J. Agric. Food Chem. 2000, 48, 3597-3604
`
`3597
`
`Assessing Antioxidant and Prooxidant Activities of Phenolic
`Compoundsi
`
`L. R. Fukumoto and G. Mazza*
`
`Food Research Program, Pacific Agri—Food Research Centre, Agriculture and Agri—Food Canada,
`Summerland, British Columbia VOH 1Z0, Canada
`
`Methods for determining primary antioxidant activity were evaluated. A fl—carotene bleaching method
`and a free radical method using 2,2—diphenyl-1—picry1hydrazyl (DPPH') were modified to rapidly
`test samples for potential antioxidant activity. Malonaldehyde production in a linoleic acid emulsion
`system assayed by an HPLC method was also used to determine antioxidant and prooxidant activities
`initiated by a metal catalyst (Cu2+). All methods were used to assess activity of selected phenolic
`compounds including several anthocyanidins/anthocyanins and selected berry extracts. Most phenolic
`compounds had prooxidant activity at low concentrations, unlike synthetic antioxidants (BHA and
`BHT). Compounds with similar structures exhibited comparable trends in antioxidant activity.
`Antioxidant activity usually increased with an increase in the number of hydroxyl groups and a
`decrease in glycosylation.‘ The antioxidant activity of many phenolic compounds and extracts was
`comparable to those of synthetic antioxidants using the [3-carotene bleaching and HPLC methods.
`
`Keywords: Antioxidant activity,’ prooxidant activity,’ phenolics,’ flavonoids,‘ anth0cyan1'ns,' berry
`extracts
`
`INTRODUCTION
`
`Lipid oxidation occurs when oxygen reacts with lipids
`in a series of free radical chain reactions that lead to
`complex chemical changes. Oxidation of lipids in foods
`causes quality losses. In vivo, lipid oxidation may play
`a role in coronary heart disease, atherosclerosis, cancer,
`and the aging process (Jadhav et al., 1996).
`Antioxidants are compounds that can delay or inhibit
`lipid oxidation. When added to foods, antioxidants
`minimize rancidity, retard the formation of toxic oxida-
`tion products, maintain nutritional quality, and increase
`shelf life (Jadhav et al., 1996). Recently, interest has
`been growing in finding naturally occurring antioxi-
`dants for use in foods to replace synthetic antioxidants
`and for possible in vivo use. As one potential source,
`plant phenolics have primary (chain—breaking) antioxi-
`dant activity (Shahidi and Wanasundara, 1992). To
`evaluate compounds for antioxidant activity, a reliable
`in vitro method is needed.
`
`Most antioxidant activity assays consist of accelerat-
`ing oxidation in a lipid system, usually by heat, and then
`monitoring oxygen consumption, substrate loss, or
`product formation. Because many factors affect oxida-
`tion,
`including temperature, oxygen pressure, metal
`catalysts, fat composition, and form of fat, results can
`vary depending on the oxidation conditions used (Frankel,
`1993). Assays to measure substrates or products can
`also give varying results depending on their specificity.
`Osawa and Shibamoto (1992) developed a high—perfor—
`mance liquid chromatography (HPLC) method to mea—
`sure malonaldehyde formed in lipid emulsion systems
`oxidized by FeCl2/H202. Malonaldehyde was derivatized
`by reaction with urea under acidic conditions to form
`2—hydroxypyrimidine, which could be measured by
`
`* Author to whom correspondence should be addressed [fax
`(250) 49-1-0755; e—mai1 Mazzag@em.agr.ca].
`1 Pariflc Agri—Food Research Centre Contribution 2061.
`
`HPLC. Tsuda et al. (1994) used this method to measure
`malonaldehyde formed in various lipid systems. Because
`the HPLC method is specific for malonaldehyde, com-
`bining this method with a model lipid oxidation system
`could be a good assay for antioxidant activity.
`Methods relying on oxidation can be time—consuming
`to perform depending on conditions used, whereas
`fl—carotene bleaching and reacting compounds with free
`radicals are quick and simple methods of measuring
`potential antioxidant activity. Marco (1968) described
`the use of ,8—carotene bleaching for ranking compounds
`for antioxidant activity. In this method, antioxidant
`activity is measured by the ability of a compound to
`minimize the loss of ,6—carotene during the coupled
`oxidation of linoleic acid and fl—carotene in an emulsified
`aqueous system. The reaction is usually initiated using
`heat
`(50 °C). Although the method is simple and
`sensitive,
`it was criticized by Frankel
`(1993) for its
`nonspecificity, being subject to interference from oxidiz-
`ing and reducing agents in crude extracts and linoleic
`acid not being representative of typical food lipids.
`Two free radicals that have been used for assessing
`antioxidant activity are 2,2’—azinobis(3—ethyl—benzothi—
`azoline—6—sulfonic acid)
`(ABTS'+) and 2,2—diphenyl—1—
`picrylhydrazyl (DPPH'), also known as 1,1—diphenyl—2—
`picrylhydrazyl or o.,o.—diphenyl~)6-picrylhydrazyl. Reduc-
`tion of DPPH' by- an antioxidant
`(DPPH‘ + A ->
`DPPH—H + A’) or by a radical species (DPPH' + R’ -’
`DPPH—R) results in a loss of absorbance at 515 nm.
`Brand—Williams et al. (1995), using DPPH', developed
`a spectrophotometric method that gave results similar
`to an oxidation method, but comparisons were not
`quantitative because reaction with DPPH' depended on
`a compound's structural conformation.
`Compounds with antioxidant activity may exhibit
`prooxidant behavior under certain conditions. Prooxi-
`dant activity can accelerate damage to molecules such
`as DNA, carbohydrates, or proteins (Aruoma et al.,
`
`10.1021/jf000220W CCC: $19.00
`Published 2000 b lhe American Chemical Society
`Published on Web 07106 2000
`
`Page 1 of 8
`
`SENJU EXHIBIT 2269
`Innopharma v Scnju,
`IPR20l5-00902 & IPR2015-00903
`
`Page 1 of 8
`
`

`
`3598
`
`J. Agric. Food Chem, Vol. 48, No. 8, 2000
`
`Fukumolo and Mazza
`
`PIIENOLIC ACIDS
`Bcn/.oIc ncld dcnrnln cs
`7
`
`I
`
`COOH
`
`Gnlllcacid
`
`Pmlncnlccluuc ncld
`3-H\dro.\1bcn'/.o1c and
`
`~.
`I = OH.
`
`I = H
`l = H
`
`2 = oH
`
`2 = OH
`2 = H_
`
`OH
`
`= = OH
`
`3 =
`3 = OH
`
`Cinmnuc ncid dcm nmcs
`OH
`
`Chlorogcmc flCld
`
`no
`
`/ \
`--
`
`I
`
`0
`\—<
`
`U
`
`FLAVONOLS
`
`(OCH OH
`
`O"
`
`OH
`
`0
`
`Synngic acid
`Vnmllic acid
`
`I 4 OCH,
`= H
`
`2 = OH.
`2 = OH.
`
`1 = OCH‘ 2:‘
`3 = ocn,
`
`"L
`
`C00“
`
`0H
`
`M_\ ncclln:
`Oucrcclm
`
`3‘ = OH
`3‘ = OH
`
`-I‘ = OH.
`4' = OH.
`
`5' = OH.
`5' = H.
`
`‘u= OH
`3: OH
`
`Cnffclc acid
`p-Counmric .‘lCld
`Clmmmc ilCldI
`
`I = OH.
`|: H.
`I = H
`
`Fcnllic ncid:
`FLAVANONE
`
`I = OCH,
`OH
`
`2 : OH
`2 : OH
`2 = H
`
`2 = OH
`
`1 = OH
`Rulm;
`Kncmpfcrolz 3' = H.
`
`4‘ = OH
`4‘ = OH
`
`i = H
`5‘ = H
`OH
`
`FLAVANOLS
`‘IH
`
`1 = nlllmsl.
`'4 = OH
`
`
`
`
`Nnnngcnin
`
`(+)-Cnlcclunor
`(-)—Epicnlccl|in
`
`H
`
`STANDARDS
`cu,
`
`Emglc and
`
`‘IE-1
`ANTIIOCYANIDINS / ANTHOCYANINS
`
`a..i
`
`IIC
`
`
`Dclplmidm 3‘ = OH.
`C_\'.'Inmdi1t
`3' = OH. '
`
`OH
`4
`4 = OH.
`
`5 = OH
`% = H
`
`3 = OH
`3 : OH
`
`3‘ = H.
`Pclnrgonidin
`.‘v‘=OCH,
`Maludln:
`3 = OCH,.
`Pconidm:
`For 3-glucosidcs:
`
`5‘ : H.
`4 = OH.
`5‘=OCH,.
`4 =OH
`5' = H
`4 = OH
`.1 = glucosndc
`
`3 = OH.
`”v=OH.
`3: OH.
`
`4 = OH
`‘=OH
`5 = OH
`
`5 = glucosudc
`3 : glucosidc
`For .‘\.5-diglucosldcs
`Figure 1. Structures of compounds tested.
`
`Ascorbic acid
`D
`
`
`
`0H
`
`(cH3}3(-
`
`(CHJHF
`
`L‘t('H3)3
`
`BHA
`
`BHT
`
`OCH3
`
`CH3
`
`1997). Potential antioxidants should therefore be tested
`for prooxidant activity as well. The deoxyribose, iron-
`bleomycin—DNA, and copper~1,10 phenanthroline—
`DNA assays have been used as prooxidant tests (Aruo—
`ma et al., 1997). Prooxidant activity has also been
`measured using metal catalysts in a fl—carotene/linoleic
`emulsion system (Pischetsrieder et al., 1998) and using
`a Cu“ catalyst in an oxygen radical absorbance capacity
`assay (Cao et al., 1997).
`The objectives of this study were to (1) adapt a
`fl—carotene bleaching method and a DPPH' method as
`fast screening assays for potential antioxidant activity
`using microplates, (2) develop an oxidation system using
`a linoleic acid emulsion so both prooxidant and anti-
`oxidant activities could be assessed by an HPLC method
`that measures malonaldehyde, (3) measure the antioxi-
`dant and prooxidant activities of several phenolic com-
`pounds including some anthocyanidins/anthocyanins by
`all methods to compare results and evaluate if the
`results can be related to compound structures and
`literature results, and (4) measure the antioxidant and
`prooxidant activities of selected berry extracts by all
`methods to compare results and determine if the
`methods are affected by various compounds present in
`extracts.
`
`MATERIALS AND METHODS
`
`(+)-
`Chemicals. L—Ascorbic acid, 3-hydroxybenzoic acid,
`catechin. 4—hydroxy—3—methoxycinnamic acid (ferulic acid),
`
`gallic acid, 4’,5,7-trihydroxyflavanone (naringenin), and rutin
`were purchased from Aldrich (Sigma-Aldrich Canada Ltd.,
`Oakville, ON, Canada). Cyanidin chloride, cyanidin 3-glucoside
`(kuromanin) chloride, Cyanidin 3,5—diglucoside (cyanin) chlo-
`ride, delphinidin chloride, malvidin chloride, malvidin 3—glu—
`coside (oenin) chloride, pelargonidin chloride, pelargonidin 3,5-
`diglucoside (pelargonin) chloride, peonidin chloride, and peonidin
`3-glucoside chloride were obtained from Extrasynthese (Cenay,
`France). Pelargonidin 3-glucoside (callistephin) chloride was
`purchased from Carl Roth (Karlsuhe, Germany). Malvidin
`3-glucoside chloride was also obtained from Professor R.
`Brouillard (Université Louis Pasteur, Strasbourg, France).
`0L,o.’—Azodiisobutyramidine dihydrochloride (ADIBA), cinnamic
`acid, kaempferol,
`linoleic acid, and myricetin were from
`Fluka (Sigma-Aldrich Canada Ltd.). Butylated hydroxyanisole
`(BHA), butylated hydroxytoluene (BHT), caffeic acid, [3-caro-
`tene, chlorogenic acid, pcoumaric acid, 1,1—dipheny1—2—picryl-
`hydrazyl (DPPI-I‘), ellagic acid,
`(—)—epicatechin, 2-hydroxy-
`pyrimidine, malvidin 3,5—diglucoside (malvin) chloride, proto-
`catechuic acid, quercetin, syringic acid, 1.1 ,3,3-tetraethoxypro-
`pane (TEP), (1-tocopherol, and vanillic acid were from Sigma
`Chemical Co. (Sigma-Aldrich Canada Ltd.). Tween 20 was
`obtained from BDH Chemicals (Toronto, ON, Canada). The
`structures of compounds tested for antioxidant and prooxidant
`activity are shown in Figure 1. Test compounds were prepared
`in methanol.
`
`fl-Carotene Bleaching Method. The fl—carotene bleaching
`methods of Marco (1968) and Velioglu et al.
`(1998) were
`modified for use with microplates. The modification consisted
`of preparing a mixture of 1 mL of [J‘~carotene (2 mg/mL in
`chloroform), 0.2 mL of linoleic acid, and 2 mL of Tween 20.
`The mixture was vortexed, and chloroform was removed using
`
`Page 2 of 8
`
`Page 2 of 8
`
`

`
`Antioxidant and Prooxidant Activities of Phenolics
`
`a stream of nitrogen for 1-1.5 h. Air-sparged distilled water
`(20 mL) was then added to the mixture, which was subse-
`quently vortexed to form a clear solution. The volume of
`solution was sufficient for 100 samples. Sample (20 /.11..) and
`200 )iL of the /1’-carotene solution were added to a well in a
`96-well flat—bottom EIA microtitration plate from ICN Bio-
`medicals Inc. (Aurora, OH). Samples were prepared in tripli-
`cate for each concentration used (0—l500 /1M), and at least
`seven different concentrations were used. To dilute the sample
`mixture, 30 )AL of the mixture was transferred to another plate
`and air—sparged distilled water (210 ;4L) was added [l:8 (v/v)
`dilution]. Because the /5-carotene bleaching reaction was
`subject to noticeable variations, the dilutions were done in
`triplicate. Marco (1968) and Velioglu et al. (1998) initiated the
`reaction by incubating mixtures at 50 “C. In the modified
`method, ADIBA (20 pL of 0.3 M) as used by Pischetsrieder et
`al. (1998) was added to each well to initiate the reaction. The
`plate was read in an MRX plate reader (Dynex Technologies
`Inc, Chantilly. VA) using a 450 nm filter at 0 min and after
`90 min of incubation in the dark at room temperature (~22
`°C). At 0 and 90 min, the /4450mm was usually around 1.0-1.2
`and 0.1—0.3, respectively, for the control (0 ;4M).
`Absorbance at 450 nm after 90 min of incubation was plotted
`against concentration of sample added. Plots either increased
`linearly with concentration and then remained constant or
`showed no change with concentration. The slope for the initial
`linear portion of the plot was calculated from the dilutions done
`in triplicate (I2 > 0.800). The average and standard deviation
`of the slopes from the three replicate measurements were
`calculated and used to compare antioxidant activities.
`DPPH° Method. Modifications were made to the original
`DPPH' method of Brand—Williams et al.
`(1995). For the
`modified procedure, a 150 ;4M solution of DPPI-I‘ was prepared
`in 80% methanol
`instead of 100% methanol. Using 80%
`methanol had the advantage of a faster reaction rate for some
`compounds such as BHA and BHT and lower evaporation
`losses. Instead of reading samples spectrophotometrically, the
`assay was performed in a microplate. To a well in a 96-well
`flat-bottom EIA microtitration plate from ICN Biomedicals Inc.
`were added 22 yL of sample and 200 yL of DPPH' solution.
`Samples were prepared in triplicate for each concentration
`used (0—500 ),¢M), and at least seven different concentrations
`were used. The plate was then covered and left in the dark at
`room temperature (~22 °C). After 30, 180, and 360 min, the
`plate was read in an MRX plate reader using a 520 nm filter.
`The incubation time for caffeic acid was increased to 48 h to
`obtain complete reaction.
`A plot of A520”... versus concentration of sample in the final
`solution was made for each time interval. Using the results
`from the time interval with the steepest slope, the initial slope
`of the curve was calculated by linear regression (I2 > 0.800).
`The antiradical activity was defined by the initial slope value
`in units of A5gu.,,,,/micromolar of sample or micromolar of
`DPPI-I‘/micromolar of sample. The units were converted from
`A52u,.,,, to micromolar of DPP1-I‘ by developing a standard curve
`for DPPH' using the plate reader. The concentration of DPPH'
`was initially determined from the calibration curve equation
`given by Brand-Williams et al. (1995), where A515,“ measured
`spectrophotometrically was equal to 12509 X concentration in
`M - 0.00258. Although Brand-Williams et al. (1995) solubi-
`lized DPPH‘ in methanol, the same equation was used because
`absorbance of DPPI-I‘ in 80% methanol was the same. The
`antiradical activity was found to be equivalent to negative half
`the antiradical power (ARP) as defined by Brand-Williams et
`al. (1995). ARP was equal to the reciprocal of the amount of
`compound required to decrease the initial DPPH' concentration
`by 50% in units of moles of DPPH' per mole of compound.
`HPLC Method. The oxidation procedure of Osawa and
`Shibaim;-to (I992) was modified by initiating oxidation with
`Cu“ instt-ad of Ft.-C12/H202, so both artl ioxitlant and pmoxidant
`activities could be measured using a linoleic acid emulsion.
`For the modified procedure, buffer was prepared with 0.218%
`(w/v) SDS in 21.8 mM Tris—HCl at pH 7.4. Linoleic acid
`emulsion was prepared by mixing linoleic acid with buffer (2.18
`mg of linoleic acid/mL of buffer). Emulsion (1 mL) was
`
`Page 3 of 8
`
`J, Agric. Food Chem., Vol. 48, No. 8, 2000 3599
`
`dispensed into a test tube followed by the addition of 70 /AL of
`sample. Samples were prepared in triplicate for each concen-
`tration used (0~4000 )4M), and at least five different concen-
`trations were used. To initiate the reaction, 20 ;1L of 10.9 mM
`CuSO4 was added to the test tube. The uncovered test tube
`was then incubated at 37 °C for 16 h in the dark on a Lab-
`Line Instruments Inc. shaker (Melrose Park, IL) rotating at
`1500 rpm.
`Test tubes were weighed before and after incubation to
`determine losses due to evaporation. The volume of each test
`tube was adjusted to the initial level with distilled water. To
`each test tube were then added 20 ;¢L of 1.2 mg/mL BHT in
`methanol, 100 ptL of 1.2 N HCl, and 100 p¢L of 120 mM urea.
`Samples were heated at 100 °C for 60 min, cooled, and then
`cleaned by applying 0.5 mL to a tC18 cartridge (Waters
`Chromatography Division, Millipore Corp., Milford, MA).
`Eluate was collected when the cartridge was washed with 1.5
`mL of distilled I-I20. The eluate (50 /,iL) was injected into a
`Waters HPLC system (Mississauga, ON, Canada) equipped
`with a Waters 990 photodiode array detector monitoring at
`309 nm. The column used was a reverse—phase Supelcosil LC-
`18 (25 cm X 2.1 mm i.d.) from Supelco Inc. (Bellefonte, PA).
`Distilled deionized water was used as the solvent at a flow
`rate of 0.35 mL/min, and the column was maintained at 25
`°C. Standard solutions of 2-hydroxypyrimidine or derivatized
`malonaldehyde prepared from TEP using the procedure of
`Csallany et al. (1984) eluted around 4.5 min. The percent
`malonaldehyde of the control at each concentration was
`calculated as area for any concentration per area for 0 /,iM x
`100. Plots of percent malonaldehyde of the control versus
`concentration of sample added were made. If the percent
`malonaldehyde of the control was >100 in the concentration
`range tested, the sample had prooxidant activity. The concen-
`tration range when the percent malonaldehyde of the control
`decreased to 0 was used as a quantitative indicator of anti-
`oxidant activity.
`Extraction of Phenolics from Berries. Phenolics were
`extracted from frozen samples of saskatoon berries (Ame1an-
`chier alnifolia Nutt.), blackberries, blackcurrants, and blue-
`berries. Saskatoon berries (cv. Smoky) were obtained from The
`Berry Basket (Clairmont, AB, Canada), and blackcurrants
`were obtained from Riverbend Country Gardens (Sylvan Lake,
`AB, Canada). Frozen blackberries and blueberries were pur-
`chased at a local supermarket.
`To extract phenolics, 10 g of berries was combined with 40
`mL of 80% methanol in a temperature—controlled (~4 °C)
`Waring blender and mixed for 8 min at low speed. The mixture
`was then filtered (Whatman No. 42) through a Biichner funnel.
`The filtrate volume was adjusted to 50 mL using 80%
`methanol, and a portion of the extract was filtered (0.45 gm
`Acrodisc LC PVDF syringe filter; Pall Gelman Laboratory,
`Montreal, PQ, Canada) prior to phenolic and antioxidant
`analyses.
`Measurement of Phenolics. Phenolics were measured
`using a modified version of the Glories’ method (Glories, 1978;
`Romani et al., 1996) described in Mazza et al. (1999). Briefly,
`the method consisted of mixing 0.25 mL of sample with 0.25
`mL of 0.1% HCl in 95% ethanol and 4.55 mL of 2% HCl. The
`absorbance of the solution was then read at 280, 320, 360, and
`520 nm to measure total phenolics, tartaric esters, flavonols,
`and anthocyanins, respectively. Standards used were chloro-
`genic acid, caffeic acid, quercetin, and malvidin 3—glucoside for
`total phenolics, tartaric esters, flavonols, and anthocyanins,
`respectively. Standards were prepared in 80% methanol except
`for quercetin, which was prepared in 100% methanol.
`
`RESULTS AND DISCUSSION
`
`Prooxidant Activity. Figure 2 shows typical effects
`of the addition of different compounds on percent
`malonaldehyde of the control derived from the HPLC
`method. Almost all phenolics exhibited some prooxidant
`behavior at
`low concentrations (Figure 2; Table 1),
`whereas the synthetic antioxidants (BHA and BHT) and
`
`Page 3 of 8
`
`

`
`3600 J. Agric. Food Chem, Vol. 48, No. 8, 2000
`
`2!) ;\lllh0('_\'2l nitlins
`
`h) .-\I||h0c_Vanins
`
`Fukumolo and Mazza
`
`
`
`
`'+<3-.u...I....\.gI.E.-snug
`H —9—(‘) nnnlun 1 <-i|:glIu\~.<ul:.-
`‘—g— M.Il\‘nlu|»‘11lIIvusnlv
`T ;I..:u..Iu.
`\ ~' ..|.,I....-...|.-
`
`- -9- I);-lgvhuunlm '
`3 —A— 1'‘ am-Inn
`: -0- Pelalgunnluu
`: —— M.|l\u|m
`—9— Pv:uIu<lm
`
`control ......
`%_\[alona|deh\'deofthe
`
`
`I000
`
`l 500
`
`2000
`
`2500
`
`3000
`
` "/n.\1alonaldeh_vdeofthemnu-ol
`
`I I
`
`-
`
`-
`
`n;._
`2500
`
`3000
`
`-
`
`._l_
`I
`3 S00
`
`n
`I
`4000
`
`3000
`
`....:
`3500
`4000
`
`()
`
`'
`0
`
`.L..§-
`500
`
`__;._._.n-
`I 000
`I 500
`
`('oncentrulion ofcompound added (p1M)
`
`(‘oust-ntralion ol'compound adcled (uM)
`
`C) Ha vonols
`
`H0
`
`1|) Slnndartls
`i-
`
`I10
`
`I00 I
`
`80
`
`%— M yncerm
`'-I-— Quelcelul
`' —X— Km:m]I[n.1l|
`-9- Rnlm
`
`
`
`60
`
`-I0
`
`20
`
`0
`
`
`
`
`
`"/a-.\1a|onaldeh.\‘denflhecontrol
`
`
`
`\¢?_——-—<Ite-*“°““--a§_.
`
`"
`
`—D— :1 ‘rm-nphulol
`-fi— BH.-\
`
`‘L’
`
`°/oI\lalonaldeh_\‘deofthecontrol
`
`-
`
`—9—,»\sco|h|<H
`
`
`
` _-...L__..._._...._.n_n....n...
`3 S00
`4000
`0
`500
`1000
`1500
`2000
`2500
`3000
`
`0
`
`500
`
`1000
`
`1500
`
`2000
`
`2500
`
`3000
`
`3500
`
`4000
`
`..a..._.j
`
`(‘onc entralion of compound added (MM)
`
`(‘onrcnlralion ofcompound uddcd (pM)
`
`Figure 2. Antioxidant and prooxidant activities of selected (a) anthocyanidins, (b) anthocyanins, (c) flavonols, and (d) standards
`using the HPLC method. Values are means :I: standard deviations, n = 3.
`
`o.—tocopherol did not. In preliminary trials, some phe-
`nolic compounds added to emulsions exposed to air
`exhibited prooxidant activity even without Cu“ added
`to initiate oxidation. The activity increased with in-
`creasing concentration until the antioxidant activity of
`the compound became dominant. When Cu“ was added,
`the concentration at which antioxidant activity became
`dominant was lower. By changing the initiation proce-
`dure for the HPLC method, the prooxidant and anti-
`oxidant activities of samples could therefore be altered.
`Pischetsrieder et al. (1998) initiated oxidation in their
`flecarotene bleaching assay using CuSO4, MnCl2, or
`FeCl3. Prooxidant activity was calculated as the percent
`of AA47gnm of the sample/A/l47onm of the blank over 20
`min. All three compounds they tested, including ascorbic
`acid, showed prooxidant activity (100—130%) at low
`concentrations. Preliminary trials were performed in
`this study using 20 yL of 0.02 M CuSO4 to initiate
`oxidation using the fl—carotene bleaching method (results
`not shown). Changes in A450nm were monitored after 90
`and 120 min of incubation. Only gallic acid, cyanidin
`3—glucoside, cyanidin 3.5—diglucoside, and ascorbic acid
`had detectable decreases in A45o,,m with concentration,
`indicating prooxidant activity.
`Compared to the HPLC method, prooxidant activity
`was difficult to measure using the fi—carotene bleaching
`method. The fl—carotene bleaching method relied on
`measurement of slight differences in absorbance with
`concentration to detect prooxidant activity. These dif-
`
`ferences were difficult to detect partly due to the high
`variability of the ,8—carotene bleaching reaction and the
`reaction conditions. Although information on the mech-
`anism of prooxidant activity in phenolic compounds is
`limited, the presence and involvement of metal ions and
`oxygen on the prooxidant activity of ascorbic acid has
`been reported (Pischetsrieder et al., 1998). Metal ions
`and oxygen may have been more limited in the fi—car»
`otene bleaching method than in the HPLC method
`because samples were not shaken and a different emul-
`sifier was used (Tween 20 instead of SDS). The reaction
`time was also reduced (2 h instead of 16 h). Therefore,
`only compounds with very high prooxidant activity
`under the conditions used may have been detected by
`this method. Further research in modifications of the
`
`fl—carotene bleaching method could be made to improve
`the method for measuring prooxidant activity.
`Cao et al. (1997) found that myricetin, quercetin, and
`kaempferol had prooxidant behavior using the oxygen
`radical absorbance capacity (ORAC) assay with Cu“ as
`a transition metal oxidant. Arouma et al. (1997) sum-
`marized results found for the prooxidant activity of
`phenolic compounds using deoxyribose and bleomycin—
`DNA assays. Many compounds including ascorbic acid,
`myricetin, quercetin, and gallic acid had prooxidant
`behavior in one or both assays. Vanillic acid tested
`negative. Delphinidin, cyanidin, malvidin, malvin. and
`pelargonidin had some prooxidant activity in human
`low—density lipoprotein (LDL) and lecithin~liposome
`
`Page 4 of 8
`
`Page 4 of 8
`
`

`
`Antioxidant and Prooxidant Activities of Phenolics
`
`J. Agric. Food Chem, Vol. 48, N0. 8, 2000 3601
`
`Table 1. Antioxidant and Prooxidant Activities of Selected Phenolic Compounds
`current study
`/i-carotene method DPPH' method
`
`HPLC method
`
`compd
`
`benzoic acid derivatives
`gallic acid
`protocatechuic acid
`3-hydroxybenzoic acid
`vanillic acid
`syringic acid
`ellagic acid
`cinnamic acid derivatives
`caifeic acid
`pcoumaric acid
`cinnamic acid
`ferulic acid
`chlorogenic acid
`
`myricetin
`quercetin
`rutin
`kaempferol
`
`(+)—catechin
`(-)-epicatechin
`
`naringenin
`
`cyanidin
`cyanidin 3-glucoside
`cyanidin 3,5—diglucoside
`delphinidin
`malvidin
`malvidin 3-glucoside
`malvidin 3,5-diglucoside
`pelargonidin
`pelargonidin 3—glucoside
`pelargonidin 3,5—diglucoside
`peonidin
`peonidin 3—glucoside
`
`ascorbic acid
`oL-tocopherol
`Bl-IA
`BHT
`
`initial slope‘
`(x 10’5)
`
`antioxidant activity” prooxidant
`antiradical
`(MM of compd added)
`activity’
`activity"
`Phenolic Acids
`
`636 :: 38
`nc’
`nc
`nc
`nc
`779 :1: 83
`
`620 :1: 36
`nc
`nc
`161 :: 14
`186 :: 26
`
`1046 :1: 83
`630 d: 42
`nc
`172 :1: 6
`
`443 :t 63
`515 :t 39
`
`nc
`
`836 :1: 69
`278 :1: 32
`220 :1: 39
`897 :1: 147
`288 :l: 34
`448 :1: 40
`266 :1: 27
`nc
`444 :t 94
`nc
`169 :t 22
`251 :: 4
`
`nc
`870 :1: 21
`835 i 50
`864 :i: 76
`
`-6.21 :1: 0.60
`-4.56 :1: 0.09
`nc
`-0.99 :1: 0.31
`-3.03 :1: 0.07
`-9.21 :1: 0.25
`
`1500-2000
`3000-3500
`>4000
`>4000
`>4000
`>4000
`
`-6.59 :1: 0.57
`-6.73 :1: 0.08
`-5.10 : 0.10
`-2.09 ::0.10
`
`-4.49 :1: 0.241
`-0.33 :1: 0.06
`nc
`-1.34 :1: 0.05
`-5.08 :1: 0.29
`
`500-1000
`>4000
`>4000
`>4000
`1000-1500
`Flavonols
`500-1000
`200-300
`500-1000
`>4000
`Flavanols
`500-1000
`-7.19 : 0.32
`500-1000
`-7.65 ::
`.17
`Flavanones
`-0.18 : 0.01
`>4000
`Anthocyanidins/Anthocyanins
`-7.40 :1: 0.15
`200-300
`-6.81 :1: 0.30
`300-400
`-3.32 :1: 0.07
`500-1000
`-8.86 :1: 0.28
`500-1000
`-4.49 :1: 0.28
`1500-2000
`-4.29 :1: 0.42
`500-1000
`-2.56 :1: 0.10
`2000-2500
`-4.63 :1: 0.25
`1500-2000
`-3.95 :1: 0.22
`2000-2500
`-2.04 :1: 0.10
`2000-2500
`-4.05 :1: 0.17
`1500-2000
`-3.38 :1: 0.15
`2500-3000
`Standards
`-1.83 :1: 0.07
`>4000
`-1.95 :1: 0.07
`2000-2500 (50%)
`-2.61 :1: 0.01
`1000-1500
`-3.17 :1: 0.07
`200-300
`
`+
`+
`+
`+
`+
`+
`
`+
`+
`nd"
`--
`+
`
`--
`--
`+
`--
`
`+
`+
`
`--
`
`+
`--
`+
`--
`--
`--
`--
`+
`+
`+
`+
`+
`
`+
`nd
`nd
`nd
`
`selected results from the literature
`
`DPPH' assay"
`ORAC assay”
`antiradical power»-'3 ORAC slope"
`[antiradical activity]
`
`12.5 [-6.25]
`7.14 [-3.57]
`
`0.17 [-0.09]
`
`9.1 [-4.55]
`0.02 [-0.01]
`
`2.33 [-1.17]
`
`4.319 :1: 0.119
`3.285 :1: 0.117
`
`2.671 :1: 0.131
`
`2.239 :t 0.029
`3.491 zt 0.011
`1.689 :1: 0.052
`1.809 :1: 0.068
`2.009 :1: 0.167
`1.404 d: 0.052
`1.550 :1: 0.062
`1.540 :1: 0.033
`1.560 :1: 0.145
`1.067 :1: 0.043
`1.693 :1: 0.035
`1.805 :1: 0.014
`
`3.7 [-1.85]
`
`4.17 [-2.09]
`4.2 [-2.1]
`
`3 The DPPH' assay used by Brand—Williams et al. (1995). “The oxygen radical absorbing capacity (ORAC) assay measuring reaction
`with peroxyl radicals expressed as ;¢M of Trolox equivalent per )4M of compound. Results for flavonols were taken from Cao et al. (1997),
`and results for anthocyanidins/anthocyanins were taken from Wang et al. (1997). ‘Values are means of slope coefficients calculated by
`linear regression :1: standard deviations (n = 3) in A450,“ after 90 min of incubation in the dark/,uM of compound added. ‘(Values are
`means of slope coefficients calculated by linear regression :1: standard deviations (n = 3) in ,uM of DPPH'/yM of compound. ” Antioxidant
`activity was defined by the concentration range of compound added needed to reach 0% malonaldehyde of the control. ’Prooxidant activity
`was positive (+) if the % malonaldehyde of the control was > 100% in the concentration range tested. gAntiradical power was defined as
`the reciprocal of the amount of antioxidant needed to decrease the initial DPPH' concentration by 50%. The antiradical activity was
`equivalent to negative half of the antiradical power. ” Values are slope coefficients calculated by linear regression :1: standard error. ‘Not
`calculated since linear regression [2 < 0.800. JValues were obtained after reaction for 48 h. "Not detected.
`
`systems using a Cu“ catalyst (Satué—Garcia et al.,
`1997). The results for prooxidant activity from these
`studies were similar to those obtained in this study.
`Some differences were expected, though, because the
`conditions in a particular assay would affect the reac-
`tions occurring. The potential prooxidant properties of
`phenolic compounds suggest that care should be taken
`when using these compounds as antioxidants.
`Antioxidant Activity. Antioxidant activity by the
`HPLC method was characterized by a decrease in
`percent malonaldehyde of the control toward 0 (Figure
`2). The antioxidant activity was quantified by the
`concentration range when the percent malonaldehyde
`decreased to 0 (Table 1). Stronger activity was indicated
`
`by a lower concentration range required. For most
`compounds, the percent malonaldehyde dropped rapidly
`within a certain concentration range. However,
`for
`myricetin, delphinidin, BHA, BHT, and o.—tocopherol,
`the decrease in percent malonaldehyde with concentra-
`tion was slow (Figure 2). The concentration range for
`these compounds, therefore, may not accurately reflect
`their potential antioxidant activity.
`Table 1 also summarizes the antioxidant activity of
`the compounds tested using the fi—carotene bleaching
`and DPPH' methods. Higher initial slope values for the
`fl—carotene bleaching method indicated less bleaching
`with increasing concentrations of compound added and,
`therefore, higher potential antioxidant activity. Higher
`
`Page 5 of 8
`
`Page 5 of 8
`
`

`
`3602 _J. Agric. Food Chem., Vol. 48, No. 8, 2000
`
`absolute values for antiradical activity by the DPPH'
`method indicated a higher concentration of DPPH'
`needed to react with each micromolar of compound and,
`therefore, higher potential antioxidant activity. Caffeic
`acid reacted very slowly with DPPH' compared to the
`other compounds, and so a longer incubation time (48
`h) was used. The DPPH' method gave values for
`antiradical activity similar to or higher than those
`reported by Brand—Williams et al. (1995) (Table 1) for
`the same compounds. Higher values were obtained
`probably due to the faster reaction rate of DPPH' in 80%
`methanol than in 100% methanol, especially for slowly
`reacting compounds such as BHA and BHT.
`The methods developed in this study to measure anti~
`oxidant activity were expressed relative to the concen-
`tration of compound added or in final solution. Often
`in antioxidant activity assays, substrates or products
`are monitored over time, and the effectiveness of an
`antioxidant at a specific concentration is determined by
`the time required to reach a certain level of substrate
`or product. Monitoring oxidation with time is time—con—
`suming and not practical when large numbers of samples
`are involved. In this study, methods were therefore
`developed using fixed reaction times with varying
`concentrations to save time while allowing comparisons
`in activity to be made based on concentration.
`Generally, for groups of compounds with similar
`structures (Figure 1) all methods followed similar
`trends. For the benzoic and cinnamic acid derivatives,
`flavonols, and anthocyanidins, an increase in the num-
`ber of hydroxyl groups led to higher antioxidant activity.
`Compounds with three hydroxyl groups on the phenyl
`ring of phenolic acids or the B ring of flavonoids had
`high antioxidant activity. The loss of one hydroxyl group
`decreased activity slightly, whereas the loss of two
`hydroxyl groups significantly decreased activity. Using
`the HPLC method, this trend was not as obvious for the
`flavonols and anthocyanidins due to the slower decrease
`in percent malonaldehyde of the control with concentra-
`tion of myricetin and delphinidin added. Dziedzic and
`Hudson (1983) found that at least two hydroxyl groups
`were required for antioxidant activity of phenolic acids.
`Pratt and Hudson (1990) noted the position and degree
`of hydroxylation of flavonoids, especially of the B ring,
`play a major
`role in antioxidant activity with all
`flavonoids, with the 3’,4’—dihydroxy configuration having
`antioxidant activity.
`The addition of methoxyl groups to phenolic acids
`increased antioxidant activity with the fi—carotene bleach-
`ing and DPPH' methods, but changes were not observed
`using the HPLC method. The ,6—carotene bleaching and
`DPPH' methods were more sensitive to small changes
`in concentration, so differences were easier to detect.
`Increases in activity were also seen upon the addition
`of methoxyl groups to anthocyanidins by the ,6—carotene
`bleaching method. Dziedzic and Hudson (1983) found
`that steric hindrance of phenolic hydroxyl groups such
`as by the addition of methoxyl groups could enhance
`activity.
`Glycosylation resulted in lower antioxidant activity
`for quercetin, cyanidin, pelargonidin, and peonidin using
`the DPPH' and HPLC methods and for quercetin and
`cyanidin using the fl—carotene bleaching method. Addi-
`tion of a sugar moiety decreased activity of the aglycon,
`and the addition of a second moiety decreased activity
`further, probably due to steric hindrance by addition of
`sugar moieties. Tsuda et al. (1994) found that cyanidin
`
`Page 6 of 8
`
`Fukumolo and Mazza
`
`had greater antioxidant activity than cyanidin 3—gluco-
`side in linoleic acid, liposome, rabbit erythrocyte mem-
`brane ghost, and rat liver microsomal systems. Pratt
`and Hudson (1990) noted that 3—glycosides of ilavonoids
`can possess the same or sometimes less activity th