Activity of essential oils from Syzygium aromaticum and Rosmarinus officinalis against growth and ochratoxin A production by Aspergillus tubingensis and Aspergillus luchuensis from Moroccan grapes

Summary. Essential oils have been used since ancient times in traditional medicine and agri-food science to preserve food, and to combat human diseases. Essential oils (EOs) from clove and rosemary, obtained by hydro-distillation, were analyzed by GC/ MS and evaluated for their antifungal activity against strains of Aspergillus tubingensis and A. luchuensis and their effect on ochratoxin A (OTA) production by these fungi. The major constituent of clove EO was eugenol (86.4%), and of rosemary EO was eucalyptol (35.9%). Mycelium growth inhibition assays showed that the EOs had dose-depend inhibition effects, which reached 36.6% for rosemary EO, and 100% for clove EO, at EO concentration of 200 μL L -1 . These reductions were different for different Aspergillus strains, the essential oil type, and the strain/oil type interaction. Analyses of OTA in the culture medium extracts of Aspergillus strains was carried out using HPLC-FLD, and was confirmed by LC-MS/MS for positive controls of two OTA-pro-ducing strains of A . tubingensis . In general, OTA was reduced (from 45 to 100% reduc-tion), except for two strains of A. tubingensis , where OTA production was stimulated.


INTRODUCTION
Aspergillus species are widely recognized as food contaminants, which may create severe economic losses, and constitute health risk for consumers due to their ability to produce mycotoxins. Ochratoxin A (OTA) is the one of the most frequent toxic secondary metabolites produced by several fungal species belonging to Aspergillus (e.g. A. ochraceus, A. carbonarius, A. niger, A. alliaceus, A. tubingensis, A. luchuensis) and Penicillium (e.g P. verrucosum, P. nordicum, P. expansum). This mycotoxin is hazardous to human health since it can be neurotoxic, carcinogenic, nephrotoxic, hepatotoxic, teratogenic and immunosuppressive (Murphy et al., 2006;IARC, 2009;Coronel et al., 2010). OTA has been classified as a possible carcinogen for humans (group 2B) by the International Agency for Research in Cancer (IARC, 1993). This mycotoxin has been frequently detected in a wide range of agricultural and food products, including cereals and derived products (Hajjaji et al., 2006;Gamza et al., 2015), cocoa (Copetti et al., 2010), grapes, grape and apple juices (Battilani et al., 2006;Leong et al., 2006;Selouane et al., 2009), and spices (Zinedine et al., 2007;Wan Ainiza et al., 2015).
To reduce potential hazards from mycotoxin-producing fungi in crops and stored agricultural products, control of mycotoxin-producing fungi using conventional fungicides is often advocated. However, the use of antifungal chemicals is progressively becoming restricted by food safety regulations, as it could lead to development of resistant strains and occurrence of residues in food, feed, crops and the environment. These fungicides may also stimulate the production of mycotoxins (Audenaert et al., 2010;Zouhair et al., 2014). Therefore, research for alternative antifungal substances is important to control the growth of spoilage fungi and risks of OTA.
Essential oils (EOs) are possible natural alternatives to synthetic fungicides. Several aromatic and medicinal plants have received attention, and have been studied for control of growth of toxigenic fungi to reduce production of mycotoxins (Lappa et al., 2017;Mateo et al., 2017;Lasram et al., 2019;Oliveira et al., 2020). EOs are complex mixtures of volatile organic compounds, mainly terpenoids, aromatic, and aliphatic components, all characterized by low molecular weights (Bassole and Juliani, 2012). The Food and Drugs Administration (FDA) classified theses oils "Generally Recognized as Safe" (GRAS) substances (Edris, 2007). Several biological activities have been attributed to EOs, including antibacterial (Risaliti et al., 2019) and antifungal properties (Bomfim et al., 2015;Santamarina et al., 2016). EOs contain major and minor components, and mixtures are more biologically active than major mixed constituents (Gill et al., 2002;Mourey and Canillac, 2002), which suggests that the minor components may have potentiating influences or synergistic effects.
The aim of the present work was to assess effects EOs from Syzygium aromaticum (clove) and Rosmarinus officinalis (rosemary) on growth and OTA production for eight strains of A. tubingensis and A. luchuensis.

Fungus strains and culture conditions
Seven strains of A. tubingensis (MUCL54478, MUCL54481, MUCL54482, MUCL54483, MUCL54484, MUCL54485 and MUCL54486) and one strain of A. luchuensis (MUCL54477) were used in this study. These strains were previously isolated from grapes produced in Morocco. All strains were identified (Qjidaa et al., 2014), and deposited at BCCM/MUCL (Mycothèque de l'Université Catholique de Louvain, Belgium). The strains (suspensions of conidia) were maintained in 25% glycerol + 0.01% Tween 80 at -20°C. Prior to the study, OTA production of the eight strains was confirmed in vitro on Czapek Yeast Autolysate agar (CYA). The identification of the OTA peak in chromatograms was confirmed by methylation according to Zimmerli and Dick, (1995). Inocula for the experiments were prepared by growing each strain on CYA at 25°C for 7 d. Suspensions of conidia (≈10 5 conidia mL -1 ) were prepared in sterile distilled water containing 0.5% Tween 80. A Thoma chamber was used to determine the final conidium concentrations.

Essential oil isolation
Dried S. aromaticum or dried leaves of R. officinalis (100 g) were hydro-distillated (3 h) using a Clevengertype apparatus. EOs were stored at 4°C in the dark before analysis.

Antifungal tests
In vitro experiments to evaluate the efficacy of EOs against A. tubingensis and A. luchuensis were carried out using a CYA medium (0.99a w ). Conidia of A. tubingensis or A. luchuensis were each exposed to a series of increasing concentrations of the tested antifungal agents (Remmal et al., 1993). The EO was emulsified with a 0.2% agar solution to disperse the constituent compounds. Volumes of this dilution were added to sterilized CYA medium and cooled to 45-50°C. The final essential oil concentrations used were 25, 50, 100 and 200 µL L -1 . Petri plates were prepared with the different amended media, and experimental controls containing the culture medium alone were also prepared. The plates were each single-point inoculated with 10 µL of a suspension of the appropriate conidia (≈ 10 5 conidia), and the plates were then incubated at 25°C. All the experiments were carried out in triplicate. The Petri plates were examined daily, and the diameter of resulting colonies was measured in two perpendicular directions during a period of 7 d. Linear regression of colony radius (mm) against time (d, from the day the colony was 5 mm diam.) was used to determine growth rates (mm d -1 ). Lag phase for growth was considered as the number of days elapsed between inoculation and the time when fungal growth became evident (colony reached 5 mm diam.). Petri plates were also used for OTA extractions and determination.

Essential oils analyses
Gas chromatography-mass spectrometry (GC/MS) analyses were carried out using a Thermo Finnigan gas chromatograph directly coupled to a mass spectrometer system (model TRACE GC ULTRA S/N 20062969; Polaris Q S/N 210729). Retention Indices and the confirmation of MS identified compounds were carried out using a gas chromatograph with flame-ionization detection (GC-FID) Thermos Finnigan Trace GC 2000, under the same conditions as used for GC/MS analyses. Compounds were separated using a TR-5 capillary column (60 m × 0.32 mm, 0.25 µm film thickness). The Oven temperature was maintained at 40°C for 2 min, the programmed to 280°C at 5°C min -1 , and the final temperature kept for 10 min. The splitless injector was maintained at 220°C and opened after 0.85 min. The flow rate of carrier gas (helium) was 1 mL min -1 , and the volume of injected specimen was one µL of diluted oil in hexane. Electron impact mass spectra were recorded at 70 eV. The ion source temperature was 200°C, the interface line temperature was 300°C, and the scan mass range m/z was of 40-650. The same method was previously used by Bennouna et al. (2018).

Extraction and OTA analyses
Ochratoxin A was extracted using the method of Bragulat et al. (2001). Briefly, three agar plugs (diam. = 7 mm) were removed from the inner, middle, and outer areas of each fungus colony. The plugs were weighed and then dispensed into 3 mL vials before adding 1 mL of methanol to each vial. The vials were shaken for 5 sec with an autovortex and then incubated at 25°C for 60 min. The extracts were centrifuged three times for 10 min at 13,000 rpm. The supernatant was then filter-sterilized through a PVDF hydrophilic filter (0.22 μm) and then analyzed using HPLC technology (Agilent Technologies) with fluorescence detection (excitation 333 nm, emission 460 nm; calibration with OTA standard (Sigma Aldrich). The separation of metabolites was carried out on a C18 RP column (Zorbax SB, 4.6 × 250 mm × 5 µm particle size). The mobile phase (acetonitrile-water-acetic acid; 99:99:2, v/v/v) was pumped at 0.7 mL min -1 , and the injection volume reached 20 µL. Run time for samples was 30 min, with OTA being detected at about 11 min. OTA was identified by its retention time (11-11.6 min), according to the OTA standard.
Quantification of OTA was achieved by measuring peak areas according to a linear standard curve, and the limit of detection was 0.2 ppb. All analyses were repeated three times. OTA production by ochratoxigenic fungus strains was confirmed using an UPLC-MS/MS system (Waters) coupled with a triple-quadrupole (Xevo TQ-S micro) mass spectrometer (MS). The OTA was separated using an UPLC C18 column (2.1 µm particle size, 1.7 × 100 mm) with a C18 pre-column (2.1 µm particle size, 1.7 × 20 mm). Both the column and guard column were maintained at 45°C. The flow rate was 0.4 mL min -1 and injection volume of 2 µL. The mobile phase consisted of a gradient (Table 1) achieved with the following mobile phase: phase A) water/0.1% formic acid/5 mM ammonium formate, and phase B) methanol/0.1% formic acid/5 mM ammonium formate. The mass spectrometry analysis ESI-MS/MS was carried out using the positive electrospray ionization mode (ESI+) in a multiple reaction monitoring mode (MRM) at capillary voltage of 3.5 kV, cone voltage of 8 V, source temperature of 150ºC, and desolvation temperature of 500ºC. Argon was used as desolvation with a flow rate of 1000 L h -1 . Full scan mode was employed in the mass range of 30-1250 Da. The transition of the most abundant product transition ion (m/z 404.2 < 239.1) and the second least abundant transition (target or confirmatory) ion (m/z 404.2 < 358.2) were selected for quantification and identification. MassLynx (V4.1) software (Waters Corporation) was used for the data acquisition.

Statistical analyses
Linear regressions of colony radius against time (d) was used to determine the growth rates (mm d -1 ) under each set of conditions, were obtained with the program Microsoft Excel version 2013. Analysis of variance for the different growth data sets was carried out using IBM SPSS Statistics, version 20.

Essential oil composition
Hydro-distillation of clove and rosemary tissues yielded an average, respectively, of 3 mL and 2 mL for 100 g of the dry plant material. Based on the results of the GC/MS and GC/FID analyses, different compounds were identified in the oils ( Table 2). The most abundant chemical components were eugenol (86.4%), eugenyl acetate (8.9%) and a-humulene (1.1%) for clove EO, and eucalyptol (35.9%), camphor (21.2%) and o-cymene (4.8%) for rosemary EO.

Essential oil antifungal effects on fungus lag phases and growth rates
The lag phases observed in CYA cultures supplemented with clove EO were always greater than for rosemary EO, and this under all doses tested. A concentration of 200 µL L -1 of clove oil completely inhibited growth of two ochratoxigenic strains of A. tubingensis (MUCL54482 and MUCL54484) and two non-ochratoxingenic strains A. tubingensis MUCL54486 and A. luchuensis MUCL54477. In contrast, the lag phases for the control cultures were 1 d, and for the cultures supplemented with clove EO were about 2 to 3 d for 25 to 50 µL L -1 , and 4 to 6 d for 100 to 200 µL L -1 . However, the lag phases after rosemary EO treatments were similar for all doses tested (approx. 2 d).
Statistical analyses (Table 3) revealed that factors of the EOs, EO concentrations of EOs, fungus strains, as well as their interactions, had significant effects (P ≤ 0.001) on the antifungal activity. In control plates, mycelium growth had already reached the edges of the Petri dishes after 7 d of incubation, and the mean growth rate was 7.4 mm d -1 . Clove EO at different concentrations (0-200 µL L -1 ) decreased the radial growth rate of all the tested strains (Figure 1). Clove EO applied at 25 µL L -1 gave mean growth rates between 6.4 and 7.3 mm d -1 , and mean growth rate reductions varied between 1.2  (Figures 2 and 3). At 25 µL L -1 , mean fungus growth rates were between 5.3 and 5.7 mm d -1 and resulted in reductions from 18.9 to 33.5% compared to the control. At 200 µL L -1 of rosemary EO, the mean growth rates were between 5.1 and 5.3 mm d -1 , resulting in inhibition between 23.5 to 36.6%. Clove EO therefore displayed greater antifungal potential as a mycelium growth inhibitor than rosemary EO (Figure 3).

Ochratoxin A production
OTA production was assessed for all treatments of cultures of A. tubingensis and A. luchuensis using HPLC-FLD (Table 4). No OTA was detected in culture extracts of non OTA producing strains, with or without clove or rosemary EOs, except for the A. tubingensis MUCL54478 strain grown in presence of clove EO at 25 µL L -1 , where 8.3 µg g -1 OTA was detected. For OTA producing strain controls, the OTA amounts produced were between 0.06 to 0.14 µg g -1 . The rosemary EO reduced accumulation of OTA for all the strains stud-ied, except for A. tubingensis strains MUCL54483 and MUCL54484 at 200 µL L -1 , where OTA was increased (means of 0.44 µg g -1 for MUCL54483 and 0.22 µg g -1 for MUCL54484). Where clove EO was applied at 50, 100 or 200 µL L -1 , OTA has not been detected (depending on the strain), but exceeded the control levels of OTA, especially for strain MUCL54484 at 25 or 50 µL   L -1 (respectively, 0.41 and 0.43 µg g -1 ). Confi rmation of OTA production by A. tubingensis strains MUCL54482 and MUCL54483 was carried out by UPLC-MS/MS, with production of 1050 µg kg -1 by A. tubingensis strain MUCL54482 (Figure 4). DISCUSSION GC/MS analyses of the clove and rosemary EOs from Moroccan origins showed that eugenol was the major component of Clove EO, and eucalyptol (1,8-cin-Essential oils Doses (µL L -1 )   eole) was the major compounds in rosemary EO. For clove EO, eugenol was 86.4% of the total volatile compounds. This agrees with the results of Santamarina et al. (2016), who showed that the proportion of eugenol in commercial clove EO was 88.6%. Bozik et al. (2017) also reported that eugenol was the major constituent in two commercial clove Eos, with amounts of 81.7 and 77.8%. In addition, Boukaew et al. (2017) and Sharma et al. (2017) concluded, respectively, that eugenol constituted 62.4 and 75.4% of clove EO volatiles. Results for rosemary EO in the present study showed that eucalyptol was 35.8% of total volatiles. Risaliti et al. (2019) also reported similar results, with a proportion of 48.7%. Farhat et al. (2017) reported that the proportion of eucalyptol in Tunisian rosemary EO ranged from 20.8 to 64.7% depending according on the extraction methods used. Analysis of rosemary EO volatiles from Brazil (Takayama et al., 2016) and Morocco (Bouyahya et al., 2017) showed that the major component was eucalyptol, at 28.5% in Brazil and 23.67% in Morocco. Studies of the chemical composition of EOs from clove and rosemary revealed variability in their chemical profiles dependent on plant growing conditions, phenological stage, sampling time, and extraction and analytical methods used, and even existence of new plant chemotypes as suggested by Kokkini et al. (2004) and Sharma and Tripathi (2008). The chemical composition and antibacterial activity of clove and rosemary EOs has been described by several authors. However, studies on antifungal activity of these essential oils against A. niger aggregates and their effects on OTA synthesis are scarce.

OTA (µg g -1 ) (a) (Mean±S.D. (b) ) OTA non-producing strains OTA producing strains
In the present study, the lag phases, radial growth and OTA accumulation by strains of A. tubingensis and A. luchuensis were strongly influenced by clove and rosemary EOs. In general, the EO type, their concentrations, fungus isolate, and interactions between these factors had highly significant effects on radial fungus growth. For clove EO, mean inhibition rates greater than 90% were obtained at 200 μL L -1 , showing that this essential oil had effectiveness exceeding that for some fungicides widely used in Morocco (azoxystrobin, benomyl, hexaconazole and pyrimethanil) (Zouhair et al., 2014;Laaziz et al., 2017). Inhibition rates for clove EO were between 1.2 and 13 3% at the lowest concentration of the EO (25 μL L -1 ), and were 69.2 to 100% for 200 μL L -1 . Passone et al. (2012) reported that the application of 500 µL L -1 of clove EO inhibited the growth of Aspergillus section Nigri by 48.8 to 100%, depending on EO concentration, fungus strains and water activity levels. Santamarina et al. (2016) reported significant antifungal activity of clove EO against Fusarium spp. They demonstrated that for a concentration of 300 mg mL -1 , there was a growth reduction of 88% for F. graminearum. Boukaew et al. (2017) also studied the antifungal effect of clove EO but on A. flavus responsible of the deterioration of maize in storage in Thailand. They demonstrated that low concentrations of clove essential oil (1 and 10 µL L -1 ) induced greater growth inhibition (respectively, 85.7 and 94.0%), and complete growth inhibition (100%) was achieved at EO concentrations of 50 and 100 μL L -1 .
This study showed that A. tubingensis growth inhibition rates from treatments with rosemary EO ranged from 23.5 to 36.6%, depending on the EO concentration. Bouddine et al. (2012) also reported that rosemary EO showed a weak effect on A. niger, as assessed using the broth dilution method. Rasooli et al. (2008), using the disc diffusion method, showed antifungal activity of rosemary EO at 20 µL, giving 17% inhibition of A. parasiticus. Bomfim et al., 2015) showed that the same essential oil inhibited growth of F. verticillioides by 29.7% at a concentration of 300 µg mL -1 . The weak effect (1.6 to 24.2% at 1 µL mL -1 ) of rosemary EO on growth inhibition for fungal pathogens was also demonstrated by Matusinsky et al. (2015).
Investigation of synergistic or potential effects of different constituent compounds of EOs can increase insights into how EO composition affects antifungal activities. Morcia et al. (2012) demonstrated that eugenol, the major component of clove EO, had high activity against A. tubingensis and A. carbonarius. However, eucalyptol (the major component rosemary EO) was only active at the greatest doses tested (2%). Nazzaro et al. (2017) also reported that eucalyptol and eugenol exhibited important antifungal activity. Several studies demonstrated that the lipophilic properties and low molecular weights of essential oils allow them to cross cell membranes, causing irreversible cell wall and cellular organelle damage and affecting inorganic ion equilibria and pH homeostasis (Pawar et al., 2006;Helal et al., 2007).
High efficacy of EOs in disease control has been reported by several authors. The use of EOs in the packaging of fruits and other commodities was reported by Sivakumar et al. (2014, Guerra-Rosas et al. (2017, Munhuweyi et al. (2017), andTalebi et al. (2018). Campos et al. (2016) reported that the use of thyme and sage EOs in the headspace of strawberry packages led to decreases fruit contamination. To improve shelf-life without changing the organoleptic characteristics of jujube fruit, Nikkhah and Hashemi (2020) demonstrated that the treatments of the fruit with thyme, cinnamon or rosemary EOs (at, respectively, 0.156, 0.625, or 0.078 g L -1 ), during 60 d of storage, did not negatively impact on taste, texture, aroma or overall acceptance. They also suggested that using multiple essential oils at low concentrations may increase the antifungal activity because of synergistic actions. They demonstrated that, the concentration of EOs penetrating the fruit flesh remained under taste detection thresholds. According to the European Food Safety Authority (EFSA, 2012), the Acceptable Daily Intake (ADI) and Acceptable Operator Exposure Level (AOEL) for eugenol are 1.0 mg kg -1 body weight d -1 . In the present study, total inhibition for clove essential oil was obtained at 200 µL L -1 meaning that the use of this essential oil to conserve fruits in storage is likely to be safe for consumers. On the other hand, based on the Maximized Survey-derived Daily Intake (MSDI) approach, EFSA (2013), reported that 1,8-cineole (eucalyptol) presented no safety concern (1200 µg capita -1 d -1 ) In parallel to the fungal growth inhibitory effect by the two tested EOs, there were contradictory effects (reduction or stimulation) on OTA biosynthesis caused by the tested Aspergillus strains. Although several authors have reported that A. tubingensis produces OTA, the ability of this species to produce this toxin is controversial. Studies have shown that strains of A. tubingensis were able to produce a metabolite identified by HPLC-FLD as OTA (Medina et al., 2005;Perrone et al., 2006;Selouane et al., 2009;Chiotta et al., 2011). Other studies have considered that A. tubingensis was not an OTA producer (Abarca et al., 2004;Frisvad et al., 2011). Pantelides et al. (2017) also showed that none of 261 strains of A. tubingensis isolated from wine grapes were found to be ochratoxigenic when analyzed by UPLC-MS/MS, while some of these isolates were initially considered to be toxin producers as indicated by HPLC-FLD. In the same way, Storari et al. (2012) re-examined A. tubingensis strains previously described as ochratoxigenic using LC-MS, and could not detect the toxin, concluding that A. tubingensis cannot be considered as an ochratoxigenic species. This agrees with more recent publications (Gil Serna et al., 2019 andTavakol et al., 2020). Therefore, data related to OTA production in controls in the present study was validated using an UPLC-MS/ MS based method, in order to be sure that the observed peak was not an artifact ( Figure 4). The present results showed that the A. tubingensis strains used in this study produce OTA, and this is the first report where production of OTA by A. tubingensis has been confirmed using the UPLC-MS/MS technique.
The inhibitory effect of rosemary EO on OTA production was also reported by El , who found that this EO inhibited production of OTA by A. carbonarius. Bomfim et al. (2015) concluded that rosemary EO reduced fumonisin B1 and B2 contamination by F. verticillioides. Experiments with A. parasiticus clearly illustrate reduction of aflatoxin production from rosemary EO (Rasooli et al., 2008). The use of eugenol also reduced aflatoxin B1 production by A. parasiticus and A. flavus (Jahanshiri et al., 2015;Caceres et al., 2016).
Several studies have shown that use of some essential oils as biocontrol agents against fungi lead to stimulation of the mycotoxin biosynthesis. Stimulation of OTA production depended on the strain, nature of the essential oil and the applied dose. Results of the present study showed that sub-lethal concentrations of EOs stimulated production of OTA by A. tubingensis. Lappa et al. (2017) demonstrated that the treatment of A. carbonarius with lemongrass, cinnamon or mandarin essential oils decreased growth rate, but increased OTA production. Stimulation of OTA biosynthesis by essential oils was also reported by Passone et al. (2012), who demonstrated that species of A. niger aggregate and A. carbonarius tended to stimulate OTA production in the presence of 1000 μL L -1 of poleo essential oil. In addition, Mateo et al. (2017) also reported stimulation of AFB1, AFB2, AFG1 and AFG2 accumulation in some cultures treated with ethylene-vinyl alcohol copolymer films containing oregano, carvacrol and cinnamon. These results are similar to those reported by Bluma and Etcheverry. (2008) for the species Aspergillus section flavi.
Mycotoxin production may be stimulated when stressful environmental conditions and sub-lethal antifungal agent doses are maintained in culture media during the growth of mycotoxin-producing fungi (Mateo et al., 2013;Prakash et al., 2015).
In conclusion, the effects of clove and rosemary EOs on fungal growth and OTA production by different Aspergillus isolates have shown that these oils had significant antifungal activity. Inhibition rates varied (20 to 100%) for fungus isolate and essential oil dose. In general, the statistical analyses showed that the different concentrations of clove and rosemary EOs, the Aspergillus strains, and their interactions, had strong effects on fungus growth. Mean inhibition rates greater than 85% were obtained at 200 μL L -1 of clove EO, which exceeds some fungicides (azoxystrobin, benomyl, hexaconazole and pyrimethanil) that have been tested and are widely used in Morocco. Clove EO can be used to control fungal growth and mycotoxin production as a potential alternative to chemical control. However, more research is required to evaluate the in vivo efficacy of these EOs, and to understand their effects on the expression of OTA biosynthesis and regulation genes.