Mutation Research 585 (2005) 1–13

Cytotoxicity and gene induction by some essential oils in theyeastSaccharomyces cerevisiae

F. Bakkalia,b, S. Averbeckb, D. Averbeckb,∗, A. Zhiri c, M. Idaomara

a Universite Abdelmalek Essaadi, BCM, D´epartement de Biologie, BP 2121 T´etouan, Moroccob Institut Curie-Section de Recherche, UMR2027 CNRS/I.C., Bˆat. 110, Centre Universitaire d’Orsay, F-91405 Orsay Cedex, France

c S.A. PRANAROM International, Ghislenghien, Belgium

Received 23 August 2004; received in revised form 25 January 2005; accepted 14 March 2005Available online 21 June 2005

Abstract

In order to get an insight into the possible genotoxicity of essential oils (EOs) used in traditional pharmacological applicationswe tested five different oils extracted from the medicinal plantsOriganum compactum, Coriandrum sativum, Artemisia herbaalba, Cinnamomum camphora and Helichrysum italicum for genotoxic effectsusing the yeastSaccharomyces cerevisiae. Clear cytotoxic effects were observed in the diploid yeast strain D7, with the cellsbeing more sensitive to EOs in exponential than in stationary growth phase. The cytotoxicity decreased in the following order:Origanum compactum>Coriandrum sativum>Artemisia herba alba>Cinnamomum camphora>Helichrysum italicum. In the

grgenic

EOsthan thatiptionalinked tovolves

same order, all EOs, except that derived fromHelichrysum italicum, clearly induced cytoplasmic petite mutations indicatindamage to mitochondrial DNA. However, no nuclear genetic events such as point mutations or mitotic intragenic or interecombination were induced.

The capacity of EOs to induce nuclear DNA damage-responsive genes was tested using suitableLac-Z fusion strains forRNR3andRAD51, which are genes involved in DNA metabolism and DNA repair, respectively. At equitoxic doses, alldemonstrated significant gene induction, approximately the same as that caused by hydrogen peroxide, but much lowercaused by methyl methanesulfonate (MMS). EOs affect mitochondrial structure and function and can stimulate the transcrexpression of DNA damage-responsive genes. The induction of mitochondrial damage by EOs appears to be closely loverall cellular cytotoxicity and appears to mask the occurrence of nuclear genetic events. EO-induced cytotoxicity in

se or the

oxidative stress, as is evident from the protection observed in the presence of ROS inhibitors such as glutathione, catalairon-chelating agent deferoxamine.© 2005 Elsevier B.V. All rights reserved.

Keywords:Essential oils;Origanum compactum; Coriandrum sativum; Artemisia herba alba; Cinnamomum camphora; Helichrysum italicum;Saccharomyces cerevisiae; Cytotoxicity; Gene expression; Cytoplasmic petite mutations;RNR3; RAD51; Anti-radical effects

∗ Corresponding author. Tel.: +33 169867188; fax: +33 169869429.E-mail addresses:fadil.bakkali@curie.u-psud.fr (F. Bakkali), dietrich.averbeck@curie.u-psud.fr (D. Averbeck), abzhiri@hotmail.com

(A. Zhiri), idaomar@hotmail.com (M. Idaomar).

1383-5718/$ – see front matter © 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.mrgentox.2005.03.013

2 F. Bakkali et al. / Mutation Research 585 (2005) 1–13

1. Introduction

Essential oils (EOs) extracted from medicinal plantsare widely used in aromatherapy and in traditionalmedicine, as well as in the food and perfume industryand in agriculture. One can find them applied in AIDSand cancer therapy, as condiments and as conservingor flavoring agents in food and drinks, as insecticides(aerosols) in agriculture, as perfumes in cosmetics, asdisinfectants in toothpaste and in household prepa-rations. In this respect, their antiviral, antibacterial,antifungal, antioxidant, cytotoxic and anticarcinogenicproperties are of special importance. In addition, theycan also influence the central nervous system affect-ing humoral and cognitive functions and can exertantispasmodic effects. EOs from five medicinal plantsare the focus of this paper:Origanum compactumknown for its antibacterial, antifungic, antimalaria,anti-infectious and antihypotensive properties[1,2];Coriandrumsativumknown for its antiradical, antibac-terial, antispasmodic and expectorant properties[3];Artemisia herba albaknown for its antibacterial, anti-helminthic, antidiabetic and antihypotensive properties[2]; Cinnamomum camphoraknown for its anti-inflammatory and antiparasitic prop-erties [4–6]; Helichrysum italicumknown for its antimutagenic effects [5,7]. EOscan be classified as complex mixtures generally con-taining more than 20 different compounds belonging todifferent chemical classes. Indeed, the composition ofE statee ate,s n.

uche om-p bleg heya es ofo utedf d ass ed ass h ast y tob iento

thec veE the

yeastSacharomyces cerevisiae, which is frequentlyemployed in genotoxic risk assessment including stud-ies at the mitochondrial and nuclear DNA level[8–11].We investigate here the cytotoxic and genotoxic effectsin diploid yeast (D7) and the induction of two DNAdamage-responsive genes, the ribonucleotide reductasegeneRNR3involved in the synthesis of DNA precur-sors[12,13]and the DNA repair geneRAD51involvedin the repair of DNA double-strand breaks by homolo-gous recombination[14–16] in haploidLac-Z fusionstrains after treatments with the five EOs extractedfrom Origanum compactum, Coriandrum sativum,Artemisia herba alba, Cinnamomum camphora and Helichrysum italicum.The RNR3 gene is known to beinduced in the presence of different types of DNAdamage including oxidative damage[12,13]. Also theRAD51gene is a DNA damage-inducible gene[14–16].

Using inhibitors of oxygen radical-mediated reac-tions such as reduced glutathione (GSH)[17–20], cata-lase[20] and deferoxamine (DFO)[17] we demonstratethat the cytotoxic action of EOs fromOriganum com-pactum, Coriandrum sativum, Artemisia herba albaandCinnamomum camphorainvolves the productionof reactive oxygen species (ROS) and/or hydrogen per-oxide.

2. Materials and methods

2 s

RA-Nc ),C ),A rh s d).A se parto l, int n wasq natet eree n-s s asd

Os depends on the plant species, developmentalnvironmental growth conditions (season, climoil) and on the part of the plant used for extractio

The genotoxicity of EOs has not yet been mxplored. Because of their complex chemical cosition, the evaluation of their toxicity and possienotoxic effects is indeed a difficult matter. Since tre usually employed as extracted complete entitiily consistency and not as compounds reconstit

rom pure ingredients, EOs have to be considereubstances acting as a whole and should be testuch. Many known genotoxicity test systems, suchose with bacteria and mammalian cells, are likele affected by oily conditions because of insufficxygen supply on the tester plates.

The present study was undertaken to estimateytotoxic and possible genotoxic activities of fiOs in a typical unicellular eukaryotic system,

,.1. Essential oils (EOs) and reference compound

We used five essential oils chemotyped by PAROM (B-7822 Ghislenghien, Belgium):Origanumompactumof Moroco (extracted from flower headoriandrum sativumof Russia (extracted from fruitsrtemisia herba albaof Moroco (extracted from floweead), Cinnamomum camphora

of Madagascar (extracted from leaves), Helichry-um italicumof France (extracted from flower heaccording to the chart of PRANAROM, each EO waxtracted under the same conditions from the samef the plants, which were grown on the same soi

he same climate and season so that its compositiouite reproducible. Throughout the paper we desig

he EOs by the name of the plant from which they wxtracted.Table 1gives the list of main chemical cotituents of the EOs derived from the plant extractetermined by gas chromatographic analysis[6,7].

F. Bakkali et al. / Mutation Research 585 (2005) 1–13 3

Table 1Main ingredients of five essential oils (EOs) fromOriganum compactum,Coriandrum sativum, Artemisia herba alba,Cinnamomum camphoraand Helichrysum italicum

Ingredients Essential oils

Origanumcompactum(%)

Coriandrumsativum(%)

Artemisia herbaalba (%)

Cinnamomumcamphora(%)

Helichrysumitalicum (%)

1,8-Cineole 3.26 50.20Alpha curcumene 3.21Alpha pinene 5.31 5.19Alpha terpinene 2.59Alpha terpineol 9.89Alpha thuyone 36.67Beta caryophyllene 2.85Beta pinene 2.70Beta thuyone 20.61Camphene 3.01Camphor 5.93 24.61Carvacrol 30.53Epi-gamma-eudesmol 3.25Gamma terpinene 18.2 3.41Geranyl acetate 4.43Isoitalicene 2.98Italicene 8.28Italidione I, II, III 10.46Limonene 2.29 3.84 4.75Linalol 68.39Nerol 2.61Neryl acetate 33.90Neryl propionate 6.47p-Cymene 7.89 2.27Sabinene 6.45Terpinene-4-ol 2.65Thymol 27.5Viridiflorol 2.75Not determined 10.44 7.97 11.84 16.33 24.09

In the study of gene induction, we included hydro-gen peroxide (H2O2 solution, 30%, w/w) (CAS 7722-84-1, Sigma–Aldrich, USA) and methyl methanesul-fonate (MMS) (CAS 66-27-3, Sigma–Aldrich, USA)as DNA-damaging agents[21,22].

Antiradical effects were studied using inhibitorsof oxygen radical-mediated reactions such as defer-oxamine mesylate salt (DFO) (CAS 138-14-7,Sigma–Aldrich, USA) or reducedl-glutathione (GSH)(CAS 70-18-8, Sigma–Aldrich, USA) and catalase(product ref. C100, Sigma–Aldrich, USA).

2.2. Yeast strains

For gene expression analysis, we used the followinghaploid strains of the yeastSaccharomyces cerevisiae:

DBY747 (MAT a RAD+ his3-11, leu2-3, 112 trpl-289, ura3-52, carrying theRNR3–lacZfusion) kindlyprovided by Dr. Wei Xiao (Department of Microbi-ology and Immunology, University of Saskatchewan,Saskatoon, Canada, SK S7N 5E5)[12].

FF1082 (MATa RAD+ leu2, trpl, ura3, his7,lysl, ura3+-51LACZ LEU2, carrying theRAD51–lacZfusion) kindly provided by Dr. Francis Fabre (DSV,CEA, Fontenay aux Roses, France).

For the detection of cytotoxic effects, the induc-tion of cytoplasmic petite mutations, ILV+ point muta-tions, TRP+ mitotic gene convertants and mitoticcrossing-over involving theAde2 gene, we usedthe known diploid yeast tester strain D7 (a/�,Ade2-40/Ade2-119,Trp5-12/Trp5-27, Ilv1-92/Ilv1-92)[9].

4 F. Bakkali et al. / Mutation Research 585 (2005) 1–13

2.3. Growth media

The complete growth medium (YEPD) contained2%d-glucose (Sigma–Aldrich), 1% bacto yeast extract(Difco), 2% bacto peptone (Difco), 2.6% bacto agar(Difco).

The minimum medium complemented or not withthe appropriate amino acids was composed of 2%d-glucose, 3.2% bacto agar complemented with 0.67%of bacto yeast nitrogen base without aminoacids and5 mg adenine/L medium, 60 mg isoleucine/L for thedetection ofTRP+ gene convertants (intragenic recom-binants), 10 mg tryptophan/L for the detection ofILV+

revertants.

2.4. Treatments with essential oils (EOs) or withH2O2 or MMS as reference compounds

Cells of the strain D7 were grown in YEPD at30◦C up to exponential (2× 107 cells/mL) or station-ary phase of growth (2× 108 cells/mL). Cells in expo-nential phase or stationary phase were sedimented andresuspended in distilled water. Generally, cell suspen-sions of 2× 107 cells/mL were treated with EOs orthe reference compounds at the desired concentrationsfor a given time as indicated in the figure legends.First, EOs were carefully diluted in absolute ethanol(Merck) at different concentrations, and identical vol-umes were added to the aqueous cell suspensions insmall flasks to give the chosen concentrations of EOsa %).P ss-c con-t ingt con-t eed( t tok pen-s tsw ands s orr ost-t

2

s tod nce

compounds H2O2 or MMS at a given incubation timeor to a given concentration of EOs at different incu-bation times, at room temperature, were determinedas previously described[8,23,24]. Clonogenic survivalwas determined, and among the surviving coloniescytoplasmic petite mutants were detected either byreplication on glycerol medium or by the triphenyltetrazolium chloride overlay technique[25]. Mitoticintergenic recombinants involving theade2locus weredetected by accumulation of a red pigment (imidazole)in purine metabolism on complete growth medium.Mitotic gene convertants (TRP+) and gene revertants(ILV +) were detected by their growth on the appro-priate selective media. Experiments were performed atleast in triplicate with stationary phase cells and withexponential phase cells.

2.6. Detection of reactive oxygen species usinginhibitors

The involvement of radical-mediated reactions inthe cytotoxic effects of four EOs (Origanum com-pactum, Coriandrum sativum, Artemisia herba alba,Cinnamomum camphora) was determined using l mMDFO or 1 mM GSH during 1 h pre-incubation of cellsbefore treating with EOs for 30 min. One thousand unitsper millilitre of catalase were added at the same time asthe EOs. For the control experiments with EOs aloneand the experiments with catalase, cells were also pre-incubated for 1 h to obtain conditions comparable tot eastc e toOi aa Oss ryg t leastt

2R

ingtRg tratea den-s

t a given same final concentration of ethanol (1.25ipetman filter tips (ART) were used to avoid croontamination between different EO preparationsaining some volatile molecules. The flasks containhe cell suspensions with EOs were agitated atrolled room temperature and at constant low sp200 rpm in a New Brunswick agitator) just sufficieneep the oil/ethanol mixture in the aqueous cell susion well mixed. After 30 min incubation, cell aliquoere diluted and spread onto YEPD solid mediumelective minimal media to select for recombinantevertants. Colonies were counted after 6 days of preatment incubation at 30◦C.

.5. Detection of cytotoxic and genotoxic effects

Cytotoxic and genotoxic effects after exposureifferent concentrations of EOs or of the refere

hose used for the experiments with DFO or GSH. Yells (D7) were exposed in stationary growth phasriganum compactumandCoriandrum sativumand

n exponential growth phase toArtemisia herba albndCinnamomum camphorabecause these latter Ehowed too little cytotoxicity to cells in the stationarowth phase. Results presented are based on a

hree independent experiments.

.7. Test of gene expression in RNR3–lacZ andAD51–lacZ fusion strains

Gene induction was analysed by determinhe �-galactosidase activity in theRNR3–lacZandAD51–lacZfusion strains usingo-nitro-phenyl-�-d-alactopyranoside (ONPG) as a colorimetric subsnd measuring spectrophotometrically the opticality at 420 nm[26,27].

F. Bakkali et al. / Mutation Research 585 (2005) 1–13 5

Briefly, yeast cells were grown in YEPD up to expo-nential phase (1.1–1.3× 107cells/mL), sedimented andresuspended in distilled water. Aliquots were exposedto the EOs or to the reference compounds at differentconcentrations under the same conditions as above, for30 min at room temperature. Cells were centrifuged,resuspended in YEPD and then incubated for the indi-cated times at 30◦C. After this post-treatment incu-bation, cells were collected by centrifugation, resus-pended in a buffer with mercaptoethanol and per-meabilized by chloroform and SDS. The activity of�-galactosidase was determined by means of ONPG[12,28]and expressed as previously described[26,27].The results reported are from at least three independentexperiments.

3. Results

3.1. Cytotoxic effects of EOs

After treatment of diploid yeast cells (D7) with EOsextracted fromOriganum, Coriandrum, Artemisia,Cinnamomum, Helichrysum, we observed an inhibi-tion of cellular growth depending on the concentrationof the EOs. To quantify the cytotoxic effects involved,we measured the clonogenic survival after treatmentof the cells with the five EOs.Fig. 1shows the resultsobtained for cells treated in exponential and stationary

growth phase at different EO concentrations for 30 min.In most cases, shouldered survival curves are obtained.Stationary phase cells are clearly more sensitive toOriganum (LD50 = 0.45�L/mL) and Coriandrum(LD50 = 1.6�L/mL) than toArtemisia, Cinnamomumand Helichrysum (LD50 > 8�L/mL). The cytotoxiceffectiveness decreases in the following order:Ori-ganum>Coriandrum�Cinnamomum>Artemisia>Helichrysum.

A different pattern is seen for cells treatedin the exponential phase of growth. Cells areclearly more sensitive, especially toArtemisia andCinnamonum so that the cytotoxic effectivenessdecreases in a slightly different order:Origanum>Coriandrum>Artemisia>Cinnamomum�Helichry-sumthan in stationary phase.

The results suggest that the cytotoxic effects of EOsare facilitated by the less thicker cell wall at buddingsites of exponential phase cells and may be mediatedby effects on the cell membranes.

3.2. Induction of cytoplasmic petite mutations

Yeast has the advantage that mitochondrial dam-age can be easily assessed by detection of cytoplas-mic petite mutations among the surviving cells.Fig. 2shows the results obtained as a function of the EO con-centration. The induction of cytoplasmic petite muta-tions follows the same order as that for cytotoxicity:

F the s tment withfi sativumc italicum withs

ig. 1. Survival curves obtained for diploid yeast cells (D7) inve EOs fromOriganum compactum(ORI) (©, �); Coriandrumamphora(CIN) (Ravintsara aromatica) (�, �); andHelichrysumtandard deviations.

tationary phase (a) and the exponential phase (b), after trea(COR) (�, �); Artemisia herba alba(ART) (♦, �); Cinnamomum(HEL) (�, �) at different concentrations. Data are presented

6 F. Bakkali et al. / Mutation Research 585 (2005) 1–13

Fig. 2. Induction of cytoplasmic petites in diploid yeast cells (D7)in stationary (broken lines) and exponential phase (solid lines) aftertreatment with five EOs fromOriganum compactum(ORI) (©, �);Coriandrum sativum(COR) (�,�);Artemisia herba alba(ART) (♦,�); Cinnamomum camphora (CIN) (�, �);andHelichrysum italicum(HEL) (�, �) at different concentrations.Data are presented with standard deviations.

Origanumbeing the most active andHelichrysumbeingnearly inactive. Also, cells treated in the exponentialphase are clearly more sensitive to the induction ofcytoplasmic petites than cells treated in the station-ary phase of growth. The differential effect of the twogrowth phases is slightly less pronounced in the caseof Coriandrum.

Also, when compared at equitoxic levels clearlymore cytoplasmic petite mutations are induced by EOssuch asArtemisia, CinnamomumandCoriandruminexponential phase cells than in stationary phase cells(data not shown). Most cytoplasmic petites form non-sectored colonies suggesting that treatments with EOsdamage already the mitochondria of mother cells, andthe damage is directly transmitted to daughter cells.

3.3. Induction of point mutations and mitoticrecombination

Using the diploid yeast strain D7, several importantgenotoxic endpoints such as induction of point muta-tions and mitotic intragenic recombination (gene con-version) and intergenic recombination (crossing-over)can be detected simultaneously. Exposure to the fiveEOs did not provide evidence for any induction of thistype of genetic events in cells in the stationary or expo-

nential phase of growth (data not shown). These typesof events were also not induced when cells were treatedat given concentrations for different incubation times,at least for the three EOs tested (Origanum,Artemisia,Cinnamomum) (data not shown). Genotoxicity (induc-tion of mutations and mitotic recombinants) was alsonot observed in cells that grew in liquid medium inthe presence of EOs fromOriganum, Coriandrum,Artemisiaand were harvested at the end of growth (datanot shown). Also, there was no induction of CAN®

mutants after exposures of the haploid wild type strainBY4741 to the EO ofOriganum.

Taking MMS and H2O2 as reference compoundsand in line with previous work[26–28], MMS effec-tively induced point mutations and mitotic intra- andintergenic recombination, but only a few cytoplas-mic petites. H2O2 induced strong lethality, just a fewnuclear genetic events and no cytoplasmic petite muta-tions.

Thus, in contrast to MMS and H2O2, the five EOsare unable to induce significant nuclear genetic eventsin yeast. Induction of cytotoxicity appears to be pre-dominant.

3.4. Modification of the expression of DNAdamage-responsive genes after treatments withEOs

Changes in gene expression can be taken as a verysensitive measure of genotoxic insult. We tested thee ge-i eL e-v ring�

w ase,a ofpF s ofif OsfCc iono tert ursa ent

ffects of EOs on the induction of two typical damanducible genesRNR3andRAD51using appropriatacZ fusion strains of the yeastSaccharomyces cerisiae. Gene induction was monitored by measu-galactosidase activity.

We first studied the induction of the geneRNR3,hich encodes a subunit of ribonucleotide reductn important enzyme involved in the synthesisrecursors for DNA replication and repair[13].ig. 3a–e shows the results obtained for the kinetic

nduction of theRNR3gene in the haploidRNR3–LacZusion strain using different concentrations of Erom Origanum (a), Coriandrum (b), Artemisia (c),innamomum(d) andHelichrysum(e) giving rise toomparable levels of cytotoxicity. With the exceptf Helichrysum, gene induction is starting 4 h af

reatment, and maximum gene induction occpproximately at the same time of post-treatm

F. Bakkali et al. / Mutation Research 585 (2005) 1–13 7

Fig. 3. Induction of theRNR3gene in the haploidRNR3–LacZfusion yeast strain by five EOs. (a)Origanumcompactumat different concentrationsgiving rise to (©) 61.9± 2.7%; (�) 23.2± 3%; (♦) 2.3± 0.3% survival. (b)Coriandrum sativumat different concentrations giving rise to (©)75.1± 3.5%; (�) 54.5± 1.5%; (♦) 19.2± 8.2% survival. (c)Artemisia herba albaat different concentrations giving rise to (©) 79.1± 2.6%; (�)47.3± 11.3%; (♦) 11.3± 4% survival. (d)Cinnamomum camphoraat different concentrations giving rise to (©) 72.6± 3.6%; (�) 40± 9.5%;(♦) 13.9± 3.6% survival. (e)Helichrysum italicumat different concentrations giving rise to (©) 65.6± 5.5%; (�) 39.5± 5.7%; (♦) 17.1± 2.4%survival. Data are presented with standard deviations.

incubation, i.e. at 10–12 h after treatment in ourexperimental conditions. In the case ofHelichrysum,the start of induction is clearly delayed up to 10 hdepending on survival levels. For all EOs, geneinduction is clearly concentration-dependent. Atcomparable cytotoxic levels (Fig. 4a), the inducingcapacity is greatest forCinnamomumand lowestfor Helichrysum and follows the orderCinnamo-mum>Origanum>Coriandrum>Artemisia>Helich-rysum.

For comparison, we also included in this study twoknown DNA-damaging agents, the methylating agentMMS and the oxidizing agent H2O2. Gene inductionwas found to be concentration-dependent with a shift ofmaximum induction from 5 to 10 h post-treatment incu-bation for survival levels from approximately 70 down

to 6% (data not shown). At comparable cytotoxic lev-els, gene induction occurs for both agents a few hoursearlier than for EOs (Fig. 4b). H2O2 reaches inductionlevels close to that of the most effective EOs (Cinnamo-mumandCoriandrum). MMS, a DNA double-strandbreak-inducing agent, is strikingly more effective (twoto three-fold). The comparable responses for some EOsand H2O2 suggest that in both cases oxidative radical-mediated DNA lesions rather than DNA double-strandbreaks are initiatingRNR3gene induction.

In order to check whether the geneRAD51, aDNA damage-inducible gene involved in double-strandbreak repair and homologous recombination wouldalso respond to treatments by EOs, we tested its induc-tion in aLacZfusion strain after treatments withOrig-anumandCinnamomum(Fig. 5a and b). In contrast to

8 F. Bakkali et al. / Mutation Research 585 (2005) 1–13

Fig. 4. Induction of theRNR3gene in the haploidRNR3–LacZfusion strain by the five EOs fromOriganum compactum(ORI), Coriandrumsativum (COR), Artemisia herba alba (ART),Cinnamomum camphora (CIN) and Helichrysum italicum (HEL) at quasi-equitoxic concentrations: ORI (23.2% survival), COR (19.2% survival), ART (11.3% survival), CIN (13.9% survival), HEL (17.1% survival)(a); and by MMS (21.4% survival) and H2O2 (14.9% survival) together with the five EOs for better comparison (b). Data are presented withoutstandard deviations for better clarity.

the results obtained with theRNR3–LacZfusion strain,after treatment with the two EOs the peaks of induc-tion forRAD51are shifted from 8 to 12 h dependent onthe survival level (from 64.8 down to 6.2%). The max-ima of induction are clearly much lower forRAD51(Fig. 5a and b) than forRNR3(Fig. 3a and d) but stillsignificant for the two EOs used. At comparable sur-vival levels (Fig. 5c)RAD51induction by the two EOsis approximately seven-fold lower than that caused byMMS and H2O2.

As forRNR3(Fig. 4b),RAD51gene induction peaksoccur approximately at the same time (8–10 h aftertreatment) for the two EOs tested, whereas the peaksof induction are reached 2–3 h earlier after treatmentswith MMS and H2O2 (Fig. 5c). In contrast to the resultsobtained on the induction ofRNR3(Fig. 4b), the max-imal responses forRAD51are comparable for MMSand H2O2 (Fig. 5c) suggesting that induction of thesetwo DNA damage-inducible genes involves differentsignals for gene induction.

3.5. The effects of ROS inhibitors such asdeferoxamine (DFO), reduced glutathione GSHand catalase

In order to get further insight into the possiblemechanisms involved in the cytotoxic and cytoplas-

mic effects of EOs we checked the possible contribu-tion of radical-mediated reactions. For this, we usedthree inhibiting agents: (1) the iron-chelating agentDFO, an efficient inhibitor of Fenton reactions andproduction of hydroxyl radicals; (2) GSH, an efficientscavenger of ROS such as O2

•−; and (3) catalase, ahydrogen peroxide deactivating enzyme. The activityof these inhibitors is partially overlapping. As shownin Fig. 6, the inhibitors protect in a rather specific man-ner against the cytotoxic effects of EOs (Origanum,Coriandrum, ArtemisiaandCinnamomum). The EOfromHelichrysumwas not included because of its rel-atively weak cytotoxicity. GSH and, to a lesser extent,DFO protected against treatment with the EO fromOriganum, whereas catalase had no effect. DFO andcatalase protected efficiently against treatment with theEO fromCoriandrum, whereas GSH increased cyto-toxicity. DFO protected against treatment with the EOfrom Artemisia, whereas GSH and catalase were inef-fective. Catalase efficiently protected against treatmentwith the EO fromCinnamomum, and DFO was lesseffective, whereas GSH increased cytotoxicity. Theresults show that part of the oxygen radical-mediatedreactions produced by EO treatments can be inhib-ited. Thus, depending on the EO used, reactive oxygenspecies produced appear to contribute to the observedcytotoxicity.

F. Bakkali et al. / Mutation Research 585 (2005) 1–13 9

Fig. 5. Induction of theRAD51gene in the haploidRAD51–LacZfusion yeast strain by two EOs. (a)Origanum compactumat differentconcentrations giving rise to (©) 64.8 ± 7.5%; (�) 26.3 ± 8.8%; (♦) 5.5 ± 2.3% survival. (b)Cinnamomum camphoraat different concentrations giving rise to (©) 61.0 ± 3.4%; (�) 18.9 ± 5.7%; (♦) 1.16 ± 0.44% survival. (c) Induction of theRAD51 gene byOriganum compactum (ORI), Cinnamomum camphora (CIN) , MMS and H 2O2 at quasi-equitoxic concentrations: ORI(26.3% survival) (©), CIN (18.9% survival) (�), MMS (24.2% survival) (�) and H2O2 (21.3% survival) (�). Data are presented with standarddeviations in (a) and (b) and without in (c) for better clarity.

4. Discussion

The five EOs used in this study are complex mix-tures extracted fromOriganum compactum, Corian-drum sativum, Artemisia herba alba, CinnamomumcamphoraandHelichrysum italicum, which are partof traditional human pharmacopeia. EOs are usuallyemployed as whole entities. In the present study, weattempted to get further insight into their possible cyto-toxic and genotoxic effects in eukaryotic cells takingthe yeastSaccharomyces cerevisiaeas a model system.

All five EOs induced cytotoxic effects that werestronger in exponential than in stationary phase cells.The more fragile cell wall structure of budding cellsin the exponential growth phase appears to facilitatepenetration of EOs. Treatment of yeast cells in theexponential phase with the four most active EOs (Ori-ganum compactum, Coriandrum sativum, Artemisiaherba alba, Cinnamomum camphora) gave rise toshouldered survival curves. The steep final slopes of thecurves indicate strong cytotoxicity, possibly involvingdamage to cellular membranes.

10 F. Bakkali et al. / Mutation Research 585 (2005) 1–13

Fig. 6. Modifications of cytotoxic effects of four EOs: (�) (a) Origanum compactum(ORI); (b) Coriandrum sativum(COR); (c) Artemisiaherba alba (ART); and (d) Cinnamomum camphora (CIN) in the presence of 1 mM reduced glutathione (GSH) ( ♦);l mM deferoxamine (DFO) (�); or 1000 units/mL catalase (�). Data are presented with standard deviations.

Treatment with EOs damages cell walls and mito-chondria, which could also be observed by microscopicanalysis (data not shown).

The cytotoxicity of the five EOs was accompa-nied by the induction of cytoplasmic petite mutations,indicating mitochondrial damage and impairment ofoxidative metabolism[11]. Origanumwas the mostactive EO,Helichrysumwas the least active. Morecytoplasmic petites were induced in exponential thanin stationary phase cells, probably also due to easierpenetration of EOs in exponential cells. The absence

of sectored colonies[11] suggested that EOs producedysfunction of mitochondria in the treated mother cells.Damage to yeast mitochondria may, at least in part, belinked to programmed cell death or apoptosis[29–31].We found yeast cells lacking mitochondria to be moresensitive to EOs than cells with mitochondria (data notshown).

Absence of induction of genotoxic damage innuclear DNA, i.e. absence of the induction of pointmutations, mitotic gene conversion and mitotic inter-genic recombination (crossing-over) in the diploid

F. Bakkali et al. / Mutation Research 585 (2005) 1–13 11

tester strain D7 suggested that either the EOs do notreach nuclear DNA and cause cell death by solely dam-aging cellular membranes, or that they cause damageto cellular organelles such as mitochondria, involvingoxidative stress that may lead to cytotoxicity.

In an attempt to better understand the cellularresponses after treatment with EOs we also studiedgene induction. In fact, not only bacteria but alsoeukaryotic cells such as yeast and mammalian cellsrespond sensitively to external stress from exposures tovarious genotoxic agents by the transcription of a largenumber of genes[32,33]. Many genes are inducibleby agents that induce damage to nuclear DNA. UsingRNR3-and RAD51–LacZfusion strains ofSaccha-romyces cerevisiaewe confirm here that the genesRNR3[12,13]andRAD51[14–16]can be induced byDNA-damaging agents such as MMS and H2O2 pro-ducing DNA double-strand breaks and oxidative dam-age, respectively. Furthermore, we show that EOs arealso able to induce these two genes, which play impor-tant roles in DNA metabolism and repair. Interestingly,at comparable cytotoxic levels,RNR3 induction bysome EOs was close to that of H2O2 although muchweaker than that of MMS, and in comparison to MMSand H2O2, the peak of induction by EOs was delayed bya few hours. These findings indicate that with EOs thesignal of induction may be comparable in amplitude butmay be produced rather delayed when compared withthat of H2O2. These results suggest that treatments withEOs may trigger gene induction via a similar type ofd art-m ucedb[ uta-g D7)[

aagg thatp oRe tingt twog thatM tlya

Since treatments by EOs caused a variety of cellularresponses including cytotoxicity, induction of mito-chondrial damage and gene induction, but no nucleargenetic effects, we asked the question whether theseeffects could be due to cellular stress responses inducedby EOs. Mitochondria are cellular organelles of energyproduction, but at the same time they are importantsources of free radicals and oxidative stress includingROS such as superoxide radicals (O2

•−), hydroxyl rad-icals (OH•) and the termination product hydrogen per-oxide (H2O2) [34,35]. H2O2 can be also produced byoxidative enzymes in peroxisomes and easily permeatemembranes[33]. Exogenous stress disrupts the equilib-rium between the production of such reactive speciesand the cellular defense[34,35]. When mitochondriaare affected, loss of membrane potential, cytochromeC release and cell death by apoptosis can occur[34].Here, we tested the induction of oxidative stress byEOs using suitable inhibitors such as GSH, DFO andcatalase. Reduced glutathione easily reacts with hydro-gen peroxide and superoxide radicals (O2

•−), leadingto less reactive oxidized glutathione or thiol radicals[17–20]. Deferoxamine inhibits the Fenton reaction,i.e. the production of hydroxyl radicals via the reac-tion of Fe2+ iron with hydrogen superoxide[17], andcatalase inactivates hydrogen peroxide[17,20].

Indeed, we were able to demonstrate that ROS andhydrogen peroxide are involved in the biological effectsof EOs. GSH and DFO, but not catalase, protectedagainst the EO ofOriganumsuggesting that this EOp butn teda isE b-a ofA thatt lasean inlyH ox-i tt s-e ciesi alsa hei m-p renti

amage. However, slightly different cellular compents may be affected. Oxidative damage prody treatments with H2O2 can induce theRNR3gene

12] and can also cause cytotoxic and weak menic and recombinogenic effects in diploid yeast (

22,24].Comparison ofRAD51induction by two EOs (Orig-

numandCinnamomum) with that produced by MMSnd H2O2 revealed that the induction of theRAD51ene by EOs was even weaker than that of theRNR3ene and was also delayed when compared withroduced by MMS and H2O2. However, in contrast tNR3induction, the maximalRAD51 induction lev-ls for these latter agents were comparable indica

hat the pathways for gene induction differ for theenes and are not identical. From this, it appearsMS and H2O2 induce the two genes more direcnd earlier than EOs.

roduces hydroxyl radicals and superoxide anionsot H2O2. DFO and catalase, but not GSH, protecgainst the EO ofCoriandrum suggesting that thO produces H2O2 and hydroxyl radicals, but probly not much O2•−. DFO protected against the EOrtemisia, but not GSH and catalase, suggesting

his EO mainly produces hydroxyl radicals. Catand DFO, but not GSH, protected against the EO ofCin-amomum, suggesting that this EO produces ma2O2. The fact that addition of GSH increased cytot

city for CoriandrumandCinnamomummay suggeshat GSH oxidized by H2O2 reacted further in the prence of oxygen giving rise to reactive oxygen spe

ncluding sulphur peroxyl radicals, sulphinyl radicnd O2

•− radicals[18]. Since the EOs and also tnhibitors may differently affect various cellular coartments, these results are likely to reflect the diffe

nteractions at various sites.

12 F. Bakkali et al. / Mutation Research 585 (2005) 1–13

Altogether, the production by EOs of highly dam-aging agents such as OH• and O2

•− radicals and H2O2provides an explanation for the observed cytotoxiceffects, cytoplasmic petite induction and gene induc-tion. In this respect, the quasi-simultaneous expressionof strong cytotoxic effects and mitochondrial damagedoes not appear to be fortuitous. Exposure to EOsstrongly affects the cell wall and membranes and dam-ages mitochondria. This may lead to mitochondrialdysfunction and to a radical burst of reactive oxy-gen species that triggers gene induction and apoptoticcell death. Gene induction by EOs may be broughtabout by the activation of mitogen-activated proteinkinases of signaling cascades involved in the activationof c-Jun like transcription factors (Parl)[36]. In fact,mitochondrial dysfunction is known to increase intra-cellular concentrations of DNA-damaging species suchas superoxide and peroxide ions linked to apoptoticdeath[34,35]. After treatment with EOs, apoptotic-likecell killing appears to be predominant in yeast, and iteffectively avoids nuclear genetic events and genotoxi-city. Thus, the beneficial effects of the five EOs for usein humans are likely based on their capacity to inducecytotoxicity via oxidative stress involving ROS andH2O2. Because of the absence of nuclear mutagenic orrecombinogenic events in yeast after treatments withEOs, the use of these EOs is unlikely to involve long-term genotoxic risks.

A

ndt hipf F).W omt gy,U , fork .

R

tict in

.i-

[3] P. Lo Cantore, N.S. Iacobellis, A. De Mareco, F. Capasso, F.Senatore, Antibacterial activity ofCoriandrum sativum L. andFoeniculum vulgare Miller Var. vulgare(Miller ) essential oils,J. Agric. Food. Chem. 52 (26) (2004) 7862–7866.

[4] Y.-C. Shen, C.-J. Chou, Y.-H. Wang, C.F. Chen, Y.-C. Chou, M.-K. Lu, Anti-inflammatory activity of the extracts from myceliaof Antrodia camphoratacultured with water-soluble fractionsfrom five differentCinnamomumspecies, FEMS Microbiol.Lett. 231 (2004) 137–143.

[5] M. Idaomar, R. El Hamss, F. Bakkali, N. Mezzoug, A. Zhiri, D.Baudoux, A. Munoz-Serrano, V. Liemans, A. Alonso-Moraga,Genotoxicity and antigenotoxicity of some essential oils evalu-ated by wing spot test ofDrosophila melanogaster, Mutat. Res.513 (2002) 61–68.

[6] D. Baudoux, Process analysis and definition of essential oil,Aroma News, Nat. Aromatherap. Dev. 4 (1998) 1–4.

[7] I.B. Chinou, V. Roussis, D. Perdetzoglou, A. Loukis, Chemicaland biological studies on twoHelichrysumspecies of Greekorigin, Planta Med. 62 (1996) 377–379.

[8] D. Averbeck, Relationship between lesions photoinduced bymono- and bi-functional furocoumarins in DNA and genotoxiceffects in diploid yeast, Mutat. Res. 151 (1985) 217–233.

[9] F.K. Zimmermann, R. Kern, H. Rasenberger, A yeast strain forthe simultaneous detection of induced mitotic crossing over,mitotic gene conversion and reverse mutation, Mutat. Res. 28(1975) 381–388.

[10] F.K. Zimmermann, R.C. von Borstel, E.S. von Halle, J.M. Parry,D. Siebert, G. Zetterberg, R. Barale, N. Loprieno, Testing ofchemicals for genetic activity withSaccharomyces cerevisiae:a report of the U.S. Environmental Protection Agency Gene-ToxProgram, Mutat. Res. 133 (3) (1984) 199–244.

[11] L.R. Ferguson, R.C. von Borstel, Induction of the cytoplasmic‘petite’ mutation by chemical and physical agents inSaccha-romyces cerevisiae, Mutat. Res. 265 (1992) 103–148.

[12] X.M. Jia, Y. Zhu, W. Xiao, A stable and sensitive genotoxicssion

.[ ase:

992)

[ bi-ss

[ andpair

[ mi-sictsol.

[ tatessity

[ ceNA

cknowledgements

This work was supported by the Institut Curie ahe CNRS. F.B. acknowledges a doctoral fellowsrom Agence Universitaire de la Francophonie (AU

e are also very much indebted to Dr. Wei Xiao frhe Department of Microbiology and Immunoloniversity of Saskatchewan, Saskatoon, Canadaindly sending us theRNR3–lacZfusion tester strain

eferences

[1] V. Jeannot, J. Chahboun, D. Russel, H. Casabianca,Origanumcompactum Bentham: composition of the hydrolat aromafraction, composition with the essential oil and its interesaromatherapy, Intern. J. Aromatherapie 11 (2) (2003) 87–96

[2] J. Bellakhadar, La Pharmacopee Marocaine Traditionnelle, Edtions Le Fennec, Casablanca, Maroc, 1997, pp. 348–349.

testing system based on DNA damage induced gene exprein Saccharomyces cerevisiae, Mutat. Res. 519 (2002) 83–92

13] S.J. Elledge, Z. Zhou, J.B. Allen, Ribonucleotide reductregulation, regulation, regulation, Trends Biochem. 17 (1119–123.

14] Y. Cohen, M. Dardalhon, D. Averbeck, Homologous recomnation is essential forRAD51up-regulation inSaccharomycecerevisiaefollowing DNA crosslinking damage, Nucl. AcidRes. 30 (2002) 1224–1232.

15] G. Basile, M. Aker, R.K. Mortimer, Nucleotide sequencetranscriptional regulation of the yeast recombinational regeneRAD51, Mol. Cell. Biol. 12 (1992) 289–295.

16] A. Abousekhra, R. Chanet, A. Adjiri, F. Fabre, Semidonant suppressors of Srs2 helicase mutations ofSaccharomycecerevisiaemap in theRAD51gene, whose sequence preda protein with similarities to procaryotic RecA proteins, MCell. Biol. 12 (1992) 3224–3234.

17] R.V. Bensasson, E.J. Land, T.G. Truscott (Eds.), Excited Sand Free Radicals in Biology and Medicine, Oxford UniverPress, Oxford, 1993, pp. 1–431.

18] D. Becker, S. Summerfield, S. Gillich, M.D. Sevilla, Influenof oxygen on the repair of direct radiation damage to D

F. Bakkali et al. / Mutation Research 585 (2005) 1–13 13

by thiols in model systems, Int. J. Radiat. Biol. 65 (5) (1994)537–548.

[19] H. Sies, Glutathione and its role in cellular functions, Free Rad.Biol. Med. 27 (9–10) (1999) 916–921.

[20] C.M. Grant, G. Perrone, I.W. Dawes, Glutathione and catalaseprovide overlapping defenses for protection against hydrogenperoxide in the yeastSaccharomyces cerevisiae, Biochem. Bio-phys. Res. Commun. 253 (1998) 893–898.

[21] IARC Monograph on the evaluation of carcinogenic risks tohumans, IARC Lyon, 71 (Part 2) (1999) 671–689.

[22] IARC Monographs on the evaluation of carcinogenic risks tohumans, IARC Lyon, 71 (Part 3) (1999) 1059–1078.

[23] M. Monaco, R. Dominici, P. Barisano, G. Di Palermo, A. Galli,G. Bronzetti, Mutagenicity of methyl methanesulfonate andcyclophosphamide in resting and growingSaccharomyces cere-visiaeD7 cells, Mutat. Res. 282 (4) (1992) 235–239.

[24] C. Della Croce, G. Bronzetti, M. Cini, L. Caltavuturo, G. Poi,Protective effect of lipoic acid against hydrogen peroxide inyeast cells, Toxicol. In Vitro 17 (2003) 753–759.

[25] M. Ogur, S. John, S. Nagai, Tetrazolium overlay technic forpopulation studies of respiration deficiency in yeast, Science125 (1957) 928–929.

[26] D. Averbeck, S. Averbeck, Induction of the genesRAD54andRNR2by various DNA damaging agents inSaccharomycescerevisiae, Mutat. Res., DNA Repair 315 (1994) 123–138.

[27] D. Averbeck, S. Averbeck, DNA photodamage, repair, geneinduction and genotoxicity following exposures to 254 nm UVand 8-methoxypsoralen plus UVA in a eukaryotic cell system,Photochem. Photobiol. 68 (1998) 289–295.

[28] L. Guarente, Yeast promoters andlacZfusions designed to studyexpression of cloned genes in yeast, Methods Enzymol. 101(1983) 181–191, 563-7.

[29] W.C. Burhans, M. Weinberger, M.A. Marchetti, L. Ramachan-dran, G. D’Urso, J.A. Huberman, Apoptosis-like yeast celldeath in response to DNA damage and replication defects,Mutat. Res. 532 (2003) 227–243.

[30] M. Weinberger, L. Ramachandran, W.C. Burhans, Apoptosis inyeasts, IUBMB Life 55 (8) (2003) 467–472.

[31] S. Manon, Utilization of yeast to investigate the role of lipidoxidation in cell death, Antioxid. Redox Signal 6 (2) (2004)259–267.

[32] A.P. Gasch, P.T. Spellman, C.M. Kao, O. Carmel-Harel,M.B. Eisen, G. Storz, D. Botstein, P.O. Brown, Genomicexpression programs in the response of yeast cells to envi-ronmental changes, Mol. Biol. Cell. 11 (12) (2000) 4241–4257.

[33] M.G. Koerkamp, M. Rep, H.J. Bussemaker, G.P.M.A. Hardy, A.Mul, K. Piekarska, C.A.-K. Szigyyarto, J.M. Teixeira de Mattos,H.F. Tabak, Dissection of transient oxidative stress responsein Saccharomyces cerevisiaeby using DNA microarrays, Mol.Biol. Cell 13 (2002) 2783–2794.

[34] E. Cadenas, Mitochondrial free radical production and cell sig-naling, Mol. Aspects Med. 25 (2004) 17–26.

[35] D.R. Green, G. Kroemer, The pathophysiology of mitochondrialcell death, Science 305 (2004) 626–629.

[36] B. Krems, C. Charizanis, K.-D. Entian, Mutants ofSaccha-romyces cerevisiaesensitive to oxidative and osmotic stress,Curr. Genet. 27 (1995) 427–434.