Original Contribution

SINGLET OXYGEN QUENCHING AND THE REDOX PROPERTIES OFHYDROXYCINNAMIC ACIDS

SARAH FOLEY,* SUPPIAH NAVARATNAM ,† DAVID J. MCGARVEY,* EDWARD J. LAND,† T. GEORGE TRUSCOTT*and CATHERINE A. RICE-EVANS

‡

*Department of Chemistry, Keele University, Keele, Staffordshire, UK;†CRC Section of Drug Development and Imaging,Paterson Institute for Cancer Research, Christie Hospital NHS Trust, Manchester, UK; and‡Free Radical Research Group,

Division of Biochemistry and Molecular Biology, UMDS–Guy’s Hospital, St. Thomas’s Street,London, UK

(Received19 June1998;Revised9 October1998;Accepted26 October1998)

Abstract—The singlet oxygen quenching rate constants (kq) for a range of hydroxycinnamic acids in acetonitrile andD2O solutions were measured using time resolved near infrared phosphorescence in order to establish their antioxidantactivity. The magnitude of kq observed depends on both the nature of the substituent groups and solvent polarity. Thevariations in kq depend on the energy of the hydroxycinnamic acid/molecular oxygen charge transfer states,(O2

d2...HCAd1). In D2O the values of kq range from 43 107 M21 s21 to 4 3 106 M21 s21 for caffeic acid ando-coumaric acid respectively. In acetonitrile, the charge transfer energy levels are raised and this is reflected in lowersinglet oxygen quenching rate constants with a kq value of 53 106 M21 s21 for caffeic acid. The phenoxyl radicalspectra derived from the hydroxycinnamic acids were determined using pulse radiolysis of aqueous solutions and thereduction potentials were found to range from 534 to 596 mV. A linear correlation is observed between reductionpotential, and hence free energy for electron transfer, and log kq. These correlations suggest a charge transfer mechanismfor the quenching of singlet oxygen by the hydroxycinnamic acids. © 1999 Elsevier Science Inc.

Keywords—Hydroxycinnamic acids, Singlet oxygen, Free radicals, Redox potentials

INTRODUCTION

A growing amount of evidence suggests that the dietaryintake of phytochemicals plays an important role inmaintaining health and protecting against degenerativeprocesses including cardiovascular disease and certaincancers. The protective effects of the antioxidant constit-uents of fruit and vegetables have been attributed to thecarotenoids, vitamins C and E and the flavonoids [1,2].

In recent years antioxidant studies have been extendedand now include simple phenolic acids or phenylpro-panoids that are present in plants at greater concentra-tions than the polyphenolic flavanoids. The hydroxycin-namic acids (HCA) are amongst the most widelydistributed phenolic acids in plant tissue [3]. They in-clude caffeic acid, chlorogenic acid (the quinic acid ester

of caffeic acid),o-, m-, p-coumaric acids, ferulic acid andsinapic acid and are produced from the shikimate path-way from either phenylalanine orL-tyrosine (Fig. 1).They most commonly occur as simple esters with quinicacid or glucose and are present in a wide variety of fruitand vegetables including tomatoes, spinach, broccoli,asparagus, white grapes, pears and peaches.

The antioxidant properties of the HCA have beendemonstrated with regard to their ability to scavengeradicals generated in the aqueous phase [4] to increasethe resistance of low density lipoprotein (LDL) to lipidperoxidation [5,6]. Their anti-inflammatory and anti-mutagenic abilities have also been demonstrated [7,8].The uptake and excretion of hydroxycinnamates from atomato rich diet have recently been demonstrated inhuman nutritional studies [9,10].

In this paper the antioxidant potential of the HCAwith respect to their ability to quench singlet molecularoxygen, a reactive oxygen species that appears to play arole in tumour promotion and carcinogenesis, has been

Address correspondence to: Professor T. G. Truscott, Department ofChemistry, Keele University, Keele, Staffordshire, ST5 5BG, UK. Tel:144 01782 583038; Fax:144 01782 712378;E-Mail: cha31@chem.keele.ac.uk.

Free Radical Biology & Medicine, Vol. 26, Nos. 9/10, pp. 1202–1208, 1999Copyright © 1999 Elsevier Science Inc.Printed in the USA. All rights reserved

0891-5849/99/$–see front matter

PII S0891-5849(98)00313-X

1202

studied [11]. In addition, the reduction potentials of thephenoxyl radicals derived from the HCA have beenestimated via pulse radiolysis and these show thequenching of singlet oxygen by the HCA occurs via acharge transfer mechanism.

MATERIALS AND METHODS

Materials

Methylene blue, potassium thiocyanate and sodiumazide (British Drug Houses Ltd., Poole, Dorset, UK),caffeic acid, chlorogenic acid,m-, o-, p-coumaric acids,ferulic acid and sinapic acid (Sigma Chemical Co.,Poole, Dorset, UK, Aldrich Chemical Co., Gillingham,Dorset, UK), Trolox (Aldrich), acetonitrile (spectroscop-ic grade, British Drug Houses Ltd.) and D2O (GossScientific Instruments Ltd., Chelmsford, Essex, UK)were all used as received.

Instrumentation

Kinetic measurements were made using a JK System2000 Q switched Ruby laser operating on the fundamen-tal at 694 nm. Time-resolved singlet oxygen phospho-rescence (Eq. 1):

O2~1Dg! 3 O2~

3(g2! 1 hn~1270nm! (1)

was detected using a Judson germanium photodiode (G-050, active diameter5 0.5 cm) coupled to a Judsonpreamplifier.

Pulse radiolysis measurements were made using a9–12 MeV Vickers linear accelerator as described pre-viously [12] with pulses of 10–100 ns duration and dosesof between 1 and 10 Gy. Irradiation was in quartz flowthrough cells with a monitoring optical path length of 2.5cm. Absorption measurements were recorded using aPerkin-Elmer Lambda-2 UV/Vis spectrophotometer.

Methods

Singlet oxygen quenching rate constant measurements.The rate constants (kq) for the quenching of singletoxygen by the HCA were determined in acetonitrile andD2O by monitoring the time-resolved phosphorescencefollowing laser excitation of methylene blue (MB) at 694nm. This choice of wavelength precludes any absorptionby the HCA. At each concentration of HCA the recordedluminescence trace was obtained by signal averaging tensingle shots. The averaged traces were fitted with a singleexponential. Linear plots of HCA concentration versusthe observed pseudo first order rate constant, k2, forsinglet oxygen deactivation were obtained, where k2 5k1 1 kq[HCA], k1 being the intrinsic first order decayconstant of singlet oxygen in the absence of HCA. Thephotosensitized production of singlet oxygen involvesenergy transfer from the excited triplet state of MB tomolecular oxygen (Eq. 2):

3MB* 1 O2~3(g

2!O¡kET

MB 1 O2~1Dg! (2)

Generation of phenoxyl radicals

The method of production of phenoxyl radicals andtheir kinetic behaviour has been described previously[13]. The HCA (13 1024 M) were dissolved in phos-phate buffer (pH5 7) solutions containing NaN3 (0.1 M)purged of oxygen and saturated with N2O in order toconvert hydrated electrons into hydroxyl radicals. Thehydroxyl radicals react with the azide to produce the lessoxidizing azide radical. The resulting ion radical reacts

Fig. 1. The synthetic pathway of the hydroxycinnamic acids in plants.

1203Hydroxycinnamic acid antioxidant properties

with the HCA to produce its phenoxyl radical cation thatrapidly deprotonates (Eqs. 3–7).

H2OMM3 e2~aq! 1 •OH (3)

N2O 1 H2O 1 e2~aq!3 N2 1 •OH 1 OH2 (4)

•OH 1 N32 3 OH2 1 N3

• (5)

HCA 1 N3• 3 HCA•1 1 N3

2 (6)

HCA•13 HCA• 1 H1 (7)

Determination of radical absorption coefficients

Absorption coefficients were obtained using thiocya-nate dosimetry. The absorbed doses were determinedfrom transient (SCN)2

•2 formation in air-saturated 1022

M KSCN solution. In such solutions, the primary•OHradicals are converted into (SCN)2

•2 via reactions (Eq.8) and (Eq. 9):

•OH 1 SCN23 OH2 1 SCN• (8)

SCN• 1 SCN23 ~SCN!2•2 (9)

The absorption coefficient of (SCN)2•2 was taken to

be 7100 M21 cm21 at 500 nm assuming a G value of2.9 [14].

In determining the HCA radical absorption coeffi-cients, it was assumed that, under the low dose (up to 6Gy) conditions employed, complete scavenging of N3

•

by the HCA occurred (G(N3•) 5 6.13) in N2O5saturated

solutions [15].

Determination of reduction potentials

The reduction potentials of the phenoxyl radicals de-rived from the HCA were determined at pH 7 withreference to the Trolox redox couple E5 1480 mV, vs.the hydrogen electrode [16], the only exception being the

reduction potential for chlorogenic acid that was deter-mined against ferulic acid.

The electron transfer reactions between electron ac-ceptor radicals and the electron donor e.g.,

HCA• 1 Troloxº HCA 1 Trolox• (10)

Fig. 2. Observed pseudo first order rate constants, k2, for decay ofsinglet oxygen as a function of the concentrations of hydroxycinnamicacids, from which the second order quenching rate constants, kq, wereobtained in acetonitrile (A) and in D2O (B).

Table 1. Singlet-Oxygen Quenching Rate Constants (kq) for theHCA in Acetonitrile and D2O

HCAAcetonitrile

kq/105 M21 s21D2O

kq/106 M21 s21

Caffeic acid 51 40Chlorogenic acid — 34m-Coumaric acid 6 7o-Coumaric acid — 4p-Coumaric acid 7 6Ferulic acid 36 20Sinapic acid 47 30

Table 2. Phenoxyl Radical Properties for the HCA in PhosphateBuffer (pH 5 7)

HCA lmax/nm «/103 M21 cm21k(N3

• 1 HCA)/109 M21 s21

Caffeic acid 380 9.6 3.9Chlorogenic acid 390 8.6 3.7o-Coumaric acid 360 3.4 0.2p-Coumaric acid 340 27 1.0Ferulic acid 360 13.2 4.3Sinapic acid 360 20.2 3.3

1204 S. FOLEY et al.

were followed by kinetic spectroscopy. The absorbanceof each of the radicals was determined separately, andthe absorbance at equilibrium allowed the determinationof the radical concentrations. It should be noted that at

pH 7 the forms of the hydroxycinnamic acids, Trolox andthe relatively stable semiquinones, are referred to asHCA, Trolox, HCA•, and Trolox•, respectively, withoutany implication as to the net charge on such species.

Fig. 3. Changes in absorption resulting from the formation of the corresponding phenoxyl radicals at pH 7 by N3•-induced oxidation

of (A) caffeic acid, (B) chlorogenic acid, (C)o-coumaric acid, (D)p-coumaric acid, (E) ferulic acid, and (F) sinapic acid.

1205Hydroxycinnamic acid antioxidant properties

RESULTS AND DISCUSSION

Singlet oxygen quenching by the HCA

The measured values for the quenching of singletoxygen by the series of HCA in acetonitrile and D2O arecollected in Table 1. In all experiments the emissionobserved at 1270 nm decayed exponentially with k1, theintrinsic first order decay rate constant, being solventdependent. The average lifetimes of 67ms and 54mscompare with the published values of 54–68ms and 68ms [17,18] for these two solvents, respectively. Plots ofk2, the observed pseudo first order rate constant for thedeactivation of singlet oxygen in the presence of theHCA were plotted against the concentration of thequencher producing straight line graphs as illustrated inFig. 2A and 2B, the gradient of that is equal to thequenching rate constant kq and the intercept is k1. Similarplots were obtained for all the HCA.

As Table 1 illustrates, the quenching rate constantsare sensitive to both the nature of the substituent and thepolarity of the solvent. Because the values of kq descendthe group of HCA in the same order irrespective of thesolvent, the observed differences are due to differencesin substituents. Quenching rate constants are higher in

D2O than acetonitrile whilst, the dihydroxyl substitutedcaffeic acid is the most efficient quencher and the mono-substituted coumaric acids are least efficient. All of theseobservations are rationalized on the basis of the electrondonating properties of the substituents (hydroxyl groupsbeing more electron donating than methoxy groups).Moreover, the results also show [19] that the position ofthe OH group has little effect on the quenching effi-ciency, as kq for o-, m-, andp-coumaric acids are verysimilar. In contrast, the number of methoxyl groups playsan important role as can be seen from the comparison ofkq for p-coumaric, ferulic, and sinapic acids.

Absorption spectra of the radicals

Table 2 gives a summary of the phenoxyl radicalproperties for each of the HCA studied. The maxima inthe phenoxyl radical spectra range from 340–390 nm andall exhibit a well defined peak as shown in Figs. 3A–F.The rate constants of quenching of azide radical andwere measured by following the kinetics of formation ofthe HCA radical at the corresponding absorption max-ima. With the exception of the o- and p-coumaric acidsthey are all approximately 3–43 109 M21 s21 andapproach diffusion control.

Reduction potentials

The reduction potentials of the phenoxyl radicals de-rived from the HCA studied were determined at pH 7with reference to Trolox (Eq. 10). The determination ofthe equilibrium constantK from the absorbance at equi-librium is demonstrated by the system for Trolox vs.caffeic acid in Fig. 4 and Table 3. The absorbance ofeach radical is determined separately and the absorbanceat equilibrium allows the determination of the ratio ofradical concentrations at equilibrium. The equilibriumconstantK is then calculated from this ratio and theinitial concentrations of the parent compounds in solu-tion (Eq. 11). As Table 3 illustrates the values ofK areconstant within experimental error and over a range ofconcentrations.

K 5[Trolox•] z[HCA]

[HCA•] z[Trolox](11)

Figure 4 shows an example of a kinetic trace that clearlyshows the establishment of an equilibrium and that al-lows determination of the absorbance. The difference inreduction potentials between the two systems is given bythe Nernst equation.

DE 5 0.059 log K (12)

Fig. 4. Time profiles of absorption changes at 380 nm showing theestablishment of equlibrium in a typical case. Because some decay ofthe absorption observed from the caffeic acid alone, and in the presenceof Trolox, occurs during equilibrium, the slow decays were extrapo-lated back to the time of the pulse (as shown by fine lines).

Table 3. Absorbances at 380 nm (Normalized to a Dose of 1.3 Gy)at Equilibrium for Varying Concentrations of Caffeic Acid and

Trolox, From Which the Constant K Was Determined

[Caffeic Acid]/mM [Trolox]/mMAbsorbance at 380 nm

at equilibrium/1023 K

8 1 8.69 6.358 6 3.96 7.64

10 5 12.0 10.56

1206 S. FOLEY et al.

The results of the various systems studied are given inTable 4 with E values ranging from 534 mV to 596 mVfor caffeic acid and o-coumaric acid, respectively. Thevalue of 534 mV for caffeic acid is in good agreementwith the value of 540 mV obtained by Jovanovic andcoworkers [20].

There is a linear correlation between log kq and thereduction potential at pH 7, the rate constants decreasingas the reduction potentials increase, as shown in Figure5. Similar linear relationships have previously been ob-served for other series of phenolic compounds andamines [21–24]. Such correlations suggest that thequenching of singlet oxygen occurs via a charge transferreaction mechanism with the following scheme (Eq. 13)being proposed.

1O2 1 HCA^1~O2d2...HCAd1!

^3~O2d2...HCAd1!3 3O2 1 HCA (13)

The free energy change,DG, that is associated with theelectron transfer process is represented by Eq. 14:

DG 5 F~E~D•1/D! 2 E~ A/A •2!! 2 E* (14)

where F is the Faraday constant (96.5 C mol21),E(D•1/D) is the reduction potential of the electron do-nor, the HCA, E(A/A•2) is the reduction potential ofoxygen (E O2/O2

•2 5 2330 mV) [25] and E* is theenergy of the excited state acceptor (94 kJ mol21).

For electron transfer reactions between members ofthe same homologous series of donors (the HCA) and thesame acceptor (O2,

1Dg) a linear relationship exists be-tween the logarithm of the quenching rate constant kq

andDG, the free energy change for the reaction. Becauseall terms relating to the acceptor remain constant then logkq is linearly dependent on E(D•1/D) as illustrated in Fig.5. Such behaviour is consistent with current theories onelectron transfer reactions.

CONCLUSIONS

The antioxidant ability of the hydroxycinnamic acidshas been demonstrated with respect to their capability toquench singlet molecular oxygen, an example of a reac-tive oxygen species. The results have shown that theelectron donating ability of the substituents and the po-larity of the solvent influences the quenching rate con-stant kq. It has also been shown that there is an inversecorrelation between log kq and the reduction potentials ofthe hydroxycinnamic acids. The above results are evi-dence for the participation of charge transfer interactionduring the quenching of singlet oxygen by the hydroxy-cinnamic acids.

Acknowledgements— The pulse radiolysis experiments were per-formed at the Paterson Institute for Cancer Research Free RadicalResearch Facility, Christie Hospital NHS Trust, Manchester. The fa-cility is supported by the European Commission T.M.R. Programme—Access to Large-Scale Facilities. E.J.L. thanks the Cancer ResearchCampaign for financial support. We also thank Dr. R. V. Benasson foruseful discussions.

REFERENCES

[1] Block, G. The data support a role for antioxidants in reducingcancer risk.Nutr. Rev.50:207–213; 1992.

[2] Block, G.; Langseth, L. Antioxidant vitamins and disease preven-tion. Food Technol.July:80–84; 1994.

[3] Herrmann, K. Occurrence and content of hydroxycinnamic acidand hydroxybenzoic acid compounds in foods.Crit. Rev. Food.Sci. Nutr.28:315–347; 1989.

[4] Rice-Evans, C.; Miller, N.; Pagana, G. Structure-antioxidant ac-tivity relationships of flavonoids and phenolic acids.Free Radic.Biol. Med.20:933–956; 1996.

[5] Castellucio, C.; Pagana, G.; Melikian, N.; Bolwell, G. P.;Pridham, J.; Sampson, J.; Rice-Evans, C. Antioxidant potential ofintermediates in phenylpropanoid metabolism in higher plants.FEBS Lett.368:188–192; 1995.

[6] Nardini, M.; D’Aquino, M.; Tomassi, G.; Gentili, V.; Di Felice,M.; Scaccini, C. Inhibition of human low density lipoproteinoxidation by caffeic acid and other hydroxycinnamic derivatives.Free Radic. Biol. Med.19:541–552; 1995.

[7] Wood, A. W.; Huang, M.-T.; Chang, R. L.; Newmark, H. L.;

Fig. 5. Plot of the logarithm of the O2,1Dg quenching rate constants for

the HCA in D2O vs. the reduction potential in phosphate buffer atpH 7.

Table 4. Reduction Potentials at pH7 and Free Energy of ElectronTransfer for the Various HCA

HCA K E7/mV DG/kJ mol21

Caffeic acid 8.18 534 210.6Chlorogenic acid 15.5 550 29.1o-Coumaric acid 92.3 596 24.6p-Coumaric acid 72.2 590 25.2Ferulic acid 90.7 595 24.7Sinapic acid 19.1 556 28.5

1207Hydroxycinnamic acid antioxidant properties

Lehr, R. E.; Yagi, H.; Sayer, J. M.; Jerina, D. M.; Cooney, A. H.Inhibition of the muagenicity of bay-region diol epoxides ofpolycyclic aromatic hydrocarbons by naturally occurring plantphenolics.Proc. Natl. Acad. Sci. USA79:5513–5517; 1982.

[8] Sud’ina, G. F.; Mirzoeva, O. K.; Pushkareva, M. A.; Korshunova,G. A.; Sumbutya, N. V.; Varfolomeev, S. D. Caffeic acid phen-ethyl ester as a lipoxygenase inhibitor with antioxidant properties.FEBS Lett.329:21–24; 1993.

[9] Bourne, L.; Rice-Evans, C. Urinary detection of hydroxycinnama-tes and flavonoids in humans after high dietary intake of fruit.Free Radic. Res. 28:429–438; 1998.

[10] Bourne, L.; Rice-Evans, C. Methods for detecting and measuringthe bioavailability of phenolics and flavonoids in humans: phar-mokinetics of urinary excretion of dietary ferulic acid.MethodsEnzymol. 299:91–106; 1998.

[11] Cerutti, P. A. Prooxidant states and tumour promotion.Science227:375–381; 1985.

[12] Butler, J.; Hodgson, B. W.; Hoey, B. M.; Land, E. J.; Lea, J. S.;Lindley, E. J.; Rushton, F. A. P.; Swallow, A. J. Experimentalstudies of some moderately fast processes initiated by radiation.Radiat. Phys. Chem. Soc., 34:633–646; 1989.

[13] Steenken, S.; Neta, P. Electron transfer rates and equilibria be-tween substituted phenoxide ions and phenoxyl radicals.J. Phys.Chem.83:1134–1137; 1979.

[14] Adams, G. E.; Boag, J. W.; Michael, B. D.; Currant, J. The pulseradiolysis of the thiocyanate ion. In: Ebert, M., Keene, J. P.;Swallow, A. J.; Baxendale, J. H., eds.Pulse radiolysis.London:Academic Press; 1965:117–129.

[15] Schuler, R. H.; Patterson, K. L.; Janata, E. Yield for the scaveng-ing of OH radicals in the radiolysis of N2O-saturated aqueoussolutions.J. Phys. Chem.84:2088–2089; 1980.

[16] Steenken, S.; Neta, P. One-electron redox potentials of phenols.Hydroxy and aminophenols and related compounds of biologicalinterest.J. Phys. Chem.86:3661–3667; 1982.

[17] Rodgers, M. A. J. Solvent-induced deactivation of singlet oxygen-additivity relationships in non-aromatic solvents.J. Am. Chem.Soc.105:6201–6205; 1983.

[18] Schmidt, R. Influence of heavy atoms on the deactivation ofsinglet oxygen in solution.J. Am. Chem. Soc.111:6983–6987;1989.

[19] We thank the referee for this comment.[20] Jovanovic, S. V.; Steenken, S.; Tosic, M.; Marjanovic, B.; Simic,

M. G. Flavonoids as antioxidants.J. Am. Chem. Soc.116:4846–4851; 1994.

[21] Scurlock, R.; Rougee, M.; Bensasson, R. V. Redox properties ofphenol, their relationship to singlet oxygen quenching and to theirinhibitory effects on benzo (a) pyrene-induced neoplasia.FreeRadic. Res. Comm. 8:251–258; 1990 and erratum,Free Radic.Res. Comm. 16:205; 1992.

[22] Scurlock, R.; Rougee, M.; Bensasson, R. V. Redox properties ofdiphenols and their correlation to induction of enzyme synthesis.Z. Phys. Chem.196:85–92; 1996.

[23] Thomas, M. J.; Foote, C. S. Chemistry of singlet oxygen XXVI:photooxygenation of phenols.Photochem. Photobiol.27:683–693; 1978.

[24] Saito, I.; Matsura, T.; Inoue, K. Formation of superoxide ion viaone-electron transfer from electron donors to singlet oxygen.J. Am. Chem. Soc.105:3200–3206; 1983.

[25] Wardman, P. Reduction potentials of one-electron couples involv-ing free radicals in aqueous solutions.J. Phys. Chem. Ref. Data18:1637; 1989.

1208 S. FOLEY et al.