Kinetic Studies on Stilbazulenyl-bis-nitrone (STAZN), a Nonphenolic Chain-Breaking Antioxidant in Solution, Micelles, and Lipid Membranes Subhash C. Mojumdar,† David A. Becker,‡ Gino A. DiLabio,§ James J. Ley,‡ L. Ross C. Barclay,*,† and K. U. Ingold| Contribution from the Departments of Chemistry, Mount Allison University, Sackville, NB, E4L 1G8 Canada, Florida International University, Miami, Florida 33199, the National Institute for Nanotechnology, National Research Council, Edmonton, AB, T6G 2V4 Canada, and the National Research Council, Ottawa, ON, K1A OR6 Canada
[email protected] Received December 29, 2003
The rate constants, kinh, for reaction of stilbazulenyl-bis-nitrone (STAZN, 1) with peroxyl radicals and the number of radicals trapped, n, are compared with those of phenolic antioxidants 2,2,5,7,8pentamethyl-6-hydroxychroman (PMHC, 4a), 2,5,7,8-tetramethyl-6-hydroxychroman-2-carboxylic acid (Trolox, 4b), and 2,6-di-tert-butyl-4-methoxyphenol (DBHA, 5). The behavior of STAZN depended markedly on the media and type of initiator used, water-soluble or lipid-soluble. In styrene/ chlorobenzene and initiation by azo-bis(isobutyronitrile) (AIBN), kinh (STAZN) ) 0.64 kinh (5) ) 0.02kinh (4a). On addition of methanol, the kinh of STAZN increased 6-fold to be four times that of 5 while that of 4a decreased 6-fold. In aqueous SDS-micelles containing methyl linoleate and initiation with water-soluble azo-bis(amidinopropane)2HCl, ABAP, the relative kinh values were 1 g 4b > 5. In dilinoleoylphosphatidyl choline (DLPC) bilayers and initiation with lipid-soluble azobis-2,4(dimethylvaleronitrile) (DMVN), the kinh order was 5 > 4b > 1. During initiation with ABAP in micelles and bilayers, the calculated values of kinh for STAZN changed during the induction period. The experimental results are interpreted in terms of the conformation of STAZN, which is transoid in homogeneous solution but cisoid in aqueous dispersions of lipids. In such dispersions, the STAZN lies at the lipid-water interface where it traps water-soluble peroxyl radicals by a single electron-transfer mechanism. The cisoid conformation at lipid-water interfaces is supported by theoretical calculations. Introduction Phenolic antioxidants, ArOH, inhibit lipid peroxidation by trapping the chain-carrying peroxyl radicals, ROO•, by a hydrogen atom transfer (HAT) process1 (reaction 1):
ROO• + ArOH f ROOH + ArO•
(1)
The phenoxyl radical, ArO•, captures a second peroxyl radical (reaction 2):
ROO• + ArO• f nonradical products
(2)
Thus each molecule of phenol breaks two peroxidation chains and the stoichiometric factor, n, for inhibition by * To whom correspondence should be addressed. Fax: (506) 3642313. † Mount Allison University. ‡ Florida International University. § National Institute for Nanotechnology. | National Research Council. (1) (a) Ingold, K. U.; Howard, J. A. Nature 1962, 195, 280-281. (b) Howard, J. A.; Ingold, K. U. Can. J. Chem. 1962, 40, 1851-1864. (c) Burton, G. W.; Doba, T.; Gabe, E. J.; Hughes, L.; Lee, F. L.; Prasad, L.; Ingold, K. U. J. Am. Chem. Soc. 1985, 107, 7053-7065. (d) Burton, G. W.; Ingold, K. U. Acc. Chem. Res. 1986, 19, 194-201.
ArOH is 2.0. The radical trapping and hence antioxidant activities of ArOH are very significantly reduced in polar, hydrogen bond accepting (HBA) solvents. This is because ArOH molecules that are hydrogen bonded (HB) to solvent (S) molecules are unreactive toward radicals (for steric reasons).2 ROO•
ROOH + ArO• 79 ROO•
ArOH + S a ArOH‚‚‚S N no reaction (3) The observed rate constants for HAT from phenolic antioxidants to radicals are decreased in aqueous disper(2) (a) Avila, D. V.; Ingold, K. U.; Lusztyk, J.; Green, W. H.; Procopio, D. R. J. Am. Chem. Soc. 1995, 117, 2929-2930. (b) Valgimigli, L.; Banks, J. T.; Ingold, K. U.; Lusztyk, J. J. Am. Chem. Soc. 1995, 117, 9966-9971. (c) MacFaul, P. A.; Ingold, K. U.; Lusztyk, J. J. Org. Chem. 1996, 61, 1316-1321. (d) Valimigli, L.; Ingold, K. U.; Lusztyk, J. J. Am. Chem. Soc. 1996, 118, 3545-3549. (e) Barclay, L. R. C.; Edwards, C. D.; Vinqvist, M. R. J. Am. Chem. Soc. 1999, 121, 6226-6231. (f) Valgimigli, L.; Banks, J. T.; Lusztyk, J.; Ingold, K. U. J. Org. Chem. 1999, 64, 3381-3383. (g) Snelgrove, D. W.; Lusztyk, J.; Banks, J. T.; Mulder, P.; Ingold, K. U. J. Am. Chem. Soc. 2001, 123, 469-477. (h) Foti, M. C.; Barclay, L. R. C.; Ingold, K. U. J. Am. Chem. Soc. 2002, 124, 12881-12888. (i) Litwinienko, G.; Ingold, K. U. J. Org. Chem. 2003, 68, 3433-3438.
10.1021/jo030390i CCC: $27.50 © 2004 American Chemical Society
Published on Web 04/06/2004
J. Org. Chem. 2004, 69, 2929-2936
2929
Mojumdar et al.
(STAZN) has been shown by NMR measurements of the rates of formation of cumene hydroperoxide during the autoxidation of cumene to be a very effective antioxidant and moreover, in contrast to ArOH, to be a better antioxidant in the polar solvent mixture cumene/methanol (80:20) than in cumene/benzene (80:20).5 Such unusual antioxidant behavior warranted the present quantitative measurements of the rate constants for peroxyl radical trapping by STAZN and the stoichiometric factors for this reaction in media ranging from homogeneous solutions to aqueous dispersions of micelles and phospholipid bilayers. Computational modeling of STAZN conformers provides an insight into its antioxidant behavior in bilayers. Results
FIGURE 1. Structures of the antioxidants used in this study.
sions of lipids in bilayers and micelles as a result of both hydrogen bonding of the phenolic hydroxyl group to H-bond acceptors in these systems and physical restrictions that make some of the antioxidant “inaccessible to the attacking radical”.2d,3 In aqueous lipid dispersions (and in HBA solvents generally) it would clearly be advantageous to have radical trapping antioxidants that were not HB donors (HBD) and that functioned by some mechanism other than HAT. One obvious approach to this goal would be to employ a “spin-trap”. That is, in many systems, particularly biological systems, the steady-state concentration of transient free radicals, X•, is too low for their direct detection by electron paramagnetic resonance (EPR) spectroscopy. Spin-traps overcome this problem as a result of their ability to react with many transient radicals and generate a persistent spin-adduct radical, the concentration of which builds up sufficiently for its detection by EPR, and often the identity of X• can also be determined. One of the most popular and useful spintraps is phenyl-tert-butyl nitrone (PBN):
1. Rate Constants and Stoichiometric Factors. The structures of STAZN, two products of its oxidation and three phenolic antioxidants that were employed for comparative purposes are shown in Figure 1. Rate Constants in Homogeneous Solution. The azo-bis(isobutyronitrile) (AIBN)-initiated autoxidation of styrene in chlorobenzene has been employed for many years to determine the rate constants for peroxyl radical trapping by phenols.1b,c,6 The relevant reactions are O2
∆
[(CH3)2C(CN)Nd]2 98 98 rOO• (rate ) Ri)
(5)
rOO• + CH2dCHPh f rOOCH2C•HPh
(6)
rOOCH2C•HPh + O2 f rOOCH2CH(OO•)Ph () ROO•) (7) kp
ROO• + CH2dCHPh 98 ROOCH2C•HPh kinh
ROO• + ArOH 98 ROOH + ArO• fast
ROO• + ArO• 98 nonradical products
(8) (1) (2)
The length of time the oxidation of styrene is suppressed, i.e., the induction period, is given by
τ)
n[ArOH] Ri
(9)
where n is normally 2.0 for phenols.1c,d,6c,7 During the induction period the rate of oxygen uptake is given by
Unfortunately, and despite claims to the contrary, one of us has shown that PBN is not an active peroxyl radical trapping antioxidant.4 PBN acts only as a moderate “retarder” of oxidation at relatively high concentrations.4 In stark contrast to PBN, stilbazulenyl-bis-nitrone (3) (a) Barclay, L. R. C.; Baskin, K. A.; Dakin, K. A.; Locke, S. J.; Vinqvist, M. R. Can. J. Chem. 1990, 68, 2258-2269. (b) Barclay, L. R. C.; Edwards, C. D.; Mukai, K.; Egawa, Y.; Nishi, T. J. Org. Chem. 1995, 60, 2739-2744. (c) Castle, L.; Perkins, M. J. J. Am. Chem. Soc. 1986, 108, 6381-6382. (4) Barclay, L. R. C.; Vinqvist, M. R. Free Radical Biol. Med. 2000, 28, 1079-1090.
2930 J. Org. Chem., Vol. 69, No. 9, 2004
(
)
-d[O2] dt
inh
)
kp[PhCdCH2]Ri nkinh[ArOH]
(10)
Values of kinh are calculated using the integrated form of eq 10, i.e., (5) Becker, D. A.; Ley, J. J.; Echegoyen, L.; Alvarado, R. J. Am. Chem. Soc. 2002, 124, 4678-4684. (6) (a) Howard, J. A.; Ingold, K. U. Can. J. Chem. 1963, 41, 17441751. (b) Howard, J. A.; Ingold, K. U. Can. J. Chem. 1963, 41, 28002806. (c) Burton, G. W.; Ingold, K. U. J. Am. Chem. Soc. 1981, 103, 6472-6477. (7) Horswill, E. C.; Howard, J. A.; Ingold, K. U. Can. J. Chem. 1966, 44, 985-991.
Kinetic Studies on Stilbazulenyl-bis-nitrone
-∆[O2]t )
(
kp[PhCHdCH2] ln 1 kinh
t τ
)
(11)
A plot of ∆[O2]t vs ln(1 - t/τ) yields a straight line with a slope equal to kp[PhCHdCH2]/kinh, from which kinh is obtained. Oxygen uptake traces for the AIBN-initiated autoxidation of styrene in chlorobenzene at 37 °C and 760 Torr air are shown for samples containing STAZN (Figure 2A) and 2,6-di-tert-butyl-4-methoxyphenol (DBHA) (Figure 2B). It is clear that STAZN, like DBHA, is an excellent peroxyl radical trapping antioxidant.STAZN becomes an even more effective inhibitor of styrene autoxidation upon the addition of 5 vol % methanol to the system (Figure 2C), with kinh increasing 6-fold to 48 × 104 M-1 s-1 (Table 1). Methanol had no significant effect on kinh for DBHA, presumably because the phenolic hydroxyl group in this compound is sterically too well protected for a significant fraction of this phenol to HB with 5% methanol.8,9 However, with the less sterically well protected 2,2,5,7,8pentamethyl-6-hydroxychroman (PMHC, a vitamin E analogue), the same concentration of methanol induced a 6-fold decrease in kinh (Table 1) because of HB formation (see reaction 3). The stoichiometric factor for STAZN, like that for DBHA and PMHC, is 2.0 both in chlorobenzene and in chlorobenzene/methanol (Table 1). The monoaldehyde STAZN oxidation product, 2, did not give well-defined induction periods in the styrene autoxidation system. However, cumene has such a low kp (0.18 M-1 s-1 at 30 °C10) that even relatively poor antioxidants give well-defined induction periods.7 Using the AIBN-initiated autoxidation of cumene in chlorobenzene the stoichiometric factors of STAZN and 2 were found to be 2 and 1, respectively. The dialdehyde, 3, did not inhibit the autoxidation of cumene. Rate Constants in Micelles. Sonication of solid STAZN with 0.5 M SDS in phosphate buffer, to solubilize it, was slow (∼1 h) and caused its partial decomposition (as judged by the UV-vis spectrum of the sonicated material). Coevaporation of STAZN and SDS from methanol followed by dispersion in buffer gave undecomposed STAZN/SDS micelles. These were injected into solutions of SDS micelles containing methyl linoleate as the peroxidizable substrate.11 Lipid peroxidation was initiated at 37 °C using mainly the water-soluble initiator azo-bis(amidinopropane) dihydrochloride, ABAP, and to a lesser extent the lipid-soluble initiator azo-bis(2,4-dimethylvaleronitrile), DMVN (which was incorporated into the micelles by coevaporation with SDS from methanol). STAZN inhibited the ABAP-initiated peroxidation of linoleate/SDS micelles (Figure 3A), but plots according to eq 11 were somewhat curved (Figure 3A inset, cor(8) 2,6-Di-tert-butylphenols are weak HBDs,2i,9 and intermolecular H-bonds for these phenols have been observed in only a few neat solvents.2i,9 The kinetic solvent effect for DBHA in methanol(5%)/ chlorobenzene was not observed in the present kinetic study. (9) Bellamy, L. J.; Williams, R. L. Proc. R. Soc. London, Ser. A 1960, 254, 119-128; Ingold, K. U.; Taylor, D. R. Can. J. Chem. 1961, 39, 481-487; Wawer, I.; Kecki, Z. Ber. Bunsen-Ges. 1976, 80, 522-525. Litwinienko, G.; Megid, E.; Wojnicz, M. Org. Lett. 2002, 4, 2425-2428. (10) Howard, J. A.; Ingold, K. U. Can. J. Chem. 1965, 43, 27292736. (11) Xi, F.; Barclay, L. R. C. Can. J. Chem. 1998, 76, 171-182; Barclay, L. R. C.; Crowe, E.; Edwards, C. D. Lipids 1997, 32, 237245.
FIGURE 2. Oxygen uptake traces for the inhibited oxidation of styrene under 760 Torr air in chlorobenzene, 2.21 mL, initiated with azo-bis(isobutyronitrile), AIBN, 52.5 µmol. (A) Styrene, 4.27 mmol, inhibited by STAZN, 3.39 × 10-8 mol, at 37 °C, Ri ) 7.86 × 10-9 M s-1. Inset: plot according to eq 11 using data from curve A. (B) Styrene, 4.32 mmol, inhibited by 2,6-di-tert-butyl-4-methoxyphenol, DBHA, 6.60 × 10-8 mol, at 30 °C, Ri ) 1.01 × 10-8 M s-1. (C) Styrene, 4.32 mmol in chlorobenzene containing 3.08 mmol (5 vol %) methanol, inhibited by STAZN, 2.03 × 10-8 mol, at 37 °C, Ri ) 7.27 × 10-9 M s-1.
TABLE 1. Rate Constants for Inhibition, kinh, and Stoichiometric Factors, n, of STAZN, DBHA, and PMHC for Inhibited Oxidation of Styrene Initiated with AIBN inhibitora
kinhb (M-1 s-1 × 10-4)
STAZN PMHC DBHA
I. In Chlorobenzene 7.6 ( 0.4 380e 11.9 ( 0.8
STAZN PMHC DBHA
nc 2.0d 2.0 2.0
II. In Chlorobenzene/Methanol (1.24 M) 48 ( 2 1.95 ( 0.02d 65 2.0 11.6 ( 1.3 2.0
a Experiments using STAZN or PMHC were done at 37 °C, those with DBHA at 30 °C. b Results from at least three experiments except for PMHC in II (two experiments). c The stoichiometric factor for almost all phenols is 2.0; see, e.g., refs 1c,d, 6c, 7. d The stoichiometric factor for STAZN was calculated from n ) Ri‚τ/ [STAZN] where Ri ) 2[ArOH]/τ. e Reference 1c.
relation coefficient, R2 ) 0.97) compared with those for oxidations inhibited with Trolox (R2 ) 1.0) or DBHA (R2 ) 0.99). The results obtained using either a pressure transducer or an oxygen electrode to monitor oxygen uptake in the STAZN-inhibited reactions are in fairly good agreement (Table 2). The mean value for kinh (3.7 × 104 M-1 s-1) is comparable to that found for the watersoluble antioxidant, Trolox, and higher than that for the lipid-soluble DBHA. The low stoichiometric factor for STAZN (1.4 ( 0.2) is probably a consequence of the physical location and mechanism of action of this bisnitrone (see Discussion). A typical oxygen uptake trace for a STAZN-inhibited, DMVN-initiated SDS/linoleate peroxidation is shown in Figure 3B. In these systems STAZN is less efficient in suppressing oxidation than in the ABAP-initiated reacJ. Org. Chem, Vol. 69, No. 9, 2004 2931
Mojumdar et al.
FIGURE 3. Oxygen uptake traces for the inhibited oxidation of methyl linoleate (ML) at 37 °C under 760 Torr air in SDS micelles in phosphate buffer, 2.00 mL, pH 7.4. (A) ML, 7.54 × 10-5 mol, initiated with azo-bis(amidinopropane)2HCl, 26.3 µmol, and inhibited by STAZN, 2.11 × 10-8 mol, Ri ) 3.93 × 10-8 M s-1. Inset: plot according to eq 11 using the data from curve A. Curve (B) ML, 7.54 × 10-5 mol, initiated with azobis(2,4-dimethylvaleronitrile), DMVN, 4.1 µmol, and inhibited by STAZN, 2.23 × 10-8 mol, Ri ) 9.89 × 10-9 M s-1. Inset: plot according to eq 11 of the data from curve B.
TABLE 2. Rate Constants for Inhibition, kinh, and Stoichiometric Factors, n, of STAZN, DBHA, and Trolox for the Inhibited Oxidation of Methyl Linoleate in 0.50 M SDS Micelles Initiated with ABAP or DMVN at 37 °C inhibitor
kinha (M-1 s-1 × 10-4)
PTcSTAZN PTcDBHA PTcTrolox OEcSTAZN
I. Initiator, ABAP 4.32 ( 0.02 2.7 3.9 3.5
PTcSTAZN
II. Initiator, DMVN 0.91
nb
of dilinoleoylphosphatidylcholine, DLPC, 63.7 µmol, unilamellar vesicles, ULVs, in phosphate buffer at 37 °C under 760 Torr air, 2.00 mL, pH 7.4. (A) Initiated with ABAP, 27.5 µmol, and inhibited by STAZN, 11.2 × 10-8 mol, Ri ) 3.94 × 10-7 M s-1. Inset: plot according to eq 11 of the data from curve A. (B) DLPC ULVs initiated with ABAP, 27.5 µmol, and inhibited by 2,5,7,8-tetramethyl-6-hydroxychroman-2-carboxylic acid, Trolox, 2.43 × 10-8 mol, Ri ) 3.99 × 10-7 M s-1. Inset: plot according to eq 11 of data from curve B.
TABLE 3. Range of Inhibition Rate Constants, kinh, and Stoichiometric Factors, n, for STAZN and Trolox for Inhibited Oxidation of DLPC Bilayers Initiated with ABAP at 37 °C inhibitor
kinha (M-1 s-1 × 10-4) I.
1.2 ( 0.1d 2.0 2.0 1.6d 0.7d
a Results from two experiments except for the first entry (five experiments). b See footnote c, Table 1. c Oxygen uptake was measured using either a pressure transducer system (PT) or an oxygen electrode (OE). d See footnote d, Table 1.
tions, but the plots according to eq 11 are almost linear (R2 ) 0.997). The kinh and n values are significantly lower than with ABAP initiation (see Table 2). Rate Constants in Phospholipid Bilayers. (i) ABAP-Initiated Peroxidation of Dilinoleoylphosphatidylcholine (DLPC). Multilamellar vesicles (MLVs) of DLPC containing STAZN were prepared by coevaporation of STAZN and DLPC from methanol followed by vortex stirring in phosphate buffer and freeze-thaw cycles in liquid nitrogen under argon. STAZN is a very effective antioxidant during the first part of the induction period, which is not well defined, the oxidation rate slowly increasing to that for the uninhibited reaction (Figure 4A). Plots according to eq 11 were curved (Figure 4A inset), and hence single kinh values could not be obtained. At the start of the reaction chain lengths were very low, there was little peroxidation, and “initial” kinh values were comparable to those found for Trolox. Trolox, 2932 J. Org. Chem., Vol. 69, No. 9, 2004
FIGURE 4. Oxygen uptake traces for the inhibited oxidation
STAZN Trolox STAZN Trolox
MLV,c
DLPC 2.8-1.5 3.7
II. ULV,c DLPC 2.6-0.8 2.7 ( 0.2
nb 2.0 ( 0.1d 2.0 1.6 ( 0.1d 2.0
a Range of k inh values calculated from the start of the inhibition period. Results from at least three experiments except for the second entry (two experiments). b See footnote c, Table 1. c MLV or ULV, multilamellar or unilamellar vesicles of DLPC, 63.7 µmol. d See footnote d, Table 1.
however, gave well-defined induction periods and linear plots according to eq 11 (Figure 4B and inset). Table 3 contains the range of STAZN kinh values calculated throughout the induction period. Nonideal antioxidant kinetic behavior has previously been observed in ABAP-initiated peroxidation of DLPC MLVs containing R-tocopherol and was attributed to inhomogeneous distribution of the antioxidant and initiator.12 In the hope of obtaining a more homogeneous distribution, the STAZN-containing MLVs were converted to small unilamellar vesicles, ULVs, by the extrusion method,13 but these showed the same oxygen (12) Doba, T.; Burton, G. W.; Ingold, K. U. Biochim. Biophys. Acta. 1985, 835, 298-303. (13) (a) Hope, M. J.; Nayar, R.; Mayer, L. D.; Cullis, P. R. In Liposome Technology; CRC Press: Boca Raton, FL, 1992; Vol. 1. (b) Barclay, L. R. C.; Cameron, R. C.; Forrest, B. J.; Locke, S. J.; Nigam, R.; Vinqvist, M. R. Biochim. Biophys. Acta 1990, 1047, 255-263.
Kinetic Studies on Stilbazulenyl-bis-nitrone TABLE 4. Inhibition Rate Constants, kinh, and Stoichiometric Factors, n, of STAZN, Trolox, and DBHA for Oxidation of DLPCa Bilayers Initiated with DMVN at 37 °C inhibitor
kinh (M-1 s-1 × 10-4)
nb
STAZN Trolox DBHA
0.39 0.99 2.1
0.9c 2.0 2.0
a The DLPC, 61.8 µmol, was used as multilamellar vesicles in experiments with STAZN and Trolox; unilamellar vesicles with DBHA. The DBHA was added by liposomal transfer from dimyristoylphosphatidylcholine vesicles.15 b See footnote c, Table 1. c See footnote d, Table 1.
FIGURE 5. Oxygen uptake traces for inhibited oxidation of DLPC, 61.8 µmol, in phosphate buffer at 37 °C under 760 Torr air, 2.00 mL, pH 7.4. (A) DLPC MLVs initiated with DMVN, 27.5 µmol, and inhibited by STAZN, 1.12 × 10-7 mol, Ri ) 4.82 × 10-7 M s-1. Inset: plot according to eq 11 of data from curve A. (B) DLPC MLVs initiated with DMVN, 8.20 µmol, and inhibited by Trolox, 4.58 × 10-9 mol, Ri ) 2.43 × 10-8 M s-1. (C) DLPC ULVs initiated with DMVN, 8.20 µmol and inhibited by DBHA, 6.55 × 10-8 mol, added by liposomal transfer from dimyristoylphosphatidylcholine bilayers (DMPC), Ri ) 2.03 × 10-7 M s-1.
uptake profiles as the MLVs. Despite the ill-defined induction periods it was possible to estimate approximate stoichiometric factors for STAZN of about 2 and 1.6 for the MLVs and ULVs, respectively (see Table 3). (ii) DMVN Initiated Peroxidation of DLPC. In these experiments the initiator and STAZN were incorporated into the lipid by coevaporation of both from methanol followed by the usual conversion to multilamellar vesicles by freeze-thaw cycles under argon. To minimize the error involved during equilibration at 37 °C, the amount of STAZN and initiator used were such as to give relatively long inhibition periods, preferably several hours, a method that proved useful in earlier experiments using phenolic antioxidants in bilayers.3a An experimental trace of a reaction profile under these conditions is shown in Figure 5A. Several differences were observed for the inhibiting effect of STAZN during initiation by DMVN compared to ABAP, namely: (i), the oxygen uptake trace was more regular, resulting in an “ideal” plot of eq 11 (R2 ) 1); (ii), the kinetic chain length was significantly higher and the rate constant dropped to 0.4 × 104 M-1 s-1, approximately 0.1-0.5 of that found during initiation with ABAP (Tables 3 and 4), and (iii), the stoichiometric factor, n, decreased to 1.14 Both Trolox and DBHA gave regular induction periods (Figure 5B and C, respectively) and linear plots according to eq 11 (not shown) for inhibited DLPC samples prepared under these conditions containing DMVN. DBHA was introduced by the “liposomal transfer” method15 from (14) On initiation of oxidation of PLPC bilayers with the lipid-soluble initiator, DMVN, the stoichiometric factor decreases to n ) 1. In these experiments the concentrations of STAZN and initiator that had to be used were such as to produce very long induction periods (see Figure 5), and we suspect that during this time some of the STAZN was lost by hydrolysis.
FIGURE 6. UV-vis spectra: (A) STAZN, (B) the peroxidation product of STAZN, and (C) the dialdehyde, 3. All in chlorobenzene and all 1.69 × 10-5 M.
dimyristoylphosphatidylcholine (DMPC) liposomes containing a known amount of DBHA. It is interesting to note that the antioxidant activity of DBHA is about the same in bilayers during initiation with DMVN as it is in micelles during initiation with ABAP (see Table 2, I). 2. Some Spectroscopic Properties of STAZN and Derived Products. Solutions of STAZN are highly colored and exhibit characteristic UV-vis spectra. The spectrum of STAZN in chlorobenzene, shown in Figure 6 (A), exhibits absorption in the visible at 629 nm, ) 2.2 × 103 and at 424 nm, ) 5.4 × 104, and in the UV at 337 nm, ) 6.9 × 104 and at 298 nm, ) 7.0 × 104. Figure 6 (B) shows the spectrum of a final reaction mixture where STAZN was the inhibitor in styrene/ chlorobenzene, giving absorption at 442 nm, ) 2.0 × 104 and at 309 nm, ) 5.1 × 104. Figure 6 (C) shows the spectrum of the dialdehyde, 3, in chlorobenzene with maxima at 440 nm, ) 2.0 × 104 and at 307 nm, ) 5.3 × 104. The similarity between the spectra of B and C indicates that peroxidation of STAZN converts it into the dialdehyde, 3. 3. Calculations. Methods. The cisoid and transoid structures of STAZN were optimized at the AM1 level of theory.16 Frequency calculations verified that the AM1 optimized structures are local minima. Molecular energies were then computed at the B3LYP/6-31G(d)17,18 level of theory and used, in conjunction with the AM1 fre(15) Barclay, L. R. C.; Antunes, F.; Egawa, Y.; McAllister, K. L.; Mukai, K.; Nishi, T.; Vinqvist, M. R. Biochim. Biophys. Acta 1997, 1328, 1-12. (16) Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. J. Am. Chem. Soc. 1985, 107, 3902-3909.
J. Org. Chem, Vol. 69, No. 9, 2004 2933
Mojumdar et al.
quency data, to obtain molecular enthalpies. The AM1 and B3LYP calculations were performed using the Gaussian-98 suite of programs.19 Aqueous solvation energies were computed using the SM5.42R20 method as implemented in the AMSOL21 package. To estimate the differences in aqueous solvation energies of cisoid and transoid STAZN molecules at lipid/water interfaces, one of the nitrone groups of the transoid isomer was replaced with an ethylene group. Results. The calculated gas-phase enthalpies indicate that the cisoid structure is 3.6 kcal/mol higher in energy than the transoid structure. However, the two polar nitrone groups of the cisoid can benefit from aqueous solvation at a lipid/water surface, whereas the transoid structure can have only one polar group exposed to water. Our calculations suggest that the energy of the cisoid conformer is lowered by solvation by 7.6 kcal/mol. Thus, the cisoid conformer of STAZN is expected to be ca. 4 kcal/mol more stable than the transoid structure in aqueous lipid dispersions.
SCHEME 1
Discussion The inhibition rate constant for STAZN depended on the reaction medium and type of initiator. In homogeneous solution in chlorobenzene and initiation with AIBN, STAZN trapped two peroxyl radicals and gave a classical inhibition profile of suppressed oxygen uptake with a rate constant slightly lower than that for the phenolic antioxidant, DBHA (Table 1). The antioxidant activity of STAZN increased 6-fold on addition of 5 vol % methanol, whereas the activity of DBHA was unchanged and the activity of PMHC decreased. The weaker antioxidant, STAZN-aldehyde, 2, had a stoichiometric factor of 1, consistent with the activity residing solely in the nitrone function. The antioxidant properties of STAZN were quite different in aqueous micelles and depended on whether the initiator was water-soluble (ABAP) or lipid-soluble (DMVN). On initiation with ABAP, STAZN exhibited activity greater than that of DBHA, its activity being comparable to that of the water-soluble phenolic antioxidant, Trolox. However, for STAZN the plots according to eq 11 were not linear and oxygen uptake was almost completely suppressed initially (Figure 3). In contrast, on initiation with DMVN, STAZN gave ideal plots ac(17) Becke, A. D. J. Chem. Phys. 1993, 98, 5648-5652. (18) Lee. C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785-789. (19) Frish, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A, Jr; Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, k. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Baboul, A. G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98, Revision A. 9; Gaussian, Inc.: Pittsburgh, PA, 1998. (20) Li, J.; Zhu, T.; Hawkins, G. D.; Winget, P.; Liotard, D. A.; Cramer, C. J.; Truhlar, D. G. Theor. Chem. Acc. 1999, 103, 9-63. (21) Hawkins, G. D.; Giesen, D. J.; Lynch, G. C.; Chambers, C. C.; Rossi, I.; Storer, J. W.; Li, J.; Zhu, T.; Thompson, J. D.; Winget, P.; Rinaldi, D.; Liotard, D. A.; Cramer, C. J.; Truhlar, D. G. AMSOL, Version 6.9; Regents of the University of Minnesota, 2003.
2934 J. Org. Chem., Vol. 69, No. 9, 2004
cording to eq 11, though the inhibition rate constant and stoichiometric factor were reduced (Table 2, II). Hydrogen bonding is well-known to reduce inhibition rate constants of phenolic antioxidants in aqueous micelles.3b In the present case, the kinh of DBHA is reduced to 23% and the kinh of STAZN is reduced to 57% of their values in chlorobenzene. The addition of 5% methanol to the chlorobenzene increased STAZN’s inhibition rate constant 6-fold (Table 1). Therefore, the decreased value of kinh for STAZN in micelles cannot be due to hydrogen bonding. The analogous decrease in inhibition rate constants for hydrophobic antioxidants, such as R-tocopherol, in SDS micelles has been attributed to ratelimiting diffusion of the antioxidant between the micelles.3c A similar explanation would apply to STAZN, which is hydrophobic and will only diffuse slowly between micelles. The inhibition rate constant for STAZN decreased further in lipid bilayers, as observed earlier for large hydrophobic inhibitors,3a and there was a further decrease when both STAZN and the initiator (DMVN) were present in the bilayer. Our kinetic data are consistent with the single electron transfer (SET) antioxidant mechanism previously proposed for STAZN,5 at least in solvents that support ionization, see Scheme 1. The initial SET intermediate, 6, is suggested to undergo a rapid nucleophilic addition of the peroxyl anion to yield the spin adduct, 7. In a nonpolar solvent 7 might form directly by initial peroxyl radical trapping. Such spin adducts are known to be unstable22 decomposing by β-scission into aldehydes,23 in our case the dialdehyde, 3, as was confirmed by the UVvis spectrum of a chlorobenzene reaction mixture after the end of the inhibition period, which closely resembles that of 3 (Figure 6). The initial homolysis products of 7, the alkoxyl radical 8 and nitroso compound 9, may combine to form a new adduct, 10.23 In addition, 8 could (22) Howard, J. A.; Tait, J. C. Can. J. Chem. 1978, 56, 176-178. (23) Janzen, E. G.; Krygsman, P. H.; Lindsay, D. A.; Haire, D. L. J. Am. Chem. Soc. 1990, 112, 8279-8284.
Kinetic Studies on Stilbazulenyl-bis-nitrone
FIGURE 7. Schematic representation of the position of STAZN in a phospholipid bilayer.
deplete STAZN by addition,24 which may account for the reduced stoichiometric factors for STAZN in aqueous micelles and bilayers (n ∼1.4-1.0). During initiation of linoleate peroxidation by ABAP in micelles and in bilayers, STAZN typically gave strong inhibition during the first part of the inhibition period, followed by a weaker inhibition and a gradual increase in the rate of oxygen uptake. This resulted in plots made according to eq 11 being nonlinear, an effect that was very pronounced in the bilayer experiments. The usual structure given for STAZN has the nitrone groups transoid with respect to one another as shown in Figure 1. This structure is probably that adopted by STAZN in homogeneous solution. However, it will be unfavorable in aqueous dispersions of lipids and particularly unfavorable in phospholipid bilayers because only one of the polar nitrone moieties could be solvated by the water, the other being deeply buried in the nonpolar lipid. The antioxidant behavior of STAZN in phospholipid bilayers suggests that both polar nitrone groups are oriented toward the aqueous phase in a cisoid structure, as shown schematically in Figure 7. In this situation, the two nitrones of each STAZN in the outer bilayer leaflets could react with the water-soluble, positively charged peroxyls derived from ABAP. However, the two nitrones of each STAZN in the inner bilayer leaflets would be isolated from these radicals because the charge on the peroxyls would prevent them from diffusing through the bilayer to reach the STAZN. The alternative way in which the inner leaflet STAZN could encounter the peroxyl radicals (24) Irradiation of a dilute solution of STAZN-aldehyde, 2, and ditert-butylperoxide in methylene chloride gave an ESR spectrum with aN ) 14.0 and aβH ) 2.0 G, typical of an oxygen-centered adduct of a nitrone.22 A Reviewer pointed out that cage-escaped alkoxyl radicals could react with the substrate, and this chain transfer step could contribute to the nonlinearity of the plots according to eq 11 for experiments in aqueous systems (see Mahoney, L. R. Angew. Chem., Int. Ed. Engl. 1969, 8, 547-555.)
would be by a “flip-flop” of STAZN from the inner to the outer leaflet. This would be a rather slow process because of the size and polarity of STAZN. Hence, although about half the STAZN (outer leaflet) can be a very effective antioxidant, once this half has been consumed, inhibition of oxidation by the remaining STAZN (inner leaflet) becomes much less efficient and will be limited by the diffusion of STAZN to the outer leaflet. Micelles are in dynamic equilibrium with monomers and, at concentrations appreciably above the critical micelle concentration (cmc), with larger aggregates such as rods and disks25 and vesicles.26 Such aggregates will be present in our 0.50 M SDS solutions since the cmc for SDS is 8.1 × 10-3 M .27 Obviously these are complex systems, but as far as inhibition by STAZN is concerned there will be some similarity to the situation in phospholipids bilayers. Computational modeling supports the proposed conformation and position of STAZN in a bilayer. The cisoid conformer (nonplanar)28 is calculated to be 3.6 kcal/mol higher in energy than the transoid conformer (planar) in the gas phase. However, solvation of both nitrone moieties of the cisoid conformer by water at lipid-water interfaces is calculated to stabilize the cisoid by about 4 kcal/mol relative to the transoid. The ability of STAZN to trap peroxyl radicals efficiently in the aqueous phases of micelles and phospholipid bilayers can be accounted for by the SET mechanism (Scheme 1) and its cisoid conformation at the lipid-water interface of these systems.29
Experimental Section Materials. Samples of STAZN (1), STAZN-aldehyde (2), and the dialdehyde (3) were prepared as described earlier.5 The commercial inhibitors 2,5,7,8-tetramethyl-6-hydroxychroman2-carboxylic acid (Trolox) (4b), 2,2,5,7,8-pentamethyl-6-hydroxychroman (4a), 2,6-di-tert-butyl-4-methoxyphenol (5), and diethylenetriamine-pentaacetic acid (Detapak) were obtained from Aldrich. The initiators ABAP and DMVN and AIBN, sodium dodecyl sulfate (electrophoresis purity), Chelex, methyl linoleate (purity > 99%), and phospholipids (shipped on dry ice) were obtained commercially. Styrene and cumene were obtained commercially and passed through a column of silica gel under argon immediately before use. Solvents used were HPLC purity. (25) Robinson, B. H.; Bucak, S.; Fontana, A. Langmuir 2000, 16, 8231-8237. (26) Hoffman, H.; Grabner, D.; Hornfeck, U.; Platz, G. J. Phys. Chem. B 1999, 103, 611-614. (27) Fennell, D.; Ninham, B. W. J. Phys. Chem. 1983, 87, 50255032. (28) One azulene moiety is calculated to be 14° out-of-plane and the other -34.5° out-of-plane. (29) We thank a reviewer for making the very interesting suggestion that a comparison of the UV-vis spectrum of STAZN in micelles with that in homogeneous solution could detect the proposed cisoid conformer. In the event, the spectrum of 1.83 × 10-5 M STAZN in 0.50 M SDS gave bands at 387 nm and at 306 nm with absorbance values of 0.188 and 0.428, respectively. The spectrum of this concentration in methanol gave bands at 416 and 336 nm with absorbance values of 0.563 and 0.735, respectively. That is, the absorption intensity of STAZN in micelles is reduced to between 30% and 60% of that in solution and the maxima are shifted about 30 nm to shorter wavelengths. We believe this is, indeed, consistent with the formation of cisoid STAZN in the micelles and with transoid STAZN being present in the homogeneous methanol solution, since for example, cis and trans stilbene exhibit absorption maxima at 280 and 295 nm, respectively, with the cis intensity about 50% that of the trans (see Lewis, G. N.; Magel, T. T.; Lipkin, D. J. Am. Chem. Soc. 1940, 62, 2943-2980.)
J. Org. Chem, Vol. 69, No. 9, 2004 2935
Mojumdar et al. Preparations and Analyses. Phosphate buffer, pH 7.4, commercially prepackaged, was prepared in deionized, distilled water containing 1 × 10-4 M Detapak and passed through a column of Chelex 100 to remove traces of heavy metal ions. Methyl linoleate was tested for hydroperoxides by TLC on silica gel with N,N-dimethyl-p-phenylenediamine spray. When necessary, hydroperoxides were removed by treating hexane solutions of the ester with silica gel as described earlier.30 Solutions of methyl linoleate were prepared in 0.50 SDS-buffer by vortex stirring until clear (5-10 min). Solutions of STAZN or of the initiator, DMVN, in 0.50 M SDS were prepared by mixing known concentrations in methanol and evaporating the mixture to constant weight on the rotoevaporator followed by vortex stirring with phosphate buffer. The concentration of STAZN was checked by taking UV-vis spectra of these samples diluted in methanol. A similar coevaporation procedure was used to incorporate STAZN or DMVN into DLPC bilayers. Multilamellar vesicles (MLV) of DLPC were prepared as described earlier,3a by vortex mixing a film of DLPC with phosphate buffer to obtain a dispersion. The latter was then subjected to 10 freeze-thaw cycles in a flask under argon alternately immersed in liquid nitrogen and tap water. The unilamellar DLPC was prepared by the extrusion method.13 The MLV dispersions were passed through filters in the following sequence: three passes each through 4-µm, 2-µm, and then 1-µm filters followed by one pass through two 1-µm filters. ESR spectra were obtained on a Varian E3 spectrometer at a power level of 3 mW and modulation amplitude of 3.2 G. Autoxidation/Inhibition Procedures. Autoxidations were generally carried out at 37 °C under 760 Torr of air in a dual
2936 J. Org. Chem., Vol. 69, No. 9, 2004
channel oxygen uptake apparatus equipped with a sensitive pressure transducer described earlier.31 Some experiments in SDS micelles were conducted on a two-channel oxygen electrode. The procedures for adding initiators or inhibitors for experiments employing 0.50 M SDS or lipid bilayers have been described earlier.3a,b
Acknowledgment. This research was supported by a grant to L.R.C.B. from the Natural Sciences and Engineering Council of Canada and a National Institute of Health grant NS-38221 to D.A.B. We thank Ms. Melinda Vinqvist for assistance with preparation of the manuscript and acknowledge the assistance of the computing Services Research Group, University of Alberta and Jill Innes, Computing Services, Mount Allison University, for assistance with the figures for the structures. Supporting Information Available: Tables 1-4 with more details of the experimental data; structures of transoid and cisoid conformers of STAZN by computational modeling. This material is available free of charge via the Internet at http://pubs.acs.org. JO030390I (30) Barclay, L. R. C.; Basque, M.-C.; Vinqvist, M. R. Can. J. Chem. 2003, 81, 457-467. (31) Wayner, D. D. M.; Burton, G. W. Handbook of Free Radicals and Antioxidants in Biomedicine; CRC Press: Boca Raton, FL, 1989; Vol. III, pp 223-232.
