A spectroelectrochemical and chemical study on oxidation of hydroxycinnamic acids in aprotic medium

Page 1

A

awpptwfp

t©

K

1

tcptaseoac

0d

Electrochimica Acta 52 (2007) 2461–2470

A spectroelectrochemical and chemical study on oxidation ofhydroxycinnamic acids in aprotic medium

Rita Petrucci a,∗, Paola Astolfi b, Lucedio Greci b, Omidreza Firuzi c,Luciano Saso c, Giancarlo Marrosu a

a Dipartimento di Ingegneria Chimica M.M.P.M., Universita di Roma “La Sapienza”, via del Castro Laurenziano 7, I-00161 Rome, Italyb Dipartimento di Scienze dei Materiali e della Terra, Universita Politecnica delle Marche, via Brecce Bianche, I-60131 Ancona, Italy

c Dipartimento di Farmacologia delle Sostanze Naturali e Fisiologia Generale, Universita di Roma “La Sapienza”,p.le Aldo Moro 5, I-00185 Rome, Italy

Received 12 June 2006; received in revised form 22 August 2006; accepted 25 August 2006Available online 2 October 2006

bstract

Electrochemical and chemical oxidation of hydroxycinnamic acids (HCAs) was studied to investigate the mechanisms occurring in theirntioxidant activities in a protons poor medium. Electrolyses and chemical reactions were followed on-line by monitoring the UV-spectral changesith time; final solutions were analysed by HPLC–MS. Anodic oxidation of mono- and di-HCAs, studied by cyclic voltammetry and controlledotential electrolyses, occurs via a reversible one-step two-electrons process, yielding the corresponding stable phenoxonium cation. A cyclizationroduct was also proposed, as supported by ESR studies. Chemical oxidation with lead dioxide leads to different oxidation products according tohe starting substrate. Di-HCAs like chlorogenic and rosmarinic acids and the ethyl ester of caffeic acid gave the corresponding neutral o-quinones,hile mono-HCAs like cumaric, ferulic and sinapinic acids yielded the corresponding unstable neutral phenoxyl radical, as supported by the

ormation of dimerization products evidenced by HPLC–MS. In the case of caffeic acid, traces of the dimerization product suggest that the neutralhenoxyl radical may competitively undergo dimerization or decomposition of the neutral quinone.

Chemical oxidation of HCAs was also followed by ESR spectroscopy: the di-HCAs radical anions were generated and detected, whereas amonghe mono-HCAs only the phenoxyl radical of the sinapinic acid was recorded.

2006 Elsevier Ltd. All rights reserved.

uinon

afa

[amuda

eywords: Natural antioxidants; Polyphenols oxidation; Radical anions; Semiq

. Introduction

The consumption of fruits and vegetables has been showno decrease the risk of cardiovascular diseases [1] and can-er [2]. There are strong evidences that phenolic antioxidantsresent in plants are at least in part responsible for these pro-ective effects [3]. Hydroxycinnamic acids or phenylpropanoidsre widely distributed in the plant kingdom and are importantources of antioxidants due to their free radical scavenging prop-rties [4]. It has been shown that they protect LDL against

xidation induced by metmyoglobin [5], Cu2+ ions and 2,2′-zobis(2-amidinopropane) dihydrochloride (AAPH) [6]. Theyan also act as scavengers of hypochlorite [7], peroxynitrite [8]

∗ Corresponding author. Tel.: +39 06 49766855; fax: +39 06 49766749.E-mail address: [email protected] (R. Petrucci).

cakcn

m

013-4686/$ – see front matter © 2006 Elsevier Ltd. All rights reserved.oi:10.1016/j.electacta.2006.08.053

es; Spectroelectrochemistry

nd 2,2-diphenyl-1-picrylhydrazyl radical [9]. In particular, caf-eic acid and chlorogenic acid have been proved to be highlyctive in most of these studies.

Both electrochemical [10–13] and chemical oxidation10,14–16] of these compounds have been studied by manyuthors, especially in aqueous medium at different pH, but theechanisms involved in their oxidative process are not yet well

nderstood, above all because the radical species likely formeduring their oxidation (phenoxyl radicals and radical anions)re highly reactive and unstable and may be involved in compli-ated chemical reactions [15,17,18]. Their possible pro-oxidantctivity and their metabolic fate in human body are still scarcelynown, too, limiting the possible employment of these natural

ompounds in the preventive medicine as well as in the recog-ized pathology.

In this contest, we studied, in the present work, a series ofono- and di-hydroxycinnamic acids with the aim to investigate

Page 2

2 imica Acta 52 (2007) 2461–2470

tpoo

2

2

apsQec1owTn41

2

exietpscttt20i

aAruwptiaawoc1

Table 1Electrochemical data of HCAs, C and TBC in anhydrous acetonitrile 0.1 mol L−1

NaClO4, at a static GC electrode, scan rate 0.200 V s−1, vs. Ag/AgClO4 in theabsence and in the presence (*) of 2,6-lutidine

[S] (1 × 10−3

mol L−1)Eap (V) Eap (*) (V) Ecp (V) Ecp (*) (V)

C +0.85 +0.28 +0.18 −0.30TBC +0.77 +0.26 +0.13 −0.34CA +0.79 +0.26 +0.36 −0.30CAE +0.83 +0.26 +0.12 −0.31CGAa +0.80 – +0.25 –RA +0.86 +0.25 +0.13 −0.15/−0.25FA +0.91 +0.40 +0.32 −0.27SA +0.78 +0.25 +0.49 −0.15/−0.30CUAb +1.20 – – –

a Some data undetectable because of the presence of an abundant precipitateo

s

2

ts1

2

tMM(U

ePr25ri4sbbt2i(

2

462 R. Petrucci et al. / Electroch

he mechanisms occurring in the antioxidant activity in a protonsoor medium, under electron transfer (ET) conditions (anodicxidation) and H-atom transfer (HT) conditions (chemicalxidation with a H-atom acceptor).

. Materials and methods

.1. Materials

Reagents and solvents were purchased from Sigma–Aldrichnd used without further purification. Sodium perchlorate wasurchased from BDH and dried under vacuum after cristalli-ation. Water for HPLC was deionized by Millipore Milli-

Purification System. 3-(3,4-Dihydroxy-phenyl)-acrylic acidthyl ester (CAE) was prepared according to the following pro-edure: 2 mmols of caffeic acid (360 mg) were dissolved in0 mL of ethyl alcohol in the presence of a catalytic amountf p-toluenesulfonic acid. The solution was refluxed for 6 h,ashed with NaHCO3 0.5 mol L−1 and extracted with CHCl3.he organic layer was dried over Na2SO4 and evaporated to dry-ess. 1H NMR (200 MHz, CDCl3): δ = 1.39 (t, 3H, J = 7.6 Hz);.26 (q, 2H, J = 7.6 Hz); 6.26 (d, 1H, J = 17.3 Hz); 6.81–6.89 (m,H); 6.90–7.10 (m, 1H) ppm.

.2. Electrochemical experiments

Voltammetric measurements were performed with a three-lectrode multipolarograph AMEL 472 coupled with a digital/y recorder AMEL 863, with a static glassy-carbon (GC) work-ng electrode and a glassy-carbon Radiometer rotating disklectrode (BM-EDI 101) coupled with a Radiometer speed con-rol unit (CTV 101), Ag–AgClO4 (0.1 mol L−1)/MeCN – fineorosity fritted glass disk – MeCN/NaClO4 (0.1 mol L−1) –intered glass disk [19] as reference and a platinum wire asounter electrode. All experiments were carried out at roomemperature on nitrogen purged solution of anhydrous ace-onitrile containing 0.1 mol L−1 NaClO4 as supporting elec-rolyte, 1 × 10−3 mol L−1 substrate and increasing amounts of,6-lutidine as deprotonating agent. Scan rates varied from.020 to 0.200 V s−1. The accuracy of the potentials reportedn Tables 1 and 2 is ±5 mV.

Spectroelectrochemical experiments were carried out withdiode array spectrophotometer HP 8452A and a potentiostatMEL 552 coupled with an integrator AMEL 731 and a x/y

ecorder LINSEIS L250E for controlled potential electrolyses,sing a three-electrode modified UV cell, a platinum wire asorking electrode, an Ag–AgClO4 (0.1 mol L−1)/MeCN – fineorosity fritted glass disk – MeCN/NaClO4 (0.1 mol L−1) – sin-ered glass disk as reference and a platinum wire (placed on thenner wall of a glass tube containing MeCN/NaClO4 0.1 mol L−1

nd connected to the test solution via a sintered glass-disk)

s auxiliary electrode. The solution was stirred by purgingith a continuous nitrogen flux. All experiments were carriedut at room temperature on solution of anhydrous acetonitrileontaining NaClO4 0.1 mol L−1 as supporting electrolyte and× 10−4–5 × 10−4 mol L−1 substrate.

tMt

f lutidinium salt.b Some data undetectable because of adsorption phenomena at the electrode

urface.

.3. Chemical oxidation

Chemical oxidation was carried out by adding directly inhe UV cell small amounts of freshly prepared PbO2 [20] toolutions of anhydrous acetonitrile containing 1 × 10−4–5 ×0−4 mol L−1 substrate, purged with nitrogen.

.4. HPLC–MS instrumentation and conditions

HPLC–MS analyses were carried out on an HPLC separa-ion module 1525 � Waters, linked to a Quattro Micro Tandem

S-MS with an electro-spray interface Waters (Micromass,anchester, UK). Data were processed by MassLynx software

Data Handling System for Windows, Version 4.0, Micromass,K).The solution obtained from the oxidation with PbO2 of

ach HCA dissolved in acetonitrile was filtered throughTFE 0.45 �m filter disk and injected (10 �L) into a C18eversed phase column (Waters Spherisorb 5 �m ODS2,.1 mm × 150 mm). Solvent A consisted of deionized watermM formic acid and solvent B consisted of methanol. At a flow

ate of 200 �L min−1, the following binary gradient with linearnterpolation was used: 0 min, 20% B; 2 min, 20% B; 12 min,0% B; 15 min, 40% B; 25 min, 60% B; 40 min, 60% B. Beforetarting and at the end of each gradient, the column was equili-rated to initial condition for 30 min. The effluent was analysedy an electrospray source (source temperature, 100 ◦C; desolva-ion temperature, 250 ◦C; capillary voltage, 2.7 kV; cone voltage,5 V). Nitrogen was used as nebulizing (550 L h−1) and as dry-ng gas (50 L h−1). Electrospray ionization-mass spectrometryESI-MS) data were acquired in negative mode.

.5. ESR experiments

ESR spectra were recorded on a Bruker EMX EPR spec-rometer (Bruker, Karlsruhe, Germany) equipped with an XL

icrowave frequency counter, Model 3120 for the determina-ion of the g-factors. Computer simulation of EPR spectra were

Page 3

R. Petrucci et al. / Electrochimica Acta 52 (2007) 2461–2470 2463

Table 2Electrochemical data of C, TBC, CA, CAE, FA and ferrocene as a reference, in anhydrous acetonitrile 0.1 mol L−1, at a GC RDE, vs. Ag/AgClO4

ω (rad s−1) iL/4.23Cω1/2

Ferrocene C TBC CA CAE FAa

52.50 1.70 × 10−6 3.35 × 10−6 2.98 × 10−6 2.50 × 10−6 2.39 × 10−6 2.54 × 10−6

105.00 1.70 × 10−6 3.36 × 10−6 2.98 × 10−6 2.48 × 10−6 2.44 × 10−6 2.46 × 10−6

157.50 1.70 × 10−6 3.36 × 10−6 2.94 × 10−6 2.48 × 10−6 2.44 × 10−6 –210.00 1.70 × 10−6 3.34 × 10−6 2.94 × 10−6 2.48 × 10−6 2.42 × 10−6 –262.50 1.69 × 10−6 3.32 × 10−6 2.93 × 10−6 2.51 × 10−6 2.42 × 10−6 –315.00 1.69 × 10−6 3.30 × 10−6 2.91 × 10−6 2.49 × 10−6 2.39 × 10−6 2.52 × 10−6

367.50 1.68 × 10−6 3.30 × 10−6 2.93 × 10−6 2.47 × 10−6 2.34 × 10−6 –4 −6 −6 × 10−6 −6 −6

n

urfac

cEcadtttad1Etcd

3

3

aaab

dtaapp(deTttietps(

gp

TH

C

C

C

C

R

S

20.00 1.69 × 10 3.33 × 10 2.90e− 1 1.97 1.74

a Some data undetectable because of adsorption phenomena at the electrode s

alculated using the WinSim program in the NIEHS publicSR software tools package (http://epr.niehs.nih.gov/). Radi-al anions of caffeic acid, caffeic acid ethyl ester, rosmariniccid, chlorogenic acid and 4-tert-butylcathecol were generatedirectly in the ESR tube by adding to a solution of 2 mg ofhe studied compound in 0.5 mL of DMSO 1 mg of potassiumert-butoxide. Traces of PbO2 were added in order to oxidizehe anion thus formed. Alternatively, they were generated bydding 3–4 mg of K3(FeCN)6 to a solution of 2 mg of HCAissolved in 1 mL NaOH 10%. The mixture was shaken withmL of CHCl3 and the aqueous phase was transferred into anSR tube to record the spectrum. This procedure was also used

o originate the phenoxyl radical of sinapinic acid. Hyperfineoupling constants (h.f.c.cs.) and g-factors for all the radicalsetected are reported in Table 3.

. Results and discussion

.1. Cyclic voltammetry

The electrochemical behaviour of caffeic acid (CA), caffeic

cid ethyl ester (CAE), chlorogenic acid (CGA), rosmariniccid (RA), ferulic acid (FA), sinapinic acid (SA) and cumariccid (CUA) (see Chart 1), as well as cathecol (C) and 4-tert-utylcatechol (TBC) as semplified models, was studied in anhy-

rb(w

able 3yperfine coupling constants in Gauss and g-factors of the radicals obtained from the

ompound Radical detected H.f.c.cs.

H� H�

A Radical anion 1.965 0.864Cyclic radicala

AE Radical anion 1.971 0.841Cyclic radicala

GA Radical anion 2.069 0.858

A Radical anion 3.196 0.826Cyclic radicala

A Phenoxyl radical 3.683Cyclic radicalb

a H.f.c.cs. for H5 and H6 have been arbitrary assigned.b Not identified radical.

2.48 × 10 2.31 × 10 –1.47 1.41 1.48

e.

rous acetonitrile at a static GC electrode in the absence and inhe presence of 2,6-lutidine as deprotonating agent [21,22]. Thenodic potential Eap values (see Table 1) range between +1.20nd +0.77 V versus Ag/AgClO4. The higher values for oxidationotential of mono-HCAs well correlate with those of mono-henols; the negative shift observed for FA (+0.91 V) and SA+0.78 V) with respect to CUA (+1.20 V) is due to the electron-onating effect of metoxyl groups. Lower potential values werexpected and actually found for di-phenols; the cathodic shift ofBC (+0,77 V) and di-HCAs with respect to C (+0.85 V) is due

o the electron-donating effect respectively of the tert-butyl andhe unsaturated side-chain in para position. Small differencesn the anodic potentials for the different di-HCAs could bevidenced, explained by the possibility of secondary interac-ions (hydrogen bonds) between the various functional groupsresent in the studied molecules; in the presence of a base whichuppresses all these interactions, these differences disappearsee Table 1).

All studied compounds exhibit similar cyclic voltammo-rams, characterized by a well defined irreversible anodiceak and a corresponding broadened cathodic wave in the

everse scan. In particular, in Fig. 1 (black line) the irreversiblei-electronic anodic peak at Eap = +0.79 V of CA is shownthe number of electrons was determined by comparisonith ferrocene under the same experimental conditions). The

studied compounds

g-Factor Solvent

H2 H5 H6

1.041 1.417 2.702 2.00495 DMSO2.127 0.860 0.492 2.00430 H2O

1.151 1.699 2.255 DMSO2.127 0.860 0.492 2.00430 H2O

1.335 1.550 2.335 2.00505 DMSO

1.280 1.903 3.335 2.00496 DMSO2.127 0.860 0.492 2.00430 H2O

0.848 0.848 2.00472 H2O0.547 0.547 2.00480 H2O

Page 4

2464 R. Petrucci et al. / Electrochimica Acta 52 (2007) 2461–2470

art 1.

ctif

FN(

Ch

orresponding broadened cathodic wave was observed inhe reverse scan at Ecp = +0.36 V, whose height (ic/ia = 0.25)ncreased of about 35% when the solution was electrolysedor 30 s at the starting potential of +1.0 V before the cathodic

ig. 1. Cyclic voltammetry of CA in anhydrous acetonitrile 0.1 mol L−1

aClO4, at GC electrode, vs. Ag/AgClO4, scan rate 0.200 V s−1, in the absenceblack line) and in the presence (dotted lines) of 2,6-lutidine.

si

adhiowivwwT(ntpa[

it[fObiw[

vwCs

can, suggesting that an oxidized species of CA must benvolved.

When increasing amounts of 2,6-lutidine as deprotonatinggent were added to the tested solution, the first anodic peakecreased, while a second anodic peak (Fig. 1) with increasingeight was observed at a less positive potential, indicating anncrease in the nucleophilicity of the compound [23]. In the casef CA, the complete disappearance of the first peak at +0.79 Vas observed upon addition of 1.5 equivalents of base, indicat-

ng that CA is partially dissociated even in acetonitrile (a pKaalue of 5.46 is reported for CA in a 30% acetonitrile solution inater [24]); for CAE, C and TBC two equivalents of 2,6-lutidineere necessary to completely shift the anodic peak (see alsoable 1). A cathodic shift was also observed in the reverse scanfor CA, Ecp = −0.30 V), the ratio ic/ia increased to 0.71. Theew anodic peak likely corresponds to the bi-electronic oxida-ion of the phenolate mono-anion to the corresponding quinoidroduct, as supported by the increase in the ic/ia ratio as wells by the Ecp value, falling in the typical range of quinones25].

The voltammetric behaviour observed for the studied HCAss in good agreement with that found in literature for the elec-rochemical oxidation of CA, although in aqueous solutions10], and with that reported by Parker and Eberson [26–28]or the anodic oxidation of hydroquinone in aprotic medium.n the other hand, Eggins and Chambers [29,30] reported at theeginning of the 70’s for hydroquinone an oxidation mechanismnvolving the formation of a semiquinone intermediate, whichas subsequently suggested again also for CA by Hapiot et al.

12].Since we never observed any intermediate and in order to

erify the number of electrons involved in the anodic process,e also studied the electrochemical oxidation of C, TBC, CA,AE and FA at a GC rotating disk electrode (RDE) in the

ame medium and using ferrocene as a reference. RDE data

Page 5

imica Acta 52 (2007) 2461–2470 2465

(scrtso

3

cai

vy(atfuom

epsλ

favtto(wwcSta

ttλ

wm

maettdat

Fig. 2. UV spectral changes relative to the anodic oxidation and successivereversible cathodic reduction of C (a), CA (b) and FA (c) in anhydrous acetonitrile0 −1

ap

ppfae

R. Petrucci et al. / Electroch

see Table 2) were elaborated by Levich equation assuming theame diffusion coefficient D for all studied compounds and dis-ussed according to criteria suggested by literature [28]. RDEesults confirm a bi-electronic process for C and TBC and a lesshan two-electrons process found for CA, CAE and FA whichuggests a chemical step interfering with the electrochemicalne [27,31].

.2. Electrolyses in UV cell

Controlled potential electrolyses of different substrates werearried out in anhydrous acetonitrile in a modified UV cell atplatinum electrode and spectra were recorded at short time

ntervals during electrolysis.Electrolysis of C was carried out at +0.9 V till the current

alue reached a constant plateau, corresponding to 1.9 e−. Theellowish final solution showed an UV spectrum with λmaxMeCN)/nm 250 and 378. This latter solution was then reducedt a cathodic potential of about 0 V, and after the consump-ion of an equal amount of current (1.9 e−) the starting reducedorm of C was yielded (Fig. 2a). The UV spectrum of the prod-ct obtained from the reversible bi-electronic anodic oxidationf C has been taken as a reference for the o-benzoquinoneoiety.Under the same experimental conditions, controlled potential

lectrolysis of CA was carried out at +0.9 V, till a current constantlateau corresponding to 1.6 e−; the yellow-green final solutionhowed an UV spectrum very similar to that obtained for C, withmax(MeCN)/nm 248, 300 and 394, and this product was stableor at least half an hour. This final solution was then reduced atcathodic potential of about −0.5 V, the experimental potentialalue found at the platinum electrode used for electrolyses. Inhis case, the cathodic current consumption was slightly smallerhan the anodic one and the UV spectrum of the final solutionverlapped the starting one, even if with a lower absorbanceFig. 2b). Neither stable intermediate nor transient compoundsere detected in the oxidation or reduction process (in a previousork of us, the semiquinone of p-benzoquinones electrochemi-

ally generated was evidenced via UV spectrophotometry [25]).imilar results were obtained for the electrolysis of CAE under

he same experimental conditions (λmax(MeCN)/nm 250, 302nd 394).

Controlled potential electrolysis was carried out also onhe mono-hydroxycinnamic acid FA, and after the consump-ion of 1.7 e− (Fig. 2c) yielded a stable oxidation product withmax(MeCN)/nm 248 and 392. The reversibility of the processas evidenced also in this case. The bi-electronic oxidation ofono-phenols is also reported in literature [32].From our electrochemical and spectroelectrochemical data, it

ay be assumed that the anodic oxidation in aprotic medium ofll studied compounds proceeds via a reversible one-step two-lectrons mechanism coupled with a proton transfer and leadingo the corresponding phenoxonium cation. The irreversibility of

he anodic process evidenced in cyclic voltammetry could beue to a slow proton exchange equilibrium in aprotic medium,s also supported by the broadened shape of the cathodic waves,he dependence of the Ecp values as well as the cathodic–anodic

aoor

.1 mol L NaClO4, at a Pt working electrode: starting substrates, spectra 1;nodic oxidation products, spectra 2; reversible reduction products, spectra 3;artial reversible reduction process of C, spectrum 4.

eak current ratio from scan rates and from the absence or theresence of a deprotonating agent. The anodic oxidative processor CA is summarized in Scheme 1. The polarographic study atstatic or rotating disk electrode evidenced that less than two

lectrons are actually involved in the oxidation of HCAs, as

lso confirmed by coulometric data. These results suggest theccurrence of competitive chemical processes between the firstxidation product (the phenoxonium cation) and the startingeduced form, probably involving the unsaturated side-chain.

Page 6

2466 R. Petrucci et al. / Electrochimica Acta 52 (2007) 2461–2470

eme 1

ifuU

h

Fs

Sch

The main product obtained in the anodic oxidation of all stud-

ed compounds is the corresponding o-quinone in a protonatedorm (Scheme 1). In literature, the CA quinone is reported to benstable [10,12], whereas in our experiments the correspondingV spectrum has been recorded without changes for at least

ambc

ig. 3. UV spectral changes relative to the chemical oxidation with PbO2 in anhydroubstrates, spectra 1; oxidation products, spectra 2.

.

alf an hour. Since o-quinone of CA is sufficiently stable only in

cid solution [10], it may be concluded that in our experiments itost likely exists as a protonated form, which is particularly sta-

ilized by resonance effects. The stability of the protonated formould also suggest the formation of cyclic products, as shown in

us acetonitrile of: C (a), CAE (b), CA (c), CUA (d), FA (e) and SA (f): starting

Page 7

R. Petrucci et al. / Electrochimica Acta 52 (2007) 2461–2470 2467

eme 2

StCctE

3

owtapstpttnmo

cw(ctdmt

Io(cp[pi[3miocmssnii

Fsotb

3

Sch

cheme 1, and reported in literature for CUA [33]. Noteworthy,he absorption observed at λmax(MeCN)/nm 300 for electrolysedA is very similar to the value calculated for a six-memberedyclic �,�-unsaturated ketone �-hydroxy substituted [34]. Fur-hermore, the same cyclic product has been detected also in theSR experiments, as discussed below.

.3. Chemical oxidation

All electrochemically studied compounds were chemicallyxidized with PbO2 in anhydrous acetonitrile and the reactionsere monitored via UV spectrophotometry (Fig. 3). The oxida-

ion products of C and di-HCAs, except CA, have characteristicbsorptions at about 250 nm and between 380 and 390 nm. Inarticular, the UV spectrum of the oxidated C (Fig. 3a) is theame when electrochemically or chemically produced, while inhe case of CAE (Fig. 3b) the UV spectrum of the chemicalroduct shows a blue shift (λmax(MeCN)/nm 382) with respecto the electrochemical one (λmax(MeCN)/nm 394). This seemso further support that the electrochemical product is the proto-ated form of the corresponding o-quinone, characterized by aore extended conjugation than the neutral form obtained after

xidation with PbO2.Oxidation of CA did not result in the formation of a quinoid

ompound: all starting adsorbances decreased and a productith a weak absorption in the λ range 480–500 nm was detected

Fig. 3c). According to literature [12,35,36], such an absorbanceould be attributed to the phenoxyl radical, but very unlikely

his radical could be stable enough, in these conditions, to beetected, as also supported by ESR data (see below). Whenono-HCAs were oxidized with PbO2, absorptions stronger

han for CA but in the same range of λ values were observed.dH

.

n particular, UV spectra evidenced adsorptions in the λ rangef 430–460 nm for CUA and 470–510 nm for FA and SAFig. 3d–f). This kind of absorption could be due to an extendedonjugation of dimerization products involving the side-chain,ossible when neutral phenoxyl radicals are initially formed37–39], as shown in Scheme 2. Formation of dimerizationroducts is further supported by HPLC–MS data, evidencingn all cases the presence of the corresponding molecular ioneDim-H]−, respectively, m/z 357 for CA, m/z 325 for CUA, m/z85 for FA and m/z 445 for SA, as shown in Fig. 4. Further-ore, in all cases the presence of the two carboxylic groups

s also supported by the lost (even at the low cone voltagef 25 V) of two successive m/z 44 fragments, typical for thearboxylic function. All the compounds showed in their cro-athograms, besides the peak corresponding to the unreacted

tarting substrate, another peak having the same RT and theame m/z value 293, not containing carboxylic groups, whoseature has to be clarified, at present (it could be due to a start-ng impurity common to all HCAs). No quinoid structure wasdentified.

The increasing intensity of the UV adsorptions from CUA,A and SA could be due to the increasing stability of the corre-ponding starting phenoxyl radical. Regard to CA, the formationf the unstable neutral o-quinone as main product with PbO2,hat can further fastly react giving different products [15], cannote excluded.

.4. ESR

All products were also studied by ESR spectroscopy. Oxi-ation was carried out in an alkaline milieu using DMSO or2O/NaOH 10% as the solvent; although the experiments were

Page 8

2468 R. Petrucci et al. / Electrochimica Acta 52 (2007) 2461–2470

F obtas

cits

truFsfp

Fco

rtswvaisame product was formed from the evolution of the three differ-ent radical anions (in Fig. 6, the spectrum recorded with CAE is

ig. 4. The negative ion ESI mass spectra of dimerization products [Dim-H]−eparation of CA, FA, SA and CUA.

arried out in different conditions with respect to electrochem-cal and chemical oxidation, they can give useful informationo clarify the reactivity of the compounds used in the presenttudy.

Di-HCAs (CA, RA, CGA) and CAE, when oxidized, gave ini-ial well resolved and intense ESR spectra of the correspondingadical anion. Experimental ESR spectrum and computer sim-lation of the radical anion obtained from CGA are reported inig. 5. As shown in Table 3, g-factors for all these radicals have

imilar values, indicating that the same kind of radical is formedrom the different derivatives. The values for the hyperfine cou-ling constants of the phenoxyl radical anion obtained from CA

ig. 5. ESR experimental spectrum (a) and computer simulation (b) of the radi-al anion obtained from CGA by oxidation with PbO2 in DMSO in the presencef t-BuOK.

st

FrT

ined after chemical oxidation with PbO2 in anhydrous acetonitrile and HPLC

eported in literature [40] slightly differ from those found inhis work, being the differences most likely due to the particularolvent in which the spectra are recorded, namely DMSO andater. The radical initially detected evolves into another species,ery rapidly in the case of CA, more slowly for CAE and onlyfter several hours for CGA. In particular, the spectra recordedn the three cases are exactly superimposable suggesting that the

hown). In our opinion, this new spectrum could be likely dueo a cyclization product obtained from the further oxidation of

ig. 6. ESR experimental spectrum (a) and computer simulation (b) of the cyclicadical anion recorded after oxidation of CAE by K3(FeCN)6 in NaOH/H2O.he same radical was obtained from CA and RA.

Page 9

R. Petrucci et al. / Electrochimica

tatcfooisa

ddsembc[swpramss

4

wbeof

dh

lbcetcunforOsb

A

e

R

[

[

[

[

[

[[

[[

[69 (1979) 583.

Scheme 3.

he quinonic compound initially formed from the radical anion,s shown in Scheme 3. This hypothesis is supported by the facthat longer time is required for the formation of this new radi-al from CAE and CGA; in these cases hydrolysis of the esterunction is necessary in order to have cyclization. Dimerizationf phenoxyl radical followed by cyclization is considered onef the most common mechanism of oxidation for CA and ofts derivatives [13,15]. However, this is not our case since thesetructures cannot justify the hyperfine splitting observed whichre in agreement with the cyclic product shown in Scheme 3.

Mono-HCAs (CUA, FA and SA) were also tentatively oxi-ized, but only the phenoxyl radical generated from the oxi-ation of SA was detected. This is most likely due to its highertability with respect to the other two species because of the pres-nce of the methoxyl groups in the o-positions to the hydroxyloiety. The instability of these kind of radicals is confirmed

y the fact that ESR data found in literature for this kind ofompounds have been obtained using the flow system technique41–42]. The SA phenoxyl radical fastly evolves into anotherpecies characterized by a well defined triplet of intensity 1:2:1,hose hyperfine coupling constants are reported in Table 3. Atresent, we are not able to establish the structure of this newadical but, on the basis of its hyperfine splitting, it seems to bep-semibenzoquinone anion which could take place by replace-ent of the alkyl chain by an –OH group; this cannot be excluded

ince oxidation of SA was performed in alkaline aqueousolutions.

. Conclusions

Electrochemical and spectroelectrochemical data show thathen HCAs act as electron donors, in a proton poor medium,

oth mono- and di-phenols undergo a reversible one-step two-lectrons anodic process to the corresponding stable phenox-nium cation, which might further react with the very likelyormation of a heterocyclic compound, as observed also in ESR

[[[

Acta 52 (2007) 2461–2470 2469

ata, evidencing the existence of a radical species from the sameeterocyclic compound.

When HCAs are working as H-atom donors, different pathseading to different oxidation products may occur, accordingoth to the nature of starting phenols and the stability of theorresponding neutral phenoxyl radical: while di-phenols gen-rally yielded the corresponding neutral o-quinones (donatingwo H atoms), mono-phenols yielded dimerization productsharacterized by an extended conjugation system involving thensaturated side-chain and the regeneration of the reduced phe-olic function. No quinoid structure products were observedor mono-phenols. In the case of the di-phenol CA, the neutral-quinone was not evidenced, but we remember that it shouldapidly evolve to form other products, as reported in literature.n the other hand, the CA phenoxyl radical must be enough

table to undergo a competitive radical coupling, as evidencedy its dimerization product.

cknowledgment

We thank the Ministero dell’Universita e della Ricerca Sci-ntifica e Tecnologica for financial support.

eferences

[1] L.A. Bazzano, J. He, L.G. Ogden, C.M. Loria, S. Vupputuri, L. Myers, P.K.Whelton, Am. J. Clin. Nutr. 76 (2002) 93.

[2] G. Block, B. Patterson, A. Subar, Nutr. Cancer 18 (1992) 1.[3] C.S. Yang, J.M. Landau, M.T. Huang, H.L. Newmark, Annu. Rev. Nutr. 21

(2001) 381.[4] C.A. Rice-Evans, N.J. Miller, G. Paganga, Free Radic. Biol. Med. 20 (1996)

933.[5] C. Castelluccio, G. Paganga, N. Melikian, G.P. Bolwell, J. Pridham, J.

Sampson, C. Rice-Evans, FEBS Lett. 368 (1995) 188.[6] M. Nardini, M. D’Aquino, G. Tomassi, V. Gentili, M. Di Felice, C. Scaccini,

Free Radic. Biol. Med. 19 (1995) 541.[7] O. Firuzi, L. Giansanti, R. Vento, C. Seibert, R. Petrucci, G. Marrosu, R.

Agostino, L. Saso, J. Pharm. Pharmacol. 55 (2003) 1021.[8] A.S. Pannala, R. Razaq, B. Halliwell, S. Singh, C.A. Rice-Evans, Free

Radic. Biol. Med. 24 (1998) 594.[9] J.H. Chen, C. Ho, J. Agric. Food Chem. 45 (1997) 2374.10] C. Giacomelli, K. Ckless, D. Galato, F.S. Miranda, A. Spinelli, J. Braz.

Chem. Soc. 13 (2002) 332.11] S. Kallel Trabelsi, N. Belhadj Tahar, R. Abdelhedi, Elecrochim. Acta 49

(2004) 1647.12] P. Hapiot, A. Neudeck, J. Pinson, H. Fulcrand, P. Neta, C. Rolando, J.

Electroanal. Chem. 405 (1996) 169.13] H. Hotta, H. Sakamoto, S. Nagano, T. Osakai, Y. Tsujino, Biochim. Bio-

phys. Acta 1526 (2001) 159.14] H. Fulcrand, A. Cheminat, R. Brouillard, V. Cheynier, Phytochemistry 35

(1994) 499.15] J.L. Cilliers, V.L. Singleton, J. Agric. Food Chem. 39 (1991) 1298.16] S. Foley, S. Navaratnam, D.J. MvGarvey, E.J. Land, T.G. Truscott, C.A.

Rice-Evans, Free Radic. Biol. Med. 26 (1999) 1202.17] W.A. Waters, J. Chem. Soc. B (1971) 2026.18] L. Papouchado, R.W. Sandford, G. Petrie, R.N. Adams, J. Elecroanal.

Chem. 65 (1975) 275.19] R. Andruzzi, A. Trazza, L. Greci, L. Marchetti, Ann. Chim. (Rome, Italy)

20] R. Kuhn, I. Hammer, Chem. Ber. 83 (1950) 413.21] V.D. Parker, L. Eberson, Acta Chem. Scand. 24 (1970) 3542.22] C. Amatore, G. Farsang, E. Maisonhaute, P. Simon, J. Electroanal. Chem.

462 (1999) 55.

Page 10

2 imica

[

[

[

[[[[[[

[

[

[

[

[

[[

[

[

470 R. Petrucci et al. / Electroch

23] D.T. Sawyer, A. Sobkowiak, J.L. Roberts Jr., Electrochemistry forChemists, second ed., John Wiley & Sons, Inc., New York, 1995, p. 442.

24] J.L. Beltran, N. Sanli, G. Fourodona, D. Barron, G. Ozkan, J. Barbosa,Anal. Chim. Acta 484 (2003) 253.

25] R. Petrucci, E. Giorgini, E. Damiani, P. Carloni, G. Marrosu, A. Trazza,G.P. Littarru, L. Greci, Res. Chem. Intermed. 26 (2000) 269.

26] V.D. Parker, Chem. Commun. (1969) 716.27] V.D. Parker, L. Eberson, Chem. Commun. (1970) 1289.28] V.D. Parker, Electrochim. Acta 18 (1973) 519.29] B.R. Eggins, J.Q. Chambers, J. Electrochem. Soc. 117 (1970) 186.30] B.R. Eggins, Chem. Commun. (1972) 427.31] R.N. Adams, Electrochemistry at Solid Electrodes, Marcel Dekker, New

York, 1969, p. 244.

32] S.D. Ross, M. Filkelstein, E.J. Rudd, in: A.T. Blomquist, H.H. Wasserman

(Eds.), Organic Chemistry, vol. 32, Academic Press, Inc., New York, 1975,p. 271.

33] A.I. Scott, P.A. Dodson, F. McCapra, M.B. Meyers, J. Am. Chem. Soc. 85(1963) 3702.

[

[

Acta 52 (2007) 2461–2470

34] R.M. Silverstein, G.C. Bassler, T.C. Morrill, Spectrometric Identificationof Organic Compounds, fifth ed., John Wiley & Sons Inc., Singapore, 1991,p. 301.

35] P. Hapiot, J. Pinson, C. Francesch, F. Mhamdi, C. Rolando, P. Neta, J. Phys.Chem. 98 (1994) 2641.

36] P. Hapiot, J. Pinson, P. Neta, C. Francesch, F. Mhamdi, C. Rolando, S.Schneider, Phytochemistry 36 (1994) 1013.

37] L.J. Filar, S. Winstein, Tetrahedron Lett. 25 (1960) 9.38] P. Wan, D.W. Brousmiche, C.Z. Chen, J. Cole, M. Lukeman, M. Xu, Pure

Appl. Chem. 73 (2001) 529.39] H.-H. Perkampus, UV–vis Atlas of Organic Compounds, second ed., VHC,

Weinheim, 1992, Part 2, p. 1175.40] W. Bors, C. Michel, K. Stettmaier, Y. Lu, L.Y. Foo, Biochim. Biophys. Acta

1620 (2001) 97.41] W.T. Dixon, M. Moghimi, D. Murphy, J. Chem Soc. Perkin Trans. 2 (1975)

1189.42] W.R. Russell, A.B. Hanley, M.J. Burkitt, A. Chesson, Bioorg. Chem. 27

(1999) 339.