10% in the I iquid-cry s t a1 phase .188 Phospholipidic Model Membranes NO'/O, Interaction Subczinski and Hyde14 and Kusumi et al.18 observed no (or little) discontinuity in their Pi parameter under low oxygen pressure in the transition temperature range.As they suggested in their discussion, this result could be explained by an equal and opposite change in TI and T,, resulting in a continuity of Tl T, and then in Pi. This result (fig. 12) can be extended, as described in fig. 6, to 5, 7, 10, 12 or 16 nitroxide stearic spin label, as well as androstane and cholestane spin labels.18 According to the diversity of the membrane domains explored by these probes, this assumption could be extended as a general rule. There is no clear consensus in the literature dealing with the interaction between oxygen molecules and the nitroxide species.The commonly accepted hypothesis is an exchange by Heisenberg spin-interaction.l* If oxygen diffusion in phospholipidic membranes can be described by an axially symmetric tensor with its long axis parallel to the normal of the bilayer plane, this would render the spin exchange with the label more effective along the z (vertical) axis than by an approach along x or y axes. Indeed, the nitroxide radical is protected by the sort of cage constituted by the four methyl groups and which is the very reason of its stability. Conversely, the pz orbital remains open to any interaction along the z-axis. Influence of Cystearnine By analysing the PI parameter, the cysteamine molecule has a 'quenching effect' on membrane dissolved oxygen; Pi with cysteamine has a lower value, as if the membrane was deoxygenated.Moreover, this oxygen concealment remains stable over periods of several hours and this fact rules out the hypothesis of a direct complexation (cysteamine- oxygen), as such a complexation would be followed by a reoxygenation of the medium. Although previous studies suggest that cysteamine interacts preferentially at the lipid-water interface in the ~ / 3 phase, recent experiments have indicated that in the LCI phase, cysteamine penetrates the lipidic structure, its long axis being parallel to the acyl chains and the thiol group directed toward the bilayer central core.9 In the oxygen preferential diffusion conditions, one might expect steric hindrance induced by cystearnine, which would hinder z-axis oxygen diffusion and lessen or abolish z-axis spin exchange with the p z orbital.We can indeed observe on fig. 7 that for lower temperatures (below the pre-transition) PI values are not very different for the control and cysteamine-supplemented suspension. Ths indicates that cysteamine has not yet penetrated the bilayer and remains adsorbed (by electrostatic interactiong) at the water-DPPC interface. Penetration starts at the pre-transition temperature, as defects occur in the phospholipidic organization. Differ- ence is patent at higher temperatures (over 42 "C, in the LCI phase) and results in oxygen diffusion, in agreement with our previous report. Biological Implications of the Results As briefly reported in introduction, an obvious way of decreasing the radiosensitivity of an animal is to take advantage of the oxygen effect by administering drugs which could produce anoxia or severe hypoxia of radiosensitive tissues.The presence of molecular oxygen (i.e. 150 mmHgt) increases by a factor of two to three most of the effects of X or y-irradiation on living and many non-living systems. There are a number of compounds that protect by virtue of their capacity to decrease intracellular oxygen tension, by interfering with oxygen transport in the body, among which were observed f 1 mmHg w 133.3 Pa.A . Vachon et al. 189 respiratory depression, circulatory disturbances or inhibition of the haemoglobin transport mechanism. But if this is the case for sulphydryl protectors (cysteamine, cysteine), they (paradoxically), according to the literature, do not regularly cause tissular anoxia (often 0, tension is raised).Unfortunately all 0, tension measurements were performed in interstitial plasma and not inside the cell. Moreover, in the rat, raising 0, pressure (up to 5 atmt) in the respired air during irradiation does not affect (or only reduces slightly) the radioprotection afforded by cysteamine. Apart from this complex problem, many authors have been impressed by the fact that radioprotectors can affect markedly some fundamental processes in the machinery of the ell.^^ To preserve the reader from a fastidious catalogue, we can summarize most of the cellular disturbances from biochemical processes to cell organites morphological alter- ations (mitochondria) by indicating that all these experimental data (since 1950) seem to converge toward the concept of cell suffering or adjustment, related to an intracellular oxygen depletion.This concept would receive some justification if we examine the results presented in this paper. The preferential z-axis oxygen diffusion is undoubtedly responsible for the oxygen diffusion from haemoglobin through the red cell membrane and tissular cell membrane, down to oxygen-consuming chemical processes (mitochondria). If, by any means, the cellular membrane could become much less permeable to oxygen, such cellular disturbances would be liable to occur. Conversely, a membrane impermeability would not affect extracellular (interstitial) oxygen tension. But if cysteamine (a small thiol molecule) hinders oxygen diffusion, why not larger and hydrophobic molecules? In the liquid-crystalline phase (biological membranes do not present phase transitions at physiological temperatures and can be considered as a fluid phase) aliphatic chains' trans-gauche conformations could induce ' kink and jog' axial defects23 responsible for this promoted vertical diffusion.Moreover, in biological membranes, these defects, dynamic cavities, must be enhanced by the presence of the rigid double bonds generally located around C-10. The penetration of cysteamine in the La phase would take place preferentially inside these cavities, as far as the molecular dimensions are consistent (0.5 nm), obstructing by the same the oxygen diffusion pathway. After administration, the whole-body impregnation by cysteamine during its short action period (30 min) would, as a consequence, lessen or to some extent abolish oxygen transport in all the cellular membranes.Can this fact be considered as the exclusive mechanism of radioprotection? It is too soon at this stage to give a reasonable response if we take into account the other suspected mechanisms, especially free-radical trapping and DNA protection. All the same, it is striking to note that under these conditions general hypoxia would be the price to pay to ensure radioprotection and it is to be feared that expecting a dose-reducing factor over 3 becomes unreasonable. Nevertheless, it is interesting to remark that a new methodology developed for the first time by Hyde can be successfully applied in biology and is liable to bring new lights on the mechanism of chemical radioprotection about which, up to now, opinions were rather divergent.'The authors wish to thank particularly Dr Folcher for fruitful help and remarks during the development of this work, and Prof. Bienvenue for helpful suggestions. Part of the experimental work was supported by grants from Institut Henri Beaufour, from Ipsen Laboratories and from Direction des Recherches, Etudes et Techniques. References I J. M. Backer, V. G. Budker, S. I. Eremenko and Y. N. Molin, Biochim. Biophys. Acta,, 1977,460, 152. 2 R. V. Bensasson, E. J. Land, A. L. Moore, R. L. Crouch, G. Dirks, T. A. Moore and D. Guts, Nature 3 C. S . Foots, Acc. Chem. Res., 1968, 1, 104. t 1 atm = 101 325 Pa. (London), 1981, 290, 104.190 Phospholipidic Model Membranes 4 C.Emiliani and M. Delmelle, Photochem. Photobiol., 1983, 37, 487. 5 C. E. Myers, W. P. McGuire, R. H. Liss, I. Ifrim, K. Grorzinger and R. C. Young, Science, 1977,197, 6 A. Meister, Science, 1983, 220, 472. 7 W. 0. Foye, Znt. J. Sulfur Chem., 1973,8, 161. 8 Z. M. Bacq, in Chemical Protection Against Ionizing Radiations (Charles C. Thomas, 1965). 9 F. Berleur, V. Roman, D. Jaskierowicz, M. Fatome, F. Leterrier, L. Ter-Minassian-Saraga and 165. G. Madelmont, Biochem. Pharmacol., 1985, 34, 3071. 10 V. Roman, F. Bocquier, F. Leterrier and M. Fatome, C.R. Acad. Sci. Paris, 1982, 295, 191. 1 1 G. Gregordiadis, in Liposome Technology (CRC Press, Boca Raton, Florida, 1984), vol. I, chap. 1 1 . 12 G. G. Jayson, T. C. Owen and A. C. Wilbraham, J. Chem. Soc., 1969, 3, 944. 13 L. Novak, Bull. Acad. R . Belg. U. Sci., 1966, 52, 633. 14 W. K. Subczinski and J. S. Hydes, Biochim. Biophys. Acia, 1981, 643, 283. 15 M. J. Povich, Ann. Chem., 1975, 47, 346. 16 M. P. Eastman, R. G. Kooser, M. P. Das and J. H. Freed, J. Chem. Phys., 1969, 51, 2690. 17 D. A. Windrem and W. Z. Plachy, Biochim. Biophys. Acta, 1980, 600, 655. 18 A. Kusumi, W. K. Subczinski and J. S. Hyde, Proc. Nail Acad. Sci. USA, 1982,79, 1854. 19 A. Samson, M. Ptak, J. L. Rigaud and C. M. Gary-Bobo, Chim. Phys. Lipids, 1976, 17, 435. 20 P. Bonnet, C. Lecomte, E. Sponton, C. Roumestand, V. Roman, M. Fatome and F. Berleur, unpub- 21 F. Berleur, V. Roman, D. Jaskierowicz, F. Leterrier, F. Esanu, P. Braquet, L. Ter-Minassian-Saraga 22 F. Berleur, V. Roman, L. Ter-Minassian-Saraga, F. Leterrier and M. Fatome, unpublished results. 23 J. Day and C. R. Willis, J. Theor. Biol., 1982, 94, 367. lished results. and G. Madelmont, Biochem. Pharmacol., 1984,33, 2407. Paper 6/ 12 16; Received 16th June, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300177

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1987, 83, 191-202 Use of Nitroxides to Measure Redox Metabolism in Cells and Tissues Harold M. Swartz University of Illinois, College of Medicine, 506 South Mathews, Urbana, Illinois 61801, U.S.A. The metabolic and physical interactions of the paramagnetic nitroxides with cells and tissues are becoming effective tools for measuring biologically significant processes, including some not measured adequately by other means. We have exploited the physical (magnetic) interactions of oxygen with nitroxides to obtain the first reliable and facile measurements of intra- cellular oxygen, using the effects of oxygen on the relaxation times of nitroxides via Heisenberg exchange. With this technique we have demon- strated a significant gradient between extracellular and intracellular oxygen concentrations in our experimental cell suspension system.Previously, the existence of such a gradient has been debated hotly but not tested directly. We also have determined some of the principles of interactions between nitroxides and cells and have demonstrated the feasibility of using metabolic interactions of nitroxides with cells to measure hypoxia in vivo. The principal metabolism of nitroxides by cells is reversible reduction to the hydroxyl- amine. The rate of reduction depends on the physical characteristics of the nitroxides. Reduction occurs primarily in the intracellular compartment and therefore only nitroxides that can cross the cell membrane readily (e.g. small molecules that are lipid-soluble) can be reduced readily by cells.Once nitroxides get into cells, the rate of reduction depends on the structure of the nitroxide (e.g. those with the five-membered pyrrolidine ring reduce more slowly than those with the six-membered piperidine ring). For some nitroxides the rate of reduction is up to thirty times faster in severely hypoxic cells. This latter phenomenon makes feasible the use of nitroxides to detect hypoxic areas in vivo by using the nitroxides as ‘contrast agents’ for in vivo n.m.r. studies. In principle, regions of the body with hypoxic areas will have lower concentrations of nitroxides because in these regions the nitroxides become reduced to the non-paramagnetic hydroxylamines which do not affect n.m.r. images (paramagnetic molecules affect n.m.r. images by shor- tening the relaxation times of water protons - the relaxation times of water are the principal imaging parameters for most current in vivo n.m.r.techniques). Thus it appears feasible to use nitroxides to detect and follow processes in vivo that are associated with hypoxia; these include cancer, ischemia (i.e. drastically reduced blood flow) and inflammation. - The paramagnetic nitroxides have been used extensively as biophysical probes, especially to study various types of motion in membranes and macromolecules.l* In addition they have been employed to study a number of other parameters of cells including ~olume,~ transmembrane potential^,^ oxygen c~ncentration,~ surface potential6 and P H . ~ The range of systems to which they have been applied varies from simple solutions to intact animals.The use of the nitroxides in more complex biological systems such as living cells often provides data on motion and other parameters that cannot be obtained as con- veniently and/or as accurately by other means. Recently a major new use of nitroxides in biological systems has been proposed: their use as contrast agents for in vivo n.m.r. studies. 7 191 FAR 1192 Use of Nitroxides to Measure Redox Metabolism Fe (C NIi- PDT PDT PDT Fe (C N 1;- Fe (CNI ;- P DT Fig. 1. Method of measuring intracellular oxygen. (A) Schematic representation of cell, nitroxides (PDT) and broadening agents (ferricyanide). The broadening agent effectively eliminates the e.s.r. spectra of extracellular nitroxide. (B) E.s.r. spectra used to measure intracellular oxygen : spectra of 0.10 mmol dm-3 PDT (chemical formula shown) in the presence (a) and absence (b) of air.The peak-to-peak linewidths are measured as shown by the small arrows. The difference, A W,, changes linearly with oxygen concentration. Calibrations are made in the media in which the oxygen concentration is measured.H. M . Swartz 193 Fig. 2. Method of measuring extracellular oxygen. Spectra of midfield line of CAT, (chemical formula shown) in the presence (lower spectrum) and absence (upper spectrum) of air. An empirical parameter for the oxygen concentration, C, is calculated as ( A + B)/h; in the range of oxygen concentrations of interest, C varies linearly with oxygen concentration; C is calibrated for the medium in which the measurements are made.The uses of nitroxides in complex biological systems, however, has led to concern about undesired interactions between nitroxides and cells. The metabolism of nitroxides by cells has particularly drawn attention as it has been observed that cells often reduce nitroxides to non-paramagnetic hydroxylamines. Although this loss of the e.s.r. signal usually is considered an undesirable effect, our laboratory has begun a series of studies in which we exploit the metabolism of nitroxides by cells to gain additional information from living biological systems.8-10 Similarly, the effect of parmagnetic oxygen to broaden e.s.r. spectra is often considered an undesirable source of spectral distortion in biological samples, but this phenomenon can be utilized to measure the concentration of this physiologically important s~bstance.~ The accurate measurement of hypoxia inside cells and in tissues would provide very important and useful information relevant to both basic studies and clinical applications.Many fundamental physiological and biochemical effects are dependent on the con- centration of oxygen. Many major categories of disease, such as cancer, ischemia and inflammation, characteristically involve hypoxia, therefore the ability to detect and follow hypoxic areas in uivo could provide a major advance in our ability to detect diseases in patients and to assess accurately the efforts of treatment on disease processes. In this paper I summarize some of our ongoing studies on the physical and metabolic interactions of nitroxides and cells.The studies have involved a number of colleagues and students who are listed in the acknowledgements. Some of the work has been completed and published but the major effects are clearly still before us. In addition to the data needed on the two specific problems which are described in this paper, there also is a need to obtain more basic knowledge about the location and interactions of nitroxides with cells and various types of tissues in living organisms. The existing literature on this subject is fragmentary and often contradictory. Studies in carefully chosen and controlled model systems are needed as are suitable studies in living cells and animals. We have progressed sufficiently, however, to feel confident that it is possible 7-2194 Use of Nitroxides to Measure Redox Metabolism % air in equilibrating gas 0 20 40 60 80 100 l i i l l i l i i i l l 160 1 c 0 1 ~ 0 40 80 120 160 200 [ O2 ] in perfusing gaslpmol dm-3 Fig.3. The steady-state concentration of oxygen in the intracellular and extracellular compart- ments of a suspension of mammalian cells. The intra- and extra-cellular oxygen concentrations are plotted against the oxygen concentration of the gas that is perfused around a gas-permeable Teflon tube; the perfused oxygen concentration is expressed as the concentration measured in a solution that is in full equilibrium with the perfused gas. Each sample contained lo7 cells (from a cell line derived from mouse thymus-bone marrow)* in 100 mm3. For measurement of intracellular oxygen the suspension contained 0.10 mmol dm-3 PDT and 50 mmol dm-3 potassium ferricyanide.For measurement of extracellular okygen the solution contained 0.15 mmol dm-3 CAT,. Measure- ments were made at 37 "C. to obtain useful data on the concentration of oxygen inside cells in suspension and that it is feasible to use the metabolic interactions of nitroxides to obtain in vivo nuclear magnetic resonance images that reflect tissue metabolism. Our methodological and conceptual approaches, results to date, and some future plans on both of these interrelated areas are summarized in the balance of this paper. Measurement of Intracellular Oxygen Although the concentration of oxygen within cells is a crucial variable in physiological, therapeutic and pathological processes, there are few data available on actual intra- cellular oxygen concentrations.Current knowledge of intracellular concentrations is based upon oxygen determinations involving either inference from extracellular measurementsll or direct measurement of intracellular concentrations via invasive12 or technically difficult methodologies. 13-15 Using the physical effects of oxygen on the electron spin resonance (e.s.r.) spectra of nitroxides, our laboratory has directly and non-invasively measured intracellular oxygen concentrations in mammalian cells in l6 We use a variation of the same technique to measure extracellular oxygen in the same suspension. Interactions between molecular oxygen and nitroxide free radicals broaden e.s.r. signals via Heisenberg spin-exchange. 1 7 7 l8 This oxygen-induced broadening has pre- viously been utilized to measure oxygen concentrations in biological solutions [e.g.ref. (19) and (20)] and extracellular media [e.g. ref. (21) and (22)]. Hyde and coworkers have had an especially large and successful series of such studies [e.g. ref. (23) and refer- ences therein].H. M . Swartz 195 'To measure intracellular oxygen, we use (1) a nitroxide (PDT, 4-0x0-2,2,6,6- tetramethylpiperidine-d,,- 1 -0xyl) which rapidly equilibrates across cell membranes and whose linewidth is very sensitive to oxygen concentration, and (2) a paramagnetic broadening agent [ferricyanide, Fe(CN)i-], which is membrane-impermeant and remains extracell~lar,~~ thereby broadening the e.s.r. signal due to extracellular PDT. This method allows the selective visualization of the e.s.r.signal from intracellular nitroxides, whose linewidths therefore reflect only the intracellular oxygen concentration (fig. 1). Measurements of extracellular oxygen in suspensions of metabolizing cells were made (in the absence of ferricyanide) with the membrane-impermeant nitroxide CAT, (4- t rime t hylammonium-2,2,6,6- tetramet hylpiperidine- 1 -oxyl iodide ; Molecular Probes) at a concentration of 0.15 mmol dm-3 (fig. 2). E.s.r. spectra of this nitroxide show resolvable superhyperfine structure which is broadened in the presence of oxygen and can be quantitated in terms of an empirical parameter, C, described previously5 and shown in fig. 2. We chose CAT, for these extracellular measurements because its charge prevents it from entering the cell membrane and therefore its lineshape reflects ex t racellular oxygen.Fig. 3 summarizes a set of studies under steady-state conditions in which we determined intracellular and extracellular oxygen concentrations, as a function of the oxygen content of the equilibrating gas. These concentrations differed significantly from each other and from that of the perfusing gas. When the oxygen content of the perfusing gas is < 40 pmol dm-3, the rate of oxygen diffusion across the walls of the Teflon sample tube and into the sample is not sufficient to overcome cellular oxygen consumption and the extracellular and intracellular oxygen concentrations are extremely low. From 40 to ca. 80 pmol dm-3 oxygen in the perfusing gas, the extracellular concentration increases faster than the intracellular concentration and an oxygen gradient of increasing magnitude results.Above 80 pmol dm-3 there appears to be a constant difference between intracellular and extracellular oxygen concentrations. The presence of an oxygen gradient of this magnitude has important implications for the study of cellular processes which are dependent on oxygen concentration. Analysis of results in terms of conventional approaches which measure only extracellular oxygen or oxygen utilization could lead to misleading conclusions, as illustrated later in this paper (fig. 5). The demonstration of a significant difference in average intracellular oxygen con- centration (with respect to the concentration of oxygen in the extracellular media) is a surprising phenomenon inasmuch as the observed difference is greater than would be expected in a simple model of oxygen utilization at a point, with free diffusion of oxygen.Therefore some of our future plans include: (1) demonstration of the gradient by a second method (based on the microwave power saturation of the saturable stable free-radical melanin, which can be introduced readily into cells); (2) development of probes that will report on oxygen concentrations at specific locations in the cell (e.g. in membranes, in cytoplasm and in the nucleus); (3) development of nitroxides that will readily enter cells and then localize exclusively within the cell (we hope to do this with a neutral nitroxide that becomes converted intracellularly into a charged species, in a manner analogous to that of glucose which, upon phosphorylation, does not leave the cells); and (4) improvement of instrumental sensitivity so that we can measure smaller differences in oxygen concentration inside cells and/or make measurements on smaller samples (developments by Hyde and coworkers using loop-gap may be especially useful in this regard).s n Table 1.Relative rates of reduction of nitroxides by mammalian cells (TB cells) ~- typea of air N, ratio N, after ratio (N,) 3 ni troxide nameb (intact) (intact ) N,/air freeze-thaw freeze-thaw/intact 3 uncharged, Tempone (1) 2.34 3.17 1.35 3.55 1.12 hydrophilic Tempo1 (2) 2.73 2.63 0.96 4.28 1.63 c" uncharged, 5-D.S. (8) rn = 12,n = 3 2.12 29.48 hydrophobic 12-D.S. (9) rn = 5 , n = 10 1.92 3.79 partially PCA (3) ionized 5-Tempamine (4) 6-Tempamine (5) 0.14 0.46 3.74 0.67 1.12 6.19 12.5 2.0 4.8 2.4 1.66 1 S O 1.62 5.88 h 2.24 5 R 3 1.45 0.95 0.18 0.42 2.3 6.01 14.3 x Ei $ s 1.06 3.84 11.6 charged CAT, (6) ZWIT (7) 0.3 1 0.33 ~- cs- 0 % a At ca.pH 7.4. Structural formulae are given in scheme 1 .H. M . Swartz 197 6 *I 0 0 (6) $5 *I 0 (5) ( 8 , 9 ) Scheme 1. Feasibility of Use of Nitroxides as N.M.R. Contrast Agents that Reflect Metabolism: Demonstration of Responsiveness to Hypoxia The use of contrast agents in magnetic resonance imaging (m.r.i.) is rapidly developing, clinical trials are underway and some regular clinical uses seem imminent.25 The principal thrust has been to develop contrast agents whose differential tissue uptake and passive distribution provides the desired contrast.The nature of n.m.r. and n.m.r. contrast agents makes possible another, complementary approach to contrast in which the amount of the active contrast agent reflects metabolic and pathophysiological processes. This approach is possible because the usual n.m.r. contrast agents are based on their paramagnetism, and the presence of paramagnetism can be modified by redox reactions. Molecules are paramagnetic because they possess an unpaired electron. Paramagnetic molecules can be converted into non-paramagnetic states (and hence lose their n.m.r. contrast properties) by the loss (oxidation) or addition (reduction) of an electron. Physiological and pathophysiological metabolic processes can cause such oxidations and reductions and thereby provide differential contrast on the basis of metabolism.The nitroxides have been considered previously as n.m.r. contrast agents and have been198 Use of Nitroxides to Measure Redox Metabolism Table 2. Relaxivities, in dm3 mmol-l s-l, for nitroxides in albumin at various n.m.r. field strengthsa field strength (Larmor frequency of protons in MHz) ni troxide 0.01 0.1 1 10 50 Tempone 1.4 1.3 1.2 0.7 0.4 PCA 2.3 2.2 1.9 1.1 0.6 5-doxy1 stearate 6.8 6.4 5.2 2.8 1.1 7-doxy1 stearate 7.6 6.7 5.7 3.1 1.3 10-doxy1 stearate 8.8 8.0 7.0 4.6 1.9 12-doxy1 stearate 9.0 8.2 7.1 4.8 1.6 16-doxy1 stearate 7.4 7.1 5.2 3.0 1.6 a For methodology see ref. (9); values for doxyl stearates are ca. f 20 % ; values for Tempone and PCA are ca. & 5 % .Albumin concentrations ranged from 0.2 to 1 mmol dm-3 in 50 mmol dm-3 phosphate buffer (pH 7.4). Nitroxide : albumin ratios were varied from 1:2 to 10: 1 for Tempone and PCA with no change in relaxivity. For the doxyl stearates, a 1 : 1 nitroxide: albumin ratio was used, except for 7-doxy1 stearate, for which the nitroxide: albumin ratio was varied from 1: 1 to 6: 1 with no change in relaxivity . studied in vivo and in vitro for this use.26-28 They have produced demonstrable image enhancement in the brain and kidney. Concern about their clinical utility has centred on their inactivation in vivo and on their capacity for relaxivity. The concept of using differential rates of inactivation in vivo to reflect areas of different metabolism has the potential of converting an ostensible drawback into enhanced capabilities of in vivo n.m.r.techniques. To demonstrate the feasibility of this approach one needs to show (1) that nitroxides can be effective relaxers in tissues, (2) that their concentration can be dependent on metabolic processes of interest and (3) the properties of good relaxivity and good metabolic responsiveness are mutually compatible. Our experimental approach to determine the potential effectiveness of the nitroxides as contrast agents is to measure their effect on the relaxation rates of protons over a range of fields from 0-1.2 T; those experiments have been reported el~ewhere.~ Our experimental approach to determine the metabolic responsiveness of nitroxides is to measure their concentration in cellular suspensions, using e.s.r.spectroscopy. The combination of the determinations of relaxivity and cellular metabolism of a number of different nitroxides provide the data needed to evaluate the feasibility of the concept of developing metabolically responsive contrast agents and indicate the physical-chemical properties required for optimally effective contrast agents of this type. The methodology for our approach is reported elsewhere8-lo and is summarized briefly in the figure legends. Table 1 and scheme 1 indicate the structures of some of the nitroxides used in these experiments and their initial rates of reduction by TB cells. Note that for some of the nitroxides the rate of reduction by cells is much faster if the cells are deficient in oxygen. This is a key element in the feasibility of having nitroxides reflect metabolic differences. The data in table 1 also indicate that in this experimental system the nitroxides need to get into the cell in order to be reduced rapidly; e.g.CAT, is reduced rapidly only if the cell membranes are damaged by freeze-thawing. The data also indicate that intracellular location is a necessary but not sufficient cause for rapid reduction of the nitroxides. EvenA B \I V Fig. 4. Demonstration of reduction and reoxidation of a nitroxide by mammalian cells in suspension. The sample contained lo7 TB cells and 2 nmol of 5-doxystearate in 100 mm3 in a gas-permeable tube. (A) The initial e.s.r. spectrum of the nitroxide (b) and the time course (ca. 25 min) of the magnitude of the highest peak (a) are indicated.Air initially was flowed and at the point indicated by the arrow the gas was changed to nitrogen. (B) The spectra (b) at the end of the 25 min and after reperfusion with air for an additional 25 min (c) are shown along with the time course of the magnitude of the largest peak (a) [amplitude of (B) is 4 x (A)].200 Use of Nitroxides to Measure Redox Metabolism ,-4 I I oxygen concentration/pmol dmd3 Fig. 5. The relationship between the rate of reduction of a nitroxide by mammalian cells and oxygen concentration. The figure is based on initial first-order rate constants for various condi- tions of gas perfusion of lo7 TB cells in 100 mm3 with 2 mmol of nitroxide, in a gas-permeable plastic tube. The figure indicates the relation of the rate of reduction to three different oxygen con- centrations, the perfused gas (-), extracellular oxygen (- +-) and intracellular oxygen (- --).Oxygen concentrations in the intracellular and extracellular compartment were determined by methods described in fig. 1 and 2 and in media equilibrated with the perfusing gas by a Clark electrode. after freeze-thawing the rate of reduction of the different nitroxides is quite different; presumably this reflects chemical and, perhaps, physical (e.g. access to actual reducing sites) factors that affect the rate of reduction. The time course of the metabolism of a nitroxide to a non-paramagnetic state is illustrated in fig. 4(A). Fig. 4 also indicates that cells can reoxidize nitroxides back to the paramagnetic state; in fig.4(B) the amount of paramagnetic nitroxide is seen to increase when oxygen is reintroduced into a cell suspension that has, in the absence of oxygen, eliminated all of the detectable paramagnetic nitroxide. This indicates that the apparent ‘rate of reduction’ is actually a net rate of a reversible reaction between the nitroxide and hydroxylamine derivatives of each compound. In principle there can be a second set of equilibria between the oxidation and re-reduction of the nitroxide; these reactions are less facile in cells, and because the oxidation product tends to be unstable, the reaction is not reversible directly. Fig. 5 illustrates the relationship between the rate of loss of paramagnetism of a nitroxide and oxygen concentration. The inhibition of reduction of nitroxides by oxygen occurs at very low concentrations of intracellular oxygen.If the effect of oxygen concentration on the rate of reduction had been plotted conventionally as a function of oxygen concentration in the perfusing gas or as oxygen concentration in the extracellular fluid, the effect of oxygen on decreasing the rate of reduction of the nitroxide would have been more gradual and would have appeared to occur at considerably higher oxygen concentrations.H . M . Swartz 20 1 Table 2 summarizes data that illustrate that nitroxides can be good relaxing agents for water protons. The data show the effect of several nitroxides on relaxation of water protons in cell solutions of albumin at several different magnetic fields corresponding to Larmor frequencies from 0.01 to 50 MHz. The effects of some of the nitroxides are substantial and greater than those expected on the basis of simple additivity of the relaxation effects of nitroxides 011 water protons in simple aqueous solutions plus the relaxation rates of water protons in albumin solutions.In view of the limited number of nitroxides studied so far, it seems probable that we will find other nitroxides with even greater effects on relaxation of water protons of cells. The effectiveness of various nitroxides as relaxing agents seems to be independent of their responsiveness to metabolism. Some of the nitroxides, such as PCA and 5-doxy1 stearate, appear to combine metabolic responsiveness with good relaxivity. Others, such as CAT, and Tempo1 are both poorer relaxers and not responsive to oxygen. Some other nitroxides, such as 5-Tempamine, appear to be metabolically responsive but not espe- cially good relaxers.It therefore appears that the structural characteristics of nitroxides responsible for good relaxivity and metabolic responsiveness are at least partially independent of each other. This has positive implications for the development of nitroxides that have maximum metabolic responsiveness and good relaxivity. The data summarized in this section appear to indicate clearly that the nitroxides meet the major criteria needed for the development of in vivo n.m.r. contrast agents that reflect metabolism. The next tasks will be as follows (1) to determine which molecular structures provide optimal effects in regard to the principal variables for clinical uses, i.e.amount of relaxivity, amount of responsiveness to oxygen concentration (rate of reduction), critical oxygen concentrations for reduction and minimal cellular toxicity; (2) to obtain #a more thorough understanding of nitroxide-cell interactions, including the location of the various nitroxides, the kinetics of reduction (and reoxidation and oxidation and rereduction) of various nitroxides, the mechanism of reduction and oxidation of nitroxides and their derivatives and the structure-function relationships for these various interactions; ( 3 ) to determine if metabolically responsive contrast can be observed in vivo by m.r.i., where additional variables such as selective organ concentration (e.g. by means of different lipid solubility), selective organ toxicity and differential effects in neighbouring tissues may also be important variables. The design of maximally effective nitroxide contrast agents will clearly be a complex undertaking but our data give good reasons for optimism about the potential for such developments.The data indicate that the major properties that are pertinent for clinical use are functions of the molecular structure of the nitroxides and in principle can be optimized. Each of the clinically important variables seems to be dependent on different properties of the contrast agent and there is no indication that any of the major desirable ,attributes of these contrast agents are mutually exclusive. In fact, amongst the nitroxides we have studied, the doxy1 stearates appear to be both the best relaxers and the most metabolically responsive potential contrast agents.The multiple paths of potential cellular metabolism of nitroxides (oxidation and reversible reduction) will make evaluation of clinical use of metabolically responsive contrast agents more complex but raises the possibility of some additional approaches. In principle one could administer the reduced or oxidized derivatives of nitroxides rather than the nitroxides themselves; this would permit obtaining increased contrast in regions in which there is increased oxidation or reduction, respectively. Presumably the rates of metabolism for these nitroxide derivatives would follow different structure-function relationships than those for the reduction of nitroxides to hydroxylamines.In theory it also is possible to have oxidation (instead of reduction as observed here) of nitroxides to non-paramagnetic derivatives as the principal means of conversion of paramagnetic nitroxides to nonparamagnetic derivatives, but in our experience this is not a likely reaction in cells of the currently available nitroxides; the direct reversal of this reaction202 Use of Nitroxides to Measure Redox Metabolism is also problematical because of the instability of the direct oxidation products of ni troxides. Our data are currently limited to the effects of one clinically important metabolic variable (oxygen concentration) and one class of in vivo n.m.r. contrast agents (nitroxides). However, in principle this approach could be extended to other types of metabolic parameters and to the other major class of paramagnetic contrast agents, the paramagnetic metal ions.We have some preliminary results that indicate that it may be possible to develop nitroxides whose rate of reduction depends on specific other meta- bolic factors such as pH or cellular energy source. To date paramagnetic metal com- plexes used in n.m.r. contrast agents have been selected, in part, on the basis of their stability. It should be feasible to develop paramagnetic metal complexes that will oxidize or reduce to and/or from nonparamagnetic states, in vivo. The oxidation and reduction of paramagnetic metal ions is a normal occurrence in the function of many metal containing enzymes. The work summarized in this report is based on the efforts of a number of colleagues and students, including H. Bennett, S.Bernstein, R. Brown 111, K. Chen, W. Hyslop, S. Koenig, P. Morse 11, M. Nilges, M. Pals, M. Schara and M. Sentjurc. Financial support for this work was provided by the U.S. National Institute of Health (grants nos. GM-34250, CA-40665, GM-35534 and RR 0181 1) and the U.S. National Foundation for Cancer Research. References 1 Spin Labeling: Theory and Applications, ed. L. J. Berliner (Academic Press, New York, 1986). 2 Spin Labeling Theory and Applications: II, ed. L. J. Berliner (Academic Press, New York, 1979). 3 R. Mehlhorn and L. Packer, Ann. N. Y. Acad. Sci., 1983, 46, 180. 4 R. Mehlhorn, P. Landau and L. Packer, Methods Enzymol., 1982, 88, 751. 5 P. D. Morse I1 and H.M. Swartz, Magn. Reson. Med., 1985, 2, 114. 6 R. Mehlhorn and L. Packer, Methods Enzymol., 1979, 56, 515. 7 R. Mehlhorn and I. Probst., Methods Enzymol., 1982, 88, 334. 8 H. M. Swartz, K. Chen, M. Pals, M. Sentjurc and P. D. Morse 11, Magn. Reson. Med., 1986, 3, 169. 9 H. M. Swartz, H. Bennett, R. Brown 111, P. Morse I1 and S. Koenig, Period. Biol., 1985, 87, 175. 10 H. M. Swartz, M. Sentjurc and P. D. Morse 11, Biochim. Biophys. Acta, 1986, 82, 888. 11 Y. Mendelson, P. W. Cheung, M. R. Neuman, D. G. Fleming and S. D. Cahn, Adv. Exp. Med. Biol., 12 W. J. Whalen, Exp. Med. Biol., 1973, 37A, 17. 13 B. Chance, C. Barlow, Y. Nakase, H. Takeda, A. Mayevsky, R. Fischetti, N. Graham and J. Gorge, 14 M. Tamura, N. Oshino, B. Chance and I. A. Silver, Arch. Biochem. Biophys., 1978, 191, 8. 15 D. M. Benson, J. A. Knopp and 1. S. Longmuir, Biochim. Biophys. Acfa, 1980, 591, 187. 16 H. M. Swartz and M. Pals, in Handbook of Biomedicine of Free Radicals and Antioxidants, ed. J. Miguel, H. Weber and A. Quintanilha (CRC Press, Boca Raton, Florida), in press. 17 D. A. Windrem and W. Z. Plachy, Biochim. Biophys. Acta, 1980, 600, 655. 18 W. K. Subczynski and J. S. Hyde, Biochim. Biophys. Acta, 1981, 643, 283. 19 M. J. Povich, Anal. Chem., 1975, 47, 346. 20 T. Sarna, A. Duleba, W. Korytowski and H. Swartz, Arch. Biochem. Biophys., 1980, 200, 140. 21 J. M. Backer, V. G. Budker, S. I. Eremenko and Yu N. Molin, Biochim. Biophys. Acta., 1977,460, 152. 22 C. S. Lai, L. E. Hopwood, J. S. Hyde and S. Lukiewicz, Proc. Nut1 Acad. Sci. USA, 1982, 79, 1166. 23 W. Froncisz, C. S. Lai and J. S. Hyde, Proc. Nut1 Acad. Sci. USA, 1985, 82, 41 1. 24 J. Kaplan, P. G. Canonico and W. J. Caspary, Proc. Nut1 Acad. Sci. USA, 1973, 70, 66. 25 G. L. Wolf, K. R. Burnett, E. J. Goldstein and P. M. Joseph, Magn. Reson. Annu., 1985, 231. 26 R. C. Brasch, D. A. London, G. E. Wesbey, T. N. Tozer, D. E. Nitecki, R. D. Williams, J. Doemeny, 27 R. Brasch, Radiology, 1983, 147, 781. 28 R. C. Brasch, D. E. Nitecki, M. Brant-Zawadzki, D. R. Enzmann, G. E. Wesbey, T. N. Tozer, 1981, 159, 93. Am. J. Physiol., 1978, 235, H809. L. D. Tuck and D. P. Lallemand, Radiology, 1983, 147, 773. L. D. Tuck, C. E. Cann, J. R. Fike and P. Sheldon, Am. J. Neurorad., 1983, 4, 1035. Paper 6/1132; Received 5th June, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300191

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. I , 1987, 83, 203-209 Electron Spin Resonance and Saturation Transfer Electron Spin Resonance of Virus and Membrane Systems Marcus A. Hemminga Department of Molecular Physics, Agricultural University, de Dreijen 11, 6703 BC Wageningen, The Netherlands A review is given of the application of e.s.r. and saturation transfer e.s.r. techniques to virus and membrane systems, with particular emphasis on the molecular processes of virus infection. A characteristic feature of viruses is that they are not capable of reproducing indepen- dently: virus particles need the biochemical machinery of a host cell. The genome of small viruses as plant viruses is only sufficient to code for one or a few different coat proteins. To build large structures, multiple copies of identical subunits are needed.Many plant viruses are therefore composed of nucleic acid encapsidated by a large number of identical protein subunits. The objective of our research is to obtain information about the molecular mechanism of virus infection. The following questions can then be raised : How do viruses penetrate the cell? How and where do they dissociate? How does replication take place? How and where does virus assembly occur? The answer to these questions requires a thorough knowledge of the fundamentals of protein-protein, protein-nucleic acid and protein- membrane interactions. The paper presents a review of the way in which spin label e.s.r. and saturation transfer e.s.r. techniques have contributed to answering some of the basic questions of virus infection.Spin Label E.S.R. on Virus Assembly An approach to studying the assembly of virus particles is to attach e.s.r. spin labels covalently to specific groups on the viral protein or the nucleic acid. E.s.r. can then be employed to study the structure and dynamical state of the assembly products. Tobacco mosaic virus (TMV) is a long, rod-like virus particle consisting of RNA helically surrounded by 2130 identical protein subunits. The coat protein of TMV contains a single cysteinyl residue at which maleimide spin labels can be covalently attached.' TMV protein modified in this way shows a two-component spectrum: a sharp three-line spectrum arising from mobile spin labels and a broad powder-like component stemming from immobilized spin labels.The immobile component is attributed to spin labels probing interaction with neighbouring protein subunits.' On increasing the pH, the relative amount of immobile component decreases (fig. l), indicating a reduction of the size of the protein aggregates. The association of spin-labelled TMV protein into larger aggregates can still take place, although the conditions are changed, owing to the disturbing effect of the spin label. Nevertheless, the qualitative agreement with the association of native TMV protein is good: around pH 5, TMV protein forms large helical oligomers, whereas mainly trimers are found around pH 9. The assembly of TMV protein and poly-A has been followed with iodoacetamide spin labels attached to the poly-A.2 At the start of the reaction the e.s.r.spectrum arises only from free mobile spin-labelled poly-A. In the course of the assembly process the mobile component decreases and at the same time a powder spectrum appears. The latter 203204 E.S.R. and S.T.-E.S.R. of Viruses and Membranes 3 - $ 2- k .s: 1- 5 6 7 8 9 10 PH Fig. 1. Variation of h,/h, with pH for TMV protein labelled with spin labels I and I1 (see inset). h, and h, are the heights of the low-field lines of the sharp and broad components, respectively, in the e.s.r. spectra. From ref. (1). spectrum is attributed to spin-labelled poly-A immobilized between assembled protein subunits. The decrease of the mobile component (fig. 2), therefore, is a measure for the rate of formation of nucleoprotein particles.2 When the TMV protein subunits are mainly present as trimers at the start of the experiment, the assembly reaction is fast (k = 2 x min-l).However, when the protein is in the double-disc form, consisting of 34 subunits, the reaction is much slower (k = 1 x min-l). This is a very interesting observation, since it is well known that TMV RNA assembles very fast with protein in the double-disc This is because TMV RNA forms an initiation loop which fits well in the opened binding groove on the double The secondary structure of RNA, therefore, plays an essential role in the assembly process. Poly-A, however, does not form well defined loops which fit in the double disc, thus explaining the much slower reaction rate. Saturation Transfer E.S.R. on Virus Systems From the foregoing studies it is found that, owing to the high molecular weight of the assembly products, e.s.r.powder spectra are obtained. Such spectra are insensitive to the dynamical properties of the aggregates. The introduction of saturation transfer (s.t.-e.s.r.), with sensitivity in the motional range from lo-' to s for the rotational correlation time z,, has given rise to new possibilities for the use of spin labels in the study of protein aggregation and virus assembly. The association of maleimide spin-labelled TMV protein in the pH range 5-7.5 has been monitored by s.t.-e.s.r. (fig. 3).6 Native TMV protein forms double-disc structures around pH 7 and large helical rods around pH 5. The s.t.-e.s.r. spectra of spin-labelled protein indicate the occurrence of large protein aggregates undergoing molecular motions with z, values ranging from ca.10+ s at pH 7.5 to ca. s at pH 5.2.6 The analysis of the s.t.-e.s.r. spectra further shows that the rotational motion is not isotropic. This may have two origins: (1) the anisotropic motion is due to a non-spherical shape of the oligomers; (2) the anisotropy stems from local motions of the spin label with respect to its binding site. With the present knowledge, it has not been possible to analyse the s.t.-e.s.r. spectra in fig. 3 further in terms of detailed motional models.M. A . Hemminga 205 G 0 d 3.2 - 3.0 - * a I I 0 100 200 300 400 500 Fig. 2. The time dependence of the height h of the low-field peak of the sharp component in the e.s.r. spectra for the reaction of TMV protein with spin-labelled poly-A.0, TMV protein in trimer form; 0, TMV protein in disc form. From ref. (2). tlmin 2 mT V Fig. 3. Second harmonic 200 kHz quadrature s.t.-e.s.r. spectra of maleimide spin-labelled TMV protein at various values of the pH. From ref. (5).206 E.S.R. and S.T.-E.S.R. of Viruses and Membranes Fig. 4. Second-harmonic 200 kHz quadrature s. t.-e.s.r. spectra of maleimide spin-labelled CCMV (a) and its protein dimers (b). The inset shows the schematic structure of the maleimide spin label. The axes of the nitroxide group are indicated by x and y . The z axis is perpendicular to the xy plane. From ref. (7). i I I I I 1 1 o-6 1 o - ~ 1 o-l 1 0 - ~ (77/WPa s K-’ Fig. 5. Variation of the rotational correlation time z with q/T for maleimide spin-labelled CCMV in glycerol-water mixtures, deduced from the central (@) and low-field part (A) of the s.t.-e.s.r.spectra. The dotted line gives the Stokes-Einstein relation for CCMV particles. From ref. (8). Cowpea chlorotic mottle virus (CCMV) is a spherical plant virus consisting of RNA surrounded by 180 identical, icosahedrally arranged protein subunits. Also, CCMV protein can easily be labelled with a maleimide spin label.6v Fig. 4 shows the s.t.-e.s.r. spectra of spin-labelled protein dimers and spin-labelled virus. The variations in the low- and high-field part of the spectra can be ascribed to the difference in the molecular weight of the dimer and virus particles. A striking effect, however, is that the line shape in the centre hardly changes. This is indicative for the presence of anisotropic motion.Cross-linking experiments have ruled out the possibility of anisotropic internal motion of the protein subunits. Therefore, the maleimide spin label is not rigidly attached to the subunits, but is performing a local anisotropic motion, probably in a cleft. The anisotropic motion was found to be mainly about the z-axis of the spin label (see fig. 4), which approximately coincides with the long molecular axis of the label.M . A . Hemminga Q I 207 Fig. 6. Experimental (-) and calculated (---) second-harmonic (100 kHz) quadrature (Q) and in-phase (I) spectra of the cholestane spin label in macroscopically oriented multibilayers of dimyristoylphosphatidylcholine with 33 mol % cholesterol at 0" orientation. For the calculation the program of Perkins et al." has been used.The q values at various temperatures have been deduced from the best computer fits. From ref. (10). (a) T = - 10 "C, = 3 p; (b) T = 0 "C, = 2.5 p s ; (c) T = 10 "C, q = 2 p s ; ( d ) T = 20 "C, = 1.2 PS. Since CCMV is an almost spherical particle, the overall motion can be reasonably well described by a single isotropic rotational correlation time 7,. The overall particle motion, therefore, can be varied in a well defined way in glycerol-water systems with different viscosities q according to the Stokes-Einstein relation: z, = qV/kT. Here, V is the particle volume, T the absolute temperature and k the Boltzmann constant. This enables a more quantitative analysis of the anisotropic spin-label motion in CCMV.8 The rotational correlation times can be estimated from a comparison of selected regions in the s.t.-e.s.r. spectra with those of reference spectra of isotropically reorienting spin labels. Fig. 5 shows the values of the rotational correlation time z deduced from the low-field and central regions of the spectra as a function of V / T . ~ The two different z values indicate anisotropic motion of the spin label about its z-axis. For low values of q/T the z value deduced from the low-field part of the spectrum approximately corresponds to the overall motion of the virus particle. At higher values of q/T this z value reaches a constant value of 2 x s, which is attributed to the local motion of the spin label z-axis. The rotational motion about the z-axis is reflected in the central part of the spectrum.The z value for this motion varies from ca. lo-' to ca. s with increasing macroscopical viscosity. Saturation Transfer E.S.R. of Anisotropic Motion in Membranes Anisotropic motion, as encountered in TMV and CCMV systems, is a very complicating factor in the analysis of s.t.-e.s.r. spectra. Therefore, the problem of anisotropic motion has been approached in another way using the cholestane spin label embedded in macroscopically oriented lipid sy~tems.~9 lo This spin label orients perfectly in bilayers of lecithin and cholesterol, so that the rotational motions of the spin label are limited about the long molecular axis, which is parallel to the nitroxide y-axis. When the lipid bilayers are macroscopically oriented, the anisotropic motion of the spin label is directly observable in the spectra at 90" orientation.This has been used to observe the effect of cholesterol on dimyristoylphosphatidylcholine bilayers below the phase tran~ition.~ For208 E.S.R. and S.T.-E.S.R. of Viruses and Membranes Fig. 7. E.s.r. spectra at 30 "C of phosphatidylcholine spin label (a) and phosphatidic acid spin label (b) of 30 wt % M 13 coat protein (gene 8 product) in a mixed lipid system of dimyristoylphospha- tidylcholine and dimyristoylphosphatidic acid (weight ratio 4/ 1). Both spin labels are labelled on the C14 atom of the sn-2 chain. For comparison the e.s.r. spectrum of phosphatidylcholine spin label in the lipid system without coat protein is given in (c). This spectrum is identical to that obtained with the phosphatidic acid spin label.From ref. (13). comparison, the s.t.-e.s.r. spectrum of the cholestane spin label in randomly oriented bilayers has been measured. Retrieving information on motion from such spectra is less straightforward than the analysis of the oriented-bilayer ~pectra.~ The s.t.-e.s.r. spectra at 0" orientation are insensitive to the anisotropic spin label motion. This has enabled the determination of the values at various temperature; in lipid systems (fig. 6).1° This information is valuable in a more quantitative approach to the analysis of s.t.-e.s.r. spectra using computer simulations.11 E.S.R. of Viral Protein incorporated in Lipid Bilayers M13 bacteriophage is a long, rod-like virus particle, consisting of DNA surrounded by coat proteins. In vivo bacteriophage M13 enters Escherichia coli cells, leaving its coat proteins in the cytoplasmic membrane.12 After DNA duplication and coat-protein synthesis both progeny as well as parental coat proteins are stored as integral membrane proteins.During the membrane-bound assembly of new virus particles, the viral DNA is complexed with numerous coat protein molecules, without lysis of the host cell. The e.s.r. technique has been used to study the interaction of the major (gene 8 product) M13 coat protein with a mixed lipid system (fig. 7).13 In the liquid-crystalline state of the lipids the e.s.r. spectra of phospholipid spin labels consist of two components. One component corresponds to a fluid lipid environment, similar to that found in reference lipid samples.The second component has a larger hyperfine splitting not seen in the reference samples and represents a lipid environment with a considerably more restricted mobility. This component is attributed to lipids directly interacting with the coat protein, As judged from the ratio of the left-hand peaks in the spectra, the specificity for the negatively charged phosphatidic acid spin label for binding to M 13 coat protein is approximately twice that of the phosphatidyl choline spin label. This may be ofM. A . Hemminga 209 relevance for the infection mechanism of M13, since in vivo molecular association of the M 13 coat protein with negatively charged cardiolipin in the E. coli membrane has been suggested from altered host lipid metabolism after insertion of the M13 coat protein.14 Conclusions In this review it has been shown that e.s.r.spin-label techniques can be applied to complex systems encountered in the process of virus infection. One approach, based on conventional e.s.r., has been employed to study the aggregation of virus coat protein and the kinetics of the assembly of nucleoprotein particles. This is possible because the spin label is located so that it is probing different environments, giving rise to a superposition of fluid-like and solid-like e.s.r. spectra. The solid-like spectrum represents the high- molecular-weight assembly products. A similar situation is found with spin-labelled lipids probing the interaction with virus coat proteins embedded in lipid bilayers. Here, the solid-like spectrum arises from protein-interacting spin labels, similar to those found in other syste111s.l~ S.t.-e.s.r.has been shown to be a tool which can directly characterize the high- molecular-weight assembly products and virus particles. However, since the spin labels employed may not be rigidly attached to the protein and generally perform an anisotropic motion, analysis of s.t.-e.s.r. spectra is found to be complicated. Since analysis of anisotropic motion may be helpful in further characterizing complex biosystems, this type of motion has been studied in detail for spherical virus particles and macroscopically oriented model membranes. Whereas e.s.r. is a linear technique, s.t.-e.s.r. is highly non-linear. Therefore, several factors determine the lineshape of s.t.-e.s.r. spectra, which do not show up in conventional e.s.r.These are both the value and the distribution of the microwave and modulation fields in the cavity and the dielectric properties and shape of the sample. These factors also complicate the analysis of s.t.-e.s.r. spectra. The effect of these factors has been systematically investigated in a number of model experiments, leading to a proposal for standardizing the recording of s.t.-e.s.r. spectra.16-19 I am indebted to Dr A. Watts for kindly providing the phospholipid spin labels. Part of this work is supported by the Netherlands Foundation for Biophysics, with financial aid from the Netherlands Organization for the Advancement of Pure Research (ZWO). References 1 M. A. Hemminga, T. van den Boomgaard and J. L. de Wit, FEBS Lett., 1978, 85, 171. 2 H. W. M. Hilhorst, U. D. Postma and M. A. Hemminga, FEBS Lett., 1982,142, 301. 3 P. J. G. Butler and A. C. H. Durham, Ado. Protein Chem., 1977, 31, 188. 4 P. J. G. Butler, Nature (London), 1971, 233, 25. 5 M. A. Hemminga, S. P. F. M. Roefs and J. Kruse, Bull. Mugn. Reson, 1981,2, 345. 6 J. Kriise and M. A. Hemminga, Eur. J. Biochem., 1981, 113, 575. 7 G. Vriend, J. G. Schilthuis, B. J. M. Verduin and M. A. Hemminga, J. Magn. Reson., 1984, 58, 421. 8 M. A. Hemminga and A. J. Faber, J, Mugn. Reson., 1986,66, 1. 9 P. Koole, C. Dijkema, G. Casteleijn and M. A. Hemminga, Chem. Phys. Lett., 1981, 79, 360. 10 P. Koole and M. A. Hemminga, J. Magn. Reson., 1985, 61, 1 . 11 R. C. Perkins Jr, T. Lionel, B. H. Robinson, L. A. Dalton and L. R. Dalton, Chem. Phys., 1976, 16, 12 D. A. Marvin and E. J. Wachtel, Nature (London), 1975,253, 19. 13 C. J. A. M. Wolfs, K. P. Datema and M. A. Hemminga, unpublished results. 14 B. K. Chamberlain and R. E. Webster, J. Biol. Chem., 1976, 251, 7739. 15 D. Marsh, Trends Biochem. Sci., 1983, 8, 330. 16 M. A. Hemminga, Chem. Phys. Lipids, 1983, 32, 323. 17 M. A. Hemminga, J. H. Reinders and P. A. de Jager, J . Mugn. Reson., 1984, 58, 428. 18 M. A. Hemminga, F. A. M. Leermakers and P. A. de Jager, J . Mugn. Reson., 1984,59, 137. 19 M. A. Hemminga, P. A. de Jager, D. Marsh and P. Fajer, J . Mugn. Reson., 1984,59, 160. 393. Paper 6/852; Received 1st May, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300203

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. SOC., Faraday Trans. 1, 1987, 83, 211-217 Reactions of Ozonate and Superoxide Radical Anions Alexander R. Forrester" and Vemishetti Purushotham Chemistry Department, University of Aberdeen, Meston Walk, Old Aberdeen AB9 2UE, Scotland Potassium ozonate has been prepared by reaction of ozone with potassium superoxide dispersed in Freon-1 2. Orange solutions of potassium ozonate in benzene or toluene containing 18-crown-6 react with nitrone traps to give spin adducts mainly derived from the oxide radical anion, which is a decomposition product of the ozonate radical anion. With 2-methyl-2- nitrosopropane in toluene solution the ozonate gives five nitroxide radicals in whose formation the oxide radical anion is again implicated. With nitrosobenzene electron transfer occurs.Comparable reactions with super- oxide are described. Reactions of simple oxygen-centred radicals (HO', HOO') and their radical anions (O*-, 00'-) with organic substrates are under intense study at present mainly because of their role in biological redox processes.1 Meanwhile, a close relative of these, the ozonate radical anion (Oi-), has been largely ignored. It has been postulated as an intermediate in reactions of ozone with certain easily oxidised substrates2 (aromatic amines, sulphides) and there are reports of oxidation of acetaldehyde3 and alkanes4 in the vapour phase with ozonates supported on magnesium oxide. These apart, there has been little or no attempt to investigate reactions of metal ozonates with organic substrates in the liquid phase despite the fact that alkali-metal ozonates were first prepared more than 35 years We now describe a method whereby solutions of potassium ozonate in organic solvents can be obtained and reactions of these solutions with nitrone and nitroso spin traps examined.Comparison is made with reactions of superoxide under similar conditions. Potassium ozonate was prepared' as an orange powder by stirring finely ground potassium superoxide suspended in dried Freon presaturated with ozone at -90 "C. Chromatographic silica gel, dried at 125 "C, was added to the Freon and stirring was continued while the temperature rose to room temperature and the Freon evaporated. The potassium ozonate-silica gel mixture so obtained was stable in air for ca. 3 h but decomposed more rapidly if stored in a vacuum desiccator.Under organic solvents such as benzene or toluene at room temperature it survived for ca. 24 h. Accordingly, it was normally stored under solvent at -40 "C, under which conditions it survived for months. During the preparation and in all subsequent manipulations water and other hydroxylic solvents had to be rigorously excluded.8 Reaction with water is rapid (1) 0;-+ H20 + 'OH + -OH + 0, and the purpose of the silica gel was to protect the ozonate from such reactions. The spontaneous decomposition of ozonate 0;- + 0, + 0'- (2) is accelerated predictably by an increase in temperature and a decrease in When 18-crown-6 was added to the benzene or toluene under which the ozonate was stored and the mixture was briefly shaken a bright orange solution of potassium ozonate (A,,,, = 430 nm) was produced.Such solutions were stable for only a few minutes at 21 1212 Ozonate and Superoxide Radical Anions room temperature, after which the colour faded rapidly. These coloured solutions gave rise to singlet e.s.r. spectra, g = 2.0107, whose disappearance coincided with that of the orange colour. Previous measurementsg7 lo reported a g value of 2.01 17. Decomposition of the ozonate was a first-order process [reaction (2)] k,,, = 3.16 x lop4 s-l at - 30 "C, 8.57 x s-l at 0 "C. At -30 "C ozonate solutions in toluene have a half-life of 38 min. These values differ significantly from earlier ones measured, albeit under significantly different conditions. For example, ozonate in 8 mol dm-3 potassium hydroxide solution has been reportedg* l1 to have a half-life of 12 min at - 50 "C and to decompose in aqueous alcohol at pH 13 according to reaction (2), with k = 3.3 x lo3 s-l at 25 "C. Our measurements made on solutions in non- hydroxylic organic solvents clearly relate only to reaction (2), while those on aqueous alkaline solutions may also have a contribution from reaction (1).s-l at-20 "C and 7.18 x Experimental Freon-12 from a cylinder cooled to - 18 "C was dried by passage through a drying tower containing sodamide and five towers containing molecular sieves (BDH 5A) before being condensed and stored over molecular sieves at -40 "C. It was then distilled into a three-necked flask, fitted with a mechanical stirrer. The Freon (80 cm3) was saturated at -90 "C with ozone by passage of a stream of ozonised oxygen from a commercial ozoniser.The oxygen gas was pre-dried by successive passage through a coil immersed in a methanol-liquid nitrogen slush bath, a sodamide tower and three molecular sieve towers before entering the ozoniser. The gas mixture from the ozoniser was passed through another tower containing molecular sieves (5A) before bubbling into the condensed Freon. Saturation of the Freon by ozone required 2 h. Potassium superoxide (500 mg), which had been finely ground in a nitrogen glovebox, was then added to the solution of ozone in Freon and stirring was commenced. The liquid-nitrogen cooling bath was removed and the temperature was allowed to rise to -30 "C. The flask was held at this temperature and stirring was continued for 15 min, after which the cooling bath was removed and the Freon allowed to evaporate.When the volume had been reduced to ca. 10 cm3, neutral chromatographic silica gel (5 g) (previously dried at 125 "C for 19 h) was added and the flask swirled by hand to ensure mixing of silica gel and potassium ozonate. Complete evaporation of the Freon left an orange-red solid, the colour of which always persisted for several hours when exposed to air, although the exact period of survival varied from batch to batch. Purification of the ozonate by dissolution in liquid ammonia, filtration and evaporation of ammonia filtrate always gave a product which was much shorter-lived than the material obtained directly from the Freon-1 2. With such samples loss of colour occurred within 20 min and so most of the e.s.r.experiments were made using ozonate which had not been treated with liquid ammonia. E.s.r. measurements were made on a Varian El04 e.s.r. spectrometer in conjunction with a Varian (E-4557) variable-temperature unit, the sample temperature being measured with a copper-constantan thermocouple. Spin Trapping with Nitrones Addition of phenyl t-butylnitrone (PBN) to an orange solution of ozonate in benzene-crown ether at room temperature in an e m . cavity caused rapid disappearance of the orange colour and replacement of the ozonate signal by a spectrum showing (see fig. 1) aN = 14.2 G and aH = 2.0 G. This corresponds12 to the oxide (hydroxy) adduct of PBN. Hence, unless the ozonate adduct has identical coupling constants, which is unlikely, the oxide (hydroxy) adduct must arise either by decomposition of the ozonate spin adduct or by trapping oxide (hydroxy) radical anions formed by decomposition ofA .R. Forrester and V. Purushotharn 213 03- t PBN /C,H, / I8 - crown- 6/SiO, OH I C~H~CHNBJ I 0. Fig. 1. E.s.r. spectrum of radicals present in solution of potassium ozonate, PBN and 18-crown-6 in benzene. Fig. 2. E.s.r. spectrum (a) [and simulation (b)] of radicals formed on reaction of potassium superoxide with DMPO in toluene containing 18-crown-6 at - 50 "C.214 Ozonate and Superoxide Radical Anions Fig. 3. E.s.r. spectrum of radicals formed on reaction of potassium ozonate with MNP in benzene containing 18-crown-6. ozonate. When this experiment was repeated at low temperature in toluene (- 50 to - 30 "C) an ill-resolved triplet was detected (a, 13.8 G, aH = 1.8 G).The oxygen which is produced during decomposition of the ozonate causes broadening of the lines. Degassing at room temperature and remeasurement gave a well resolved spectrum of the hydroxy adduct of PBN. Use of dimethylpyrroline-1-oxide (DMPO) as a trap for the ozonate ion in toluene- 18-crown-6 at -50 "C gave a spectrum showing a broad quartet aN x aH % 12.5 G. At room temperature in benzene or toluene-1 8-crown-6 a complicated unsymmetrical pattern of lines developed. After ca. 30 min this had simplified and two spectra were resolved, with aN = 14.4 G and aH = 16.6 G and aN = 13.2 and 1.6 G. We have detected the latter previously when solutions of either superoxide or ozonate in benzene or toluene containing DMPO are left in the cavity for some time.This spectrum appears to be due to a s-alkyl-t-alkyl nitroxide arising from a t-nitroso compound formed from the trap. The former is clearly the spectrum of a DMPO adduct, and in view of results obtained with nitroso traps (see later) is most likely to be the spin adduct obtained from a radical derived from the crown ether by hydrogen abstraction. Hence, the low-temperature spectrum is assigned to either the ozonate or more likely the oxide (hydroxy) adduct formed as previously suggested for the PBN adduct. Since the ozonate produced by ozonisation of superoxide was only ca. 70% pure (u.v.-visible measurements) it was important to show that the low-temperature spectrum in particular did not arise simply from residual superoxide in the orange solid.Hence DMPO was added to a solution of superoxide and 18-crown-6 in toluene at - 50 "C and the radicals formed were detected by e.s.r. The spectrum observed is shown in fig. 2, and is very different from that obtained from the ozonate under similar conditions. The spectrum is indeed the composite of two: the major component is the superoxide adduct aN = 12.8 G, aH = 6.0 G and aH = 2.0 G and the minor one the oxide (hydroxy) adduct aN = aH = 13.0 G. The superoxide adduct spectrum (benzene solvent) has been reported13 previously (aN = 12.9 G, aH = 6.9 G), but no mention was made of the smallA . R. Forrester and V. Purushotham 215 8-hydrogen coupling. Our attempt to detect the superoxide adduct in benzene4 8-crown-6 at room temperature was unsuccessful.When the temperature was raised stepwise to 10°C the minor component intensified as the superoxide adduct waned. This is consistent with the known1* stabilities of these two adducts and the formation of oxide (hydroxy) adduct from superoxide adduct. Spin Trapping with Nitroso Compounds Addition of 2-methyl-2-nitrosopropane (MNP) to a degassed solution of ozonate and 18-crown-6 in toluene at - 50 "C in an e.s.r. cavity gave rise to a spectrum with five broad lines 14-15 G apart. The temperature was raised stepwise to 0 "C but there was little improvement in the resolution. Brief degassing by nitrogen bubbling followed by remeasurement at 10 "C gave well resolved signals attributable to five radicals (fig.3), namely di-t-butyl nitroxide (1) (a, = 15.5 G), t-butyl-t-butoxy nitroxide (2) (fN = 27.5 G), benzyloxy-t-butyl nitroxide (5) (a, = 29.6 G, aH = 1 .O G), benzyl-t-butyl nitroxide (4) (a, = 15.0 G, agH2 = 7.3 G; weak) and a nitroxide (3) derived from the crown ether (aN = 13.2, aH = 1.6 G). Consistent with these assignments,15 the triplet with aN = 15.5 G steadily intensified as that with aN = 27.5 G waned. In addition, treat- ment of a solution of 18-crown-6 and MNP in benzene with the t-butoxyl radical source di-t-butyl peroxalate also gave inter alia a nitroxide with a, = 13.2 G and aH = 1.6 G. Using benzene as solvent in the temperature range 5-20 "C only three of the above ButNd:) ButN02 + PhCH20* - ButNO PhCHzONBut I AH (3) 0- & I BuWO ( 5 ) I ButNOOCH2Ph PhCHzOO t I8-crown-6 f 0, ButNO I ButNO I I 0.(4) Bu'NO; - Bu'N0;- -NO;- But' - BufNO I - 0 2 0- j2 ktN0 Bu'NBut I 0. (1) Bu'OO. ButNO I ButNOOBut ButNO C- BUQ. + Bu'NOz Scheme 1.216 Ozonate and Superoxide Radical Anions radicals were detected (1)-(3), and with ethylbenzene, t-butyl- 1 -methylbenzyl and t-butyl-I-methylbenzyloxy nitroxides (a, = 14.8 G, agH = 3.8 G and a, = 29.1 G, respectively) were detected in addition to (1)-(3). Any reaction scheme which seeks to account for the production of these several radicals must accommodate the following points. (i) A relatively powerful hydrogen abstractor is present in the system. (ii) The alkoxy-t-butyl nitroxides do not arise by reaction of alkyl radicals with 2-methyl-2-nitropropane.This was confirmed by allowing di-t-butyl peroxalate to react with 2-methyl-2-nitropropane in toluene. Neither t-butoxy nor benzyloxy-t-butyl nitroxide was detected. In scheme 1 it is postulated that the oxide radical anion is the hydrogen abstractor. The origin of the t-butyl and hence also t-butoxyl radicals may be attributed simply to the known homolysis of MNP or to addition of the oxide or ozonate radical anions to MNPls followed by fragmentation of the ensuing 2-methyl-2-nitropropane radical anion.17 Our system does not allow us to distinguish, although the relatively high yield of di-t-butyl nitroxide in the low- temperature (and dark) reactions suggests that the latter route makes a significant contribution. In a similar experiment with MNP, toluene, 18-crown-6 and potassium superoxide the final spectrum again showed the presence of radicals (1)--(5), although relative concentrations were different from those in which ozonate was used.Higher concentra- tions of di-t-butyl and benzyl-t-butyl nitroxides (3) and (4) were formed, with a decrease in the concentrations of the corresponding alkoxy nitroxides. This is readily understandable, since ozonate would be expected to produce more oxygen than superoxide. The nature of the hydrogen abstractor is less obvious and is unlikely to be superoxide itself. One possibility is the 2-methyl-2-peroxynitroso radical anion (6) formed by the reaction: ButN=O + 0;- + ButNOO- (3) I 0' (6) The same intermediate has previously been suggested1* to participate in the reaction of potassium t-butoxide with MNP, and a related species [which may indeed be (611 is thoughtlS to mediate in the photo-oxidation of alkylbenzenes in the presence of MNP.Its formation we depict by direct addition of superoxide to MNP, but initial electron transfer from superoxide to MNP followed by addition of oxygen to the ensuing MNP radical anion cannot be excluded. In this case also there are a number of routes to t-butyl and hence to t-butoxyl radicals in addition to simple homolysis of the nitroso compound. For example, addition of (6) to MNP, fragmentation of the ensuing adduct to 2-nitropropane and its radical anion ButNOO' + ButN = 0 + ButNOONBut --+ ButNO, + ButNO;- (4) I \ 0- 0' and homolysis of the nitro radical anion. However, there is no evidence to support this route, since 2-methyl-2-nitropropane radical anion was not detected, nor was it detected when 2-methyl-2-nitropropane was treated with superoxide.Reaction of ozonate with nitrobenzene at -40 "C gave the nitrosobenzene radical anion;20 aN = 8.8 G, aH = 4.0 G (2H), aH = 3.35 G and aH = 1.2 G (2H). The spectrum slowly faded as the temperature increased, and was finally completely replaced by that of a phenyl-s-alkyl nitroxide with aN = 10.2 G, uo,p-H = 2.3 G, am-H = 1.1 G and agH = 1.6 G. The small value of the alkyl hydrogen coupling suggests that the alkylA . R. Forrester and V. Purushotham 21 7 group is bulky and is most likely derived from the crown ether. A likely course of events is outlined in scheme 2, in which the hydrogen abstractor is the oxide radical anion.When nitrobenzene was replaced by nitrosodurene the nitrosodurene radical anion was 18-crown-6 Of'----- 0 2 + 0.- 0' H 1 PhNO Scheme 2. not detected, but only a nitroxide corresponding to (7) aN = 12.8 G, aH = 5.2 G. This assignment was supported by the detection of an identical spectrum when di-t-butyl peroxalate was added to a solution of 18-crown-6 and nitrosodurene in benzene. Electron transfer20 and hydrogen abstraction were also the main processes detected by e x . when superoxide reacted with nitrosobenzene in toluene containing crown ether. We thank the S.E.R.C. for financial support. References 1 J. A. Fee and J. S . Valentine, in Superoxide and Superoxide Dismutases, ed. A. M. Michelson, J. M. McCord and I. Fridovich (Academic Press, New York, 1977); Superoxide Dismutase, ed.L. W. Obberley (CRC Press, Boca Raton, Florida, 1982), vol. 1 and 2; I. Fridovich, in Free Radicals in Biology, ed. W. A. Pryor (Academic Press, New York, 1976), vol. 1, p. 239. 2 P. S. Bailey in Ozonation in Organic Chemistry (Academic Press, New York, 1982), vol. 2, chap. VII. 3 I. A. Kazarnovskii and N. P. Lipikhin, Dokl. Akad. Nauk SSR, 1976, 231, 1155. 4 Y. Takita and J. H. Lunsford, J. Phys. Chem., 1979, 83, 683. 5 I. A. Kazarnovskii and P. G. Nikolskii, Dokl. Akad. Nauk SSR, 1949, 64, 69; T. P. Whaley and 6 I. A. Kazamovskii, N. N. Lipikhin and M. V. Tikhomirov, 2. Fiz. Khim., 1956,30, 1429. 7 I. I. Volnov, S. A. Tokareva, V. I. Klimanov and G. P. Pilipenko, Bull. Acad. Sci. USSR, 1966, 1222. 8 A. F. Morkovnik and 0. Yu. Okhlobystin, Russ. Chem. Rev., 1979,48, 1055. 9 B. L. Gall and L. M. Dorfman, J. Am. Chem. SOC., 1969,91,2199. J. Kleinberg, J. Am. Chem. SOC., 1951, 73, 79. 10 A. D. McLachlan, M. C. R. Symons and M. G. Townsend, J. Chem. SOC., 1959,952. 1 1 P. S. Gorbenko-Germanov and I. V. Kozlova, Russ. J. Phys. Chem., 1974,48, 166. 12 E. G. Janzen, D. E. Nutter, E. R. Davis, B. J. Blackburn, J. L. Poyer and P. B. McCay, Can. J. Chem., 13 J. R. Harbour and M. L. Hair, J. Phys. Chem., 1978,82, 1397. 14 E. Finkelstein, G. M. Rosen and E. J. Rauckman, Mol. Pharmacol., 1979, 16, 676. 15 A. R. Forrester, in Landolt Bornstein, Magnetic Properties of Free Radicals (Springer-Verlag, Berlin, 16 P. S. Bailey, J. E. Keller and T. P. Carter, J. Org. Chem., 1970, 35, 2777; P. S. Bailey and J. E. Keller, 17 A. K. Hoffman, W. G. Hodgson, D. L. Maricle and W. H. Jura, J. Am. Chem. SOC., 1964,86,631. 18 H. G. Aurich and W. Dersch, 2. Naturforsch., Teil B, 1973, 28, 525. 19 J. A. Maassen and Th. J. De Boer, Red. Trav. Chim. Pays-Bas, 1973, 92, 185. 20 H. G. Aurich, W. Dersch and J-M. Hemrich, Annalen, 1981, 1271. 1978,56, 2237. 1979), vol. 9, pp. 192 ff. J. Org. Chem., 1970,35, 2782. Paper 61826; Received 28th April, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300211

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J. Chem. Soc., Faraday Trans. 1, 1987,83, 219-224 Spin-trapping Study of the Radiolysis of CCl, Alexander Halpern Institut f u r Chemie 1 (Nuklearchemie) der Kernforschungsanlage Jiilich GmbH, D-5170 Jiilich, Federal Republic of Germany This paper reports the spin trapping of radicals in liquid CC1, irradiated with low-energy X-rays of different wavelengths, using phenyl-t-butylnitrone (PBN) as a spin trap. The unstable PBN-C1 adduct and the relatively stable PBN-CCl, adduct have been detected. X-rays of energy above the K-edge for chlorine produce significantly more trapped radicals than those having energies below this K-edge value. This enhancement of radical formation is a consequence of the Auger effect, which follows the inner-shell photo- absorption. Experiments have also been made on trapping the radicals generated by the decay of tritium in a solution of tritiated methanol in CCl,.The detection and reactions of trichloromethyl radicals provide an important subject for study, stimulated by the recognition that this radical is a key intermediate in the metabolism and hepatoxicity of CCl, responsible for liver cell injury. [See ref. (1) for a recent review.] The most profitable experimental method appears to be the spin trapping of 'CCl, in vitro and in ~ i v o . ~ - ~ This method depends on the fast addition of any short- lived radical 'R to a nitroso or nitrone compound to form a relatively long-lived adduct (nitroxide), the e.s.r. spectrum of which is diagnostic of 'R.6$ Specifically, the lifetime of the trichloromethyl radical at ambient temperature is in the millisecond range,8 but its adduct with phenyl-t-butylnitrone has a half-life of many hours, so that its e.s.r.spectrum can be conveniently measured. Spin-trapping e.s.r. spectroscopy has also been successfully applied to study the radical intermediates in the radiolysis of CCl, at room temperat~re.~~ lo Previous e.s.r. studies involved experiments at liquid-nitrogen temperature, where the spectrum was merely an asymmetric broad 1ine.ll Upon heating the irradiated sample more structure was observed near the phase-transition temperature (225 K), but the radicals disappeared by recombination at a temperature at which the resolution was not yet complete.ll We report here on the spin trapping of free radicals formed in liquid CCl, irradiated with low-energy X-rays of different wavelengths, as well as in 0.08 mol dmW3 CH,OH in CCl, and 0.08 mol dm-3 solution of tritiated methanol in CCl,. In the latter case, /<--particles from tritium (mean energy 5.7 keV) served as an internal radiation source.The utilization of X-rays of different wavelengths was aimed at examining the effect of photon energy on the radical yield and the possible 'Auger enhancement' of radical formation (vide infra). Experiment a1 Carbon tetrachloride and methanol, both of spectral purity, were supplied by Merck and used without further purification. a-Phenyl-t-butylnitrone (PBN) was obtained from Aldrich, tritiated [Me-,H]rnethanol (specific activity 70 mCi mmol-l) from NEN, Boston. The samples were degassed and irradiated either with a primary X-ray beam from a Cr or Mo anode tube or secondary (fluoroescent) radiation.In the former case, a Cr tube operated at 15 kV, or a Mo tube at 30 kV, were used, thus providing 219220 Spin Trapping in Irradiated CC1, characteristic line spectra (K,, p 5.47 and 17.39 keV, respectively) superimposed on the continuous spectrum. Dose rates were 5 krad min-l in both cases. The method of generating monoenergetic X-ray beams, based on the utilization of fluorescent radiation, was described previously.12 The primary beam from a Cr anode was impinged at a S or a Sc target at an angle of 45"; the fluorescent radiation emitted perpendicular to the primary beam consisted solely of K, and Ks photons (S E& 2.32 keV; Sc Ea,B 4.12 keV).Dose rates were 275 rad min-l and 460 rad min-l, respectively. Irradiations were carried out at room temperature. The dosimetric measurements were performed using a calibrated ionization chamber (DL4, PTW Freiburg). E.s.r. spectra were recorded at room temperature with a Varian E-9 spectrometer (9.5 GHz) with 100 kHz field modulation frequency, modulation amplitude 0.1 G, and a microwave power of 10 mW. Results Fig. 1 (a) shows the e.s.r. spectrum produced in 0.25 mol dm-, PBN in CCl, irradiated with X-rays of any energy. The structure in this spectrum provides evidence for three nitroxides, namely PhCH(CCl,)NO(CMe,) (a triplet of doublets, aN = 14.0 G, a& = 1.5 G), PhCH(Cl)NO(CMe,) [a triplet of doublets, aN = 12.25 G, a& = 0.75 G, further split to a multiplet (two overlapping quartets) by 35Cl and 37Cl nuclei, = 6.25 G, a(37Cl) = 5.2 GI and benzoyl-t-butylnitroxide PhC(O)NO(CMe,) (a triplet, aN = 7.95 G).These assignments compare favourably with those reported in ref. (9) and (10) for 6oCo y-irradiation. Of these three nitroxides the most stable is the trichloromethyl adduct, and the least stable is the chlorine atom adduct. Upon storage of the irradiated sample, the lines characteristic of the C1-adduct disappear within 2 h [fig. 1 (b)]; after a prolonged storage (> 12 h) only the lines corresponding to the CCl, adduct are visible [fig. l(c)]. In the e.s.r. spectrum of the X-irradiated 0.08 mol dm-, solution of CH,OH in CCl, (fig. 2) lines belonging to the CCl, adduct and to benzoyl-t-butyl nitroxide can be unambiguously distinguished, whereas the lack of the chlorine atom adduct is evident.It was important to verify this, since our intention was to carry out an experiment aimed at the examination of the radical formation due to P-particle radiolysis, in which tritiated methanol was dissolved in CCl,, i.e. serving as an internal radiation source. Fig. 3 shows a typical spectrum observed in such experiments. Shortly after the sample had been prepared no e.s.r. signal was detected even at the highest possible amplification. The splitting pattern in this spectrum will be discussed in the following section. To compare the yields of trichloromethyl radicals from radiolysis by radiations of different quality, the area under the respective lines of the PBN-CCl, adduct, normalized to an absorbed dose of 10 krad, corrected for the different mass absorption coefficients and the decay of the adduct during the experiment, are tabulated (table I), whereas the yield from the sulphur fluorescent X-rays was taken equal to 1 .Discussion Early investigations based on the detection of stable end-products demonstrated that molecular chlorine and hexachloroethane are the sole products of the radiolysis of pure liquid CCl,. The postulated mechanism involved the reactions CCl, h h ~ ) 'CCl, + 'C1 (1) 2 'CCl, -+ C2C1, (2) 2'Cl -+ Cl,. (3) The existence of 'CCl, was indirectly indicated by scavenger experiments, e.g. the formation of CC1,Br in the CCl, + Br, system,13 while chlorine atoms were anticipatedA . Halpern 22 1 4 Fig. 1. (a) E.s.r.spectrum of de-aerated 0.25 mol dm-3 PBN in CCl, subjected to X-rays; (b) same sample, 2 h later; (c) same sample, 2 12 h later. from the yield of Cl,. The results obtained with spin traps like those in fig. 1, which coincide with the results reported by other authorsg. lo for 6oCo y-radiation, are important since they enable the observation of spectra of both radical intermediates at any temperature. In this case the evidence for C1 atoms is clear-cut because of the detection of well resolved 35Cl and 37Cl coupling. The ratio of the splitting constants a(35Cl) = 6.25 G and a(37Cl) = 5.2 G agrees with the ratio of magnetic moments of both nuclei (1.061 and 0.883, re~pectivelyl~). This hyperfine interaction is not visible in the PBN-CCl, adduct in which the three chlorine atoms are in the y-position giving splitting222 Spin Trapping in Irradiated CCI, J - 5G Fig.2. E.s.r. spectrum of de-aerated 0.25 mol dmP3 PBN in CCI, containing CH,OH subjected to X-rays. Fig. 3. E.s.r. spectrum of de-aerated 0.25 mol dmP3 PBN in CC1, containing [Me-,H]methanol. Table 1. Relative yield of PBN-CC1, upon different X-rays, normalized to an equal absorbed dose radiation p/g (CCI,) relative tube target type of radiation K,,p/keV cm2 g-l yield MO - Mo K, and Ka on continuum 17.39 20 0.8 Cr - Cr Kdl and Kp on continuum 5.45 298 1.5 Cr S fluorescent, monoenergetic 2.32 304 1 .o Cr sc fluorescent, monoenergetic 4.12 618 3.5A . Halpern 223 too small to resolve. The benzoyl-t-butylnitroxide (a, x 8 G) has been repeatedly observed upon an oxidation of nit~0nes.l~' l6 In the present case the main oxidising agent was presumably a radiolytically produced chlorine.Note that PBN is capable of trapping neutral radicals 'Cl and 'CCl,, but not their ionic precursors. By nanosecond pulse radiolysis with optical detection the cation CCli+, an ion-pair neutralization complex [CClz . - - Cl-] and an ion C1; were 0bserved.l' [In this context see, however, ref. (10) and (18).] Additional evidence for the CCl&+ cation was obtained by Symons et a l l o in the solid-state e.s.r. studies at 77 K. The striking feature of the spectrum produced in CCl, containing CH,OH (fig. 2) is the lack of the e.s.r. signal from chlorine atoms. Obviously these atoms react very fast with methanol : 'C1+ CH,OH + HCl + 'CH,OH (4) 'CH,OH + CCl, + 'CCl, + HCl + CH,O.( 5 ) It has been recently reportedlg that the presence of a small amount (2 x lop3 mol dm-3) of methanol in CCl, completely suppresses the formation of molecular chlorine upon radiolysis by removing its precursor, viz. Cl atoms. The spectrum in fig. 3 was observed in the sample containing 0.08 mol dm-, solution of tritiated methanol in CCl, stored for 2 days. Since the sensitive part of the cavity housed 3.2 mCi of produces 3.2 x lo1, decays per day, the spectrum was created by ca. 2 x 1013 decay events. These correspond to an absorbed dose (assumed to be the total energy liberated by the desintegrated nuclei) of ca. 3 krad. The more prolonged storage did not cause any change either in the hyperfine structure or in the line intensities, thus indicating that an equilibrium between the radical growth kinetics and the rate of their decay has been established.Note that in experiments based on the decay of tritium, radicals are formed at a constant rate during the entire time of the experiment, because the half-life of tritium (12.3 years) is very long compared with the duration of the experiment. Since it is obviously also much longer than that of any spin adduct, the continuously produced spin adducts and tritium remain in secular equilibrium, similar to a radioactive genetic pair with a long-lived parent and a short-lived daughter nuclide. An intense triplet of doublets characteristic of the PBN-CCl, adduct dominates the spectrum, and a small share of the signal from the hydroxymethyl radicals is also seen, but neither C1 adducts nor the acyl nitroxides were detected. The question arises as to why the PBN-CH,OH adduct has not occurred in the irradiated CC1,-CH,OH system (fig.2). Considering the mass absorption coefficients weighted in proportion to the electron contribution from both compounds making up the solution, X-rays were absorbed practically solely in CCl,. Yet, when tritiated methanol, (CH,T)OH, was present, one hydroxymethyl radical per decay necessarily emerged from the trans- mutation of tritium to helium, followed by its trapping.20 (400 mm3 of the solution), and the decay of 1 mCi of The 'CCl, Yields vs. Photon Energy Although the irradiation of liquid CCl, with the primary X-ray beam from a Cr anode tube appeared to produce twice as much trapped 'CCl, radicals as X-rays from the Mo anode tube (table l), this result might be considered as disputable.Primary beams provide sharp Ka and Ks lines superimposed on the continuous spectrum, the maximum energy of which is dependent on the voltage across the tube (in our case, 15 keV in the Cr irradiation, and 30 keV in the Mo irradiation). Consequently, a large uncertainty rests in the estimation of the absorbed dose, since the calculated mass absorption coefficient (table 1) refer to the & photons, while the radiation was polychromatic. More certain evidence is afforded by the experiments in which monoenergetic X-rays from sulphur and scandium fluorescent targets were utilized. Apparently, the latter produces 3.5 times 8 FAR 1224 Spin Trapping in Irradiated CCl, more trapped 'CCl, radicals than the former, at an equal dose absorbed.We can appreciate why it is so by considering that the weighted average K& X-ray energies are 2.32 keV for sulphur, and 4.12 keV for scandium, i.e. they are on both sides of the K-absorption edge of the absorbing chlorine atoms (2.82 keV). Therefore, Sc X-rays are absorbed predominantly via a photoeffect in the K shell of chlorine, while S X-rays are not, but can only interact, at a significantly lower cross-sections, with the L shell electrons. After a photoelectric absorption, an C1 atom is left with a vacancy in an inner shell which triggers a vacancy cascade across the atomic shells. (The K Auger yield in chlorine is 0.922.) This is consequentia121 in so far as (i) it is accompanied by the emission of the Auger electrons of energy from 190 eV to 2.35 keV and very short ranges in matter and (ii) it creates a mean charge of +3 in the outer shell.Both the internal radiolysis by the Auger electrons and the energy released in charge neutralization, may be capable of enhancing the radical yield after the photoabsorption of Sc X-rays in the K-shell of chlorine. Conclusion The enhanced radiation response to the absorption of monoenergetic X-rays of energy slightly surpassing the binding energy of the inner-shell electrons of the heavy component atom has been previously observed in solid bromodeoxyuridine,12 zinc-containing enzyme,22 and brominated DNA in 24 The present results provide additional evidence for this Auger enhancement effect and demonstrate the spin-trapping e.s.r.technique to be useful in such experiments which can be now based on the estimation of radicals in organic liquids at ambient temperature. References 1 W. J. Brattin, E. A. Glende and R. 0. Recknagel, J. Free Radical Biol. Med., 1985, 1, 27. 2 J. L. Poyer, R. A. Floyd, P. B. McCay, E. G. Janzen and E. R. Davis, Biochim. Biophys. Acta, 1978, 3 P. B. McCay, M. M. King, J. L. Poyer and E. K. Lai, Ann. N.Y. Acad. Sci., 1982, 393, 23. 4 A. Tomasi, E. Albano, K. A. K. Lott and T. F. Slater, FEBS Lett., 1980, 122, 303. 5 E. Albano, K. A. K. Lott, T. F. Slater, A. Stier, M. C. R. Symons and A. Tomasi, Biochem. J., 1982, 6 E. G. Janzen, Free Radical Biol., ed. W. A. Pryor (Academic Press, New York, 1980), vol.4. 7 M. J. Perkins, Adv. Phys. Org. Chem., 1980, 17, 1. 8 H. Paul, Int. J. Chem. Kinet., 1979, 11, 495. 9 V. N. Belevskij and U. Vestfal, Moscow Univ. Chem. Bull., 1981, 36, 41. 539, 402. 204, 593. 10 M. C. R. Symons, E. Albano, T. F. Slater and A. Tomasi, J. Chem. Soc., Faraday Trans. I , 1982, 78, 11 N. Leray and J. Roncin, J. Chem. Phys., 1965, 42, 800. 12 A. Halpern and G. Stocklin, Radiat. Res., 1974, 58, 329. 13 F. A. Abramson, B. M. Buckhold and R. A. Firestone, J. Am. Chem. SOC., 1962, 84, 2285. 14 G. H. Fuller, J. Phys. Chem. Ref Data, 1976, 5, 835. 15 E. G. Janzen and B. J. Blackburn, J. Am. Chem. SOC., 1969,91,4481. 16 A. L. Bluhm and J. Weinstein, J. Am. Chem. SOC., 1970, 92, 1444. 17 0. Brede, J. Bos and R. Mehnert, Ber. Bunsanges. Phys. Chem., 1980, 84, 63. 18 Y. Tabata, H. Kobayashi, M. Washio, S. Tagawa, Y. Yoshida, Radiat. Phys. Chem., 1985, 26, 473. 19 G. V. Kovalev, A. L. Karasev, L. T. Bugaenko and E. P. Kalyazin, High Energy Chem., 1984,18,245. 20 A. Halpern, Chem. Phys. Lett., 1984, 103, 523. 21 A. Halpern, in Uses of Synchrotron Radiation in Bidogy, ed. H. B. Stuhrmann, (Academic Press, 22 B. Diehn, A. Halpern and G. Stocklin, J. Am. Chem. SOC., 1976,98, 1077. 23 A. Halpern and B. Mutze, Int. J. Radiat. Biol., 1978, 34, 67. 24 K. Shinohara, H. Ohara, K. Kobayashi, H. Mezava, K. Hieda, S. Okada and T. Ito, J. Radiat. Res., 2205. London, 1982). 1985, 26, 334. Paper 61 1 130; Received 5th June, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300219

出版商:RSC    

年代:1987

数据来源: RSC

摘要:

J . Chem. SOC., Faraday Trans. 1, 1987, 83, 225-230 High-performance Liquid Chromatography-Electron Spin Resonance Analysis of Sugars irradiated in the Solid and Liquid Phase Jacques J. Raffi,* Patrice B. Vincent, Jean-Pierre L. Agnel, Christine M. Battesti and Corinne L. Thie'ry Laboratoire de Radiochimie des Constituants des Aliments, DBlSRA, CEN Cadarache, 13108, Saint-Paul-lez-Durance, France Radicals, produced by y-irradiation of sugars in the solid state at different temperatures and in aqueous solution, have been studied by the spin- trapping method. High-performance liquid chromatography has been used to separate the nitroxides formed and hence simplify the tota! observed electron spin resonance spectra. Several correlations between radicals formed in cc-and 8-glucose, methyl-1-a-glucose, maltose, sucrose and related compounds are presented. In order to understand the y-radiolysis of saccharides in the powder state, experiments have been carried out using the spin-trapping method? Glucose and its oligomers were chosen as models for starches.293 We converted the sugar radicals formed in the polycrystalline state into long-lived nitroxide spin adducts in the liquid phase (water- ethanol ratio 2: 1) by reaction with 2-methyl-2-nitrosopropane (MNP or But-N=O): S' + But--N=O + But-N-S.I 0' Because the action of y-radiation is relatively unselective, numerous nitroxides, with only slightly different aN and aH hyperfine coupling constants and very similar g-factors, are produced, making spectra complex. High-performance liquid chromatography (h.p.1.c.) has been used to separate this mixture of sugar-nitroxide radical^.^-^ Experiment a1 The main sugars studied here (a- and P-glucose, methyl-1-a-glucose, maltose and sucrose) were obtained from the Fluka, Merck and Prolabo Companies; MNP was obtained from Janssen Chimica.The sugars were irradiated in a 13'Cs cell, supplying a dose rate of 3.6 kGy h-l. The doses (20 kGy for irradiation in the powder state or 1 kGy for liquid-phase irradiation) were chosen in order to obtain suitably intense e.s.r. signals. The sugar (200 mg), irradiated in the solid phase, was added to a solution of 2.4 cm3 MNP [12 mg in deoxygenated ethanol (0.8 cm3) added 1 min before trapping of deoxygenated water (1.6 cm3)]. In the case of irradiation in the liquid phase, to prevent the formation of radicals derived from ethanol, the sugar was dissolved in an aqueous solution of MNP (stirred overnight), irradiated, and then added to ethanol.All volumes of solvents and weights of sugars and MNP were calculated in order to obtain sugar, MNP and solvent final concentrations the same as those used in trapping of sugars irradiated in solid state. The nitroxide solution was immediately injected onto a Waters semi-preparative C 18 225 8-2226 H. P.L.C.-E.S.R. Analysis of Irradiated Sugars Fig. 1. A typical 20 kGy; (a) 10- is 70 G and the h.p.l.c.-e.s.r. experiment: P-glucose irradiated in the solid phase at - 196 "C and .I1 min; (b) 1 4 1 5 min. The fractions are collected each minute. The scan range arrow corresponds to g = 2.0059; the e.s.r. amplification is the same for each series of spectra.micro-Bondapak column, cooled to 7 "C, in order to increase the radicals' stability. The water-ethanol eluent was delivered at a flow rate of 1 cm3 min-l and a Bruker 200 D 10 e.s.r. spectrometer was used as a detector. Collected fractions (one per minute) were cooled in an ice bath until the spectra were recorded6 (fig. 1). Each fraction contains no more than four radicals, rather than the 10-1 5 in the solution before h.p.1.c. This facilitates the simulations' of the relative e.s.r. spectra, performed on a Bruker Aspect 2000 or Hewlett Packard HP 150 computer. The time parameter is useful for spectral analysis. For instance, in the case of the solid-phase irradiation ofJ . J . Rafi, P .B. Vincent, J-P. L. Agnel, C. M. Battesti and C. L. Thikry 227 Fig. 2. Example of simulation: case of the '9-10 min' fraction of h.p.1.c. ofp-glucose in nitroxide solution. (a) Experimental spectrum, recorded just after h.p.1.c. ; (b) experimental spectrum recorded 73 min later; (c) simulation of spectrum (6); ( d ) difference spectrum: (ak(b); (e) simulation of spectrum ( d ) . /$glucose at - 196 "C, the '9-10 min' fraction was analysed by a comparison of two spectra (fig. 2); the initial one and one recorded 73 min later. The simulation of the delayed spectrum and of the computed difference spectrum gives the hyperfine constants of the following three species: a N = 15.1 G ; aH = 0.45 G aN = 15.1 G; aH1 = 2.55 and aHz = 0.4 G(4 protons) UN = 15 G ; a H = 1 G.Results and Discussion The different signals recorded for the five sugars studied are compared in fig. 3. Their e.s.r. hyperfine coupling constants, retention times (tR) and stabilities are reported as a function of the irradiation conditions, i.e. at temperatures of - 196, 25 and 100 "C for the solid phase, and ca. 1 "C for the liquid phase. From these data the following228 H. P.L.C.-E.S.R. Analysis of Irradiated Sugars 16 aN 15 14 a - - 16 aN 15 14 A 4 0 L l 1 a 1 I 1 , . I a I 0 5 10 a H Fig. 3. Radici 1s observed in the case of (a) a-glucose, (b) P-glucose, (c) a-methyl-1-glucose, ( d ) maltose and (e, sucrose. The hyperfine constants are measured in G. The irradiation is performed in the solid state at - 196 "C (O), +25 "C (0) and 100 "C (A) or in water at ca.1 "C (A). conclusions may be drawn. (a) The four types of irradiation conditions used do not lead to completely different radicals. (b) Of the radicals found in the solid phase at - 196 "C, most are also recorded in the liquid phase. Hence the radicals formed in aqueous solutions are those most easily formed in the solid state. (c) The radicals formed in the solid phase at 100 "C are the most stable of those produced at room temperature.J. J . RaB, P . B. Vincent, J-P. L. Agnel, C. M . Battesti and C. L. Thiiry Table 1. Specific parameters of the five defined types of radicals 229 alG beG alMeG 3MeG Man MeMan alG beG alMeG Ma1 Gal XYI alG beG alMeG Ma1 SUC G3 CyDex Fru Gal Man MeMan alG alMeG Ma1 SUC alG alMeG Ma1 SUC 15.45 15.45 15.3 15.4 15.5 15.4 15.5 15.4 15.5 15.5 15.5 15.5 15.1 15.1 14.8 14.7 15.2 14.7 15.1 15.2 15 15.1 14.9 14.3 14.3 14.3 14.2 14.2 14.25 14.1 14.25 type 1 4.7 4.4 4.5 4.7 4.7 4.2 1.9 2.1 1.9 1.9 1.9 2 1.15 1 1.2 1.5 1.3 1.1 1.1 1.1 1.2 1.1 1 5.8 and 0.4 5.8 and 0.4 6.1 and 0.4 5.8 and 0.4 2.1 and 0.7 2 and 0.8 1.9 and 0.8 2 and 0.7 type I1 type I11 type IV type v 11-16 12-14 12-17 16-20 15-18 17-23 1620 16-19 15-19 10-17 8-14 9-1 7 9-1 5 14-22 12-22 15-17 13-15 14-20 14-22 15-17 13-1 5 irradiation solid phase -196°C 25 "C + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + liquid phase + + + + + + + + + + + + + + + + ~ a Notation: alG = a-glucose; alMeG = methyl-1-a-glucose; beG = p-glucose; CyDex = cyclo- dextrin; Fru = fructose; G3: maltotriose; Gal = galactose; Ma1 = maltose; Man = mannose; 3MeG = 3-methylglucose; MeMan = 1-methylmannose; Suc = sucrose; Xyl = xylose.Moreover, they are found in greater quantities at 100 "C than at room temperature (no radical is produced by a 100 "C thermolysis). If we now consider the radicals formed from different sugars (table l), several correlations appear which define at least five types of radical. (a) The (I) radicals, with aN z 15.3-15.45 G and aH = 4.4-4.7 G, relatively unstable. They are generated in the liquid phase and solid phases (at - 196 and 25 "C) and have a retention time of 12-1 5 min230 I3.P.L.C.-E.S.R. Analysis of Irradiated Sugars (found in a- and 8-glucose, methyl-l-a- and 8-glucose 3-methyl-glucose, mannose and methylmannose). They have been assigned to the following part structure, (I) = 'C(6)-H(OH) I -C( 9-0- in the case of a-glucose,6 after specific 13C labelling on C(6).In the case of maltose and sucrose the assignment requires confirmation. (b) The (11) radicals, with aN z 15.4-15.5 G and aH z 1.9-2.1 G, are produced under all irradiation conditions, but in variable quantities. Since they have the same retention time (16-20 min) as the ButNCH(OH)CH, radical I 0' they are assigned to the following structure; ButNCH(OH)R. I 0' (c) The (111) radicals, with aN = 14.7-1 5.1 G and aH z 1-1.2 G, are less stable and have a retention time of 11-16 min. They are produced in the liquid and solid phases (- 196 and 25 "C) in many sugars but cannot be structurally assigned. ( d ) The structures of two other types of radicals, (IV) and (V), are still unknown.They are both relatively long-lived and are only produced in the solid phase at 25 or 100 "C from a-glucose (crystallized with one water molecule) and methyl-l-glucose, maltose or sucrose. They are characterized by a retention time of 15-20 min and the following coupling constants : (IV): aN z 14.3 G, aH1 5.8-6.1 G, aH2 x 0.4 G (V): ax z 14.2 G, aH1 z 2 G, aH2 x 0.7-0.8 G. In conclusion, y-irradiation of sugars leads to many radical, e.g., 52 spin adducts are generated from the five sugars studied here. Consequently, a complete interpretation will be very difficult. It will require the study of other sugars, the use of several specifically labelled sugars and a careful simulation work on recorded spectra. C.T. is indebted to the Institut National de la Sante et de la Recherche Medicale for financial support. We thank Dr J. Thiery for helping us in simulation work. References 1 E. Janzen, Acc. Chem. Res., 1971, 4, 31, 2 J. Raffi, J-P. Agnel, C. Boizot, C. Thitry and P. Vincent, Staerke, 1985, 7, 228. 3 C. ThiCry, J-P. Agnel, C. Frejaville and J. Raffi, J. Phys. Chem., 1983, 87, 4485. 4 N. Suzuki, K. Makino, F. Moriya, S. Rokushika and H. Hatano, J. Phys. Chem., 1981, 85, 263 and 5 H. Stronks, E. Janzen and J. Weber, Anal. Lett., 1984, 17, 321. 6 J-P. Agnel, C. ThiCry, C. Battesti, P. Vincent and J. Raffi, Anal. Lett., 1985, 18, 1013. 7 J. M. Thiiry, Voyons, in Logiciels pour la chimie, (SOC. Fr. Chim. Paris, and Ass. Nat. Logiciel, Nancy, references cited therein. 1985), p. 156. Paper 6/ 1003; Received 22nd May, 1986

ISSN:0300-9599    

DOI:10.1039/F19878300225

出版商:RSC    

年代:1987

数据来源: RSC