In Vivo Antiphytoviral and Aphid Repellency Activity of Essential Oils and Hydrosols from Mentha suaveolens and Foeniculum vulgare to Control Zucchini Yellow Mosaic Virus and Its Vector Aphis gossypii

In recent years, natural compounds have gained attention in many fields due to their wide-range biological activity. In particular, essential oils and their associated hydrosols are being screened to control plant pests, exerting antiviral, antimycotic and antiparasitic actions. They are more quickly and cheaply produced and are generally considered safer for the environment and non-target organisms than conventional pesticides. In this study, we report the evaluation of the biological activity of two essential oils and their corresponding hydrosols obtained from Mentha suaveolens and Foeniculum vulgare in the control of zucchini yellow mosaic virus and its vector, Aphis gossypii, in Cucurbita pepo plants. The control of the virus was ascertained with treatments applied either concurrently with or after virus infection; choice tests were performed to verify repellency activity against the aphid vector. The results indicated that treatments could decrease virus titer as measured using real-time RT-PCR, while the experiments on the vector showed that the compounds effectively repelled aphids. The extracts were also chemically characterized using gas chromatography–mass spectrometry. Mentha suaveolens and Foeniculum vulgare hydrosol extracts mainly comprised fenchone and decanenitrile, respectively, while essential oils analysis returned a more complex composition, as expected.


Introduction
Essential oils (EOs) are volatile, complex mixtures of strong-scented natural compounds synthesized through the specialized metabolism of aromatic plants typical of temperate and tropical regions and stored in secretory cells, glands or trichomes [1]. They are usually extracted by means of steam-or hydro-distillation (SD or HD) [2] and using liquid or supercritical carbon dioxide [3]. EO chemical composition is analyzed using gas chromatography-mass spectrometry (GC-MS); their main components are terpenes, terpenoids, aromatic and aliphatic low-molecular-weight molecules and their oxygenated derivatives. In nature, EOs are involved in plant defense systems as antibacterials, antivirals, antifungals, insecticides and repellents [4].
Because of their wide-range biological activity [5][6][7], EOs are used in the pharmaceutical and food industry and are increasingly popular as antimicrobials in plant protection. In fruits, negatively affecting their marketability. After infection, plants no longer provide marketable production within 1-2 weeks. Such a rapid settlement of severe disease causes a yield decrease of up to 90% and subsequent economic losses [42]. The use of resistant varieties is currently a valid strategy for the control of ZYMV in C. pepo, but resistancebreaking strains are increasingly widespread; hence, resistance should always be used in combination with other control methods [43]. ZYMV is transmitted by several aphid species in a non-persistent manner. Aphis gossypii Glover (Hemiptera: Aphididae) plays a key role in ZYMV spread because it is widely distributed worldwide and has a high transmission efficiency [44]. Moreover, it has been reported to rapidly develop resistance to several synthetic insecticides [45].
Given the above-described data and hypothesis, and the urgent need for natural compounds to treat plant diseases and pests, we investigated the antiphytoviral and aphidrepellent activities of either EOs or HSs extracted from M. suaveolens and F. vulgare. Plants were treated with EOs and HSs concurrently with virus inoculum to measure the EO or HS effect on virus vitality and infectivity in vitro, and after virus inoculation, to test the "curative" activity of extracts. The potential effects against ZYMV were assessed by comparing the relative virus titer as measured with real-time RT-PCR in systemically infected leaves of treated and untreated plants of C. pepo. Moreover, the expression of the phenylalanine ammonia lyase (PAL) gene was quantified to explore a possible mode of action for treatments in plants. PAL is the first enzyme of the phenylpropanoid pathway, which is involved in plant defense mechanisms; the expression of the related gene is reported to be a marker of plant response to biotic stress through transcriptional regulation [46,47]. Choice tests of A. gossypii specimens between treated and untreated leaves were also performed to measure the repellent action of EOs and HSs against the ZYMV vector; the toxicity of treatments was also evaluated by measuring mortality and fecundity in adults. Our results showed that the tested extracts displayed antiphytoviral and repellency activity against the target virus and vector.

Chemical Composition of Plant Extracts
The chemical classes identified using GC-MS in EOs and HSs are summarized in Table 1, while the detailed chemical composition is provided in Table S1. A total of 60 molecules belonging to 9 classes of compounds and 19 molecules belonging to 6 classes were identified in EOs and HSs, respectively; thus, EOs generally had a more complex composition than HSs. The identified compounds accounted for 82.38-94.75% of EO or HS composition. Monoterpene hydrocarbons and oxygenated monoterpenes were the main components of EOs; HSs were generally rich in oxygenated monoterpenes. The M. suaveolens EO composition was variegated, with 42 identified compounds showing relative abundance above 0.1%; the main compounds were isomintlactone and menthene (7.72 and 7.28%, respectively). Monoterpene hydrocarbons and oxygenated monoterpenes accounted for about 40% of the composition, followed by minor amounts of esters (about 10%). In the corresponding HS, the main component of the ether extract was identified to be decanenitrile (68.46%), which was not recognized in M. suaveolens EO. Oxygenated monoterpenes were also present in fair abundance (about 16%). Regarding F. vulgare extracts, EO contained monoterpene hydrocarbons and oxygenated monoterpenes for about 60%, while esters and aromatic compounds accounted for about 30%. Its HS displayed the main component of the extract, i.e., fenchone, at 75.93% relative abundance; it was also present in the corresponding EO but at a lower concentration (16.63%).

Treatments Applied at the Same Time as Inoculation
The evolution of the ZYMV relative titer in leaves of treated C. pepo plants over time was investigated with repeated weekly sampling between 7 and 28 days post infection (d.p.i.) as described in Section 4.4.1. In Figure 1, panel A reports the fold change (i.e., Plants 2023, 12, 1078 4 of 16 2 −∆∆Ct ) values of the relative ZYMV titer obtained at the four sampling times with all treatments. In the first sampling instance, 7 d.p.i., the results showed a 2-3-fold increase in the virus relative titer with all treatments, including ribavirin. However, 14 d.p.i., the fold change of relative virus titer was around zero with all treatments, indicating no effects of applying EOs nor HSs on the control of ZYMV. Then, 21 d.p.i, all EO and HS treatments consistently had fold changes of the order of magnitude of 10 −1 , suggesting an effect of treatments on reducing the relative virus titer. Finally, 28 d.p.i., a major influence of treatments was noticeable on the relative virus titer; with M. suaveolens HS and F. vulgare extracts, fold changes of the order of magnitude of 10 −3 were observed, while the relative virus titers were decreased by 10 −1 in M. suaveolens EO-treated plants and by 10 −2 in ribavirin-treated plants. In samples of all mock-inoculated plants in this and the following sections, the virus was not detected at any time, and such samples were not included in the graphs. Symptoms of phytotoxicity were not observed in any of the treated plants in this and the following sections, with all EOs and HSs. The evolution of the ZYMV relative titer in leaves of treated C. pepo plants over time was investigated with repeated weekly sampling between 7 and 28 days post infection (d.p.i.) as described in Section 4.4.1. In Figure 1, panel A reports the fold change (i.e., 2 −ΔΔCt ) values of the relative ZYMV titer obtained at the four sampling times with all treatments. In the first sampling instance, 7 d.p.i., the results showed a 2-3-fold increase in the virus relative titer with all treatments, including ribavirin. However, 14 d.p.i., the fold change of relative virus titer was around zero with all treatments, indicating no effects of applying EOs nor HSs on the control of ZYMV. Then, 21 d.p.i, all EO and HS treatments consistently had fold changes of the order of magnitude of 10 −1 , suggesting an effect of treatments on reducing the relative virus titer. Finally, 28 d.p.i., a major influence of treatments was noticeable on the relative virus titer; with M. suaveolens HS and F. vulgare extracts, fold changes of the order of magnitude of 10 −3 were observed, while the relative virus titers were decreased by 10 −1 in M. suaveolens EO-treated plants and by 10 −2 in ribavirin-treated plants. In samples of all mock-inoculated plants in this and the following sections, the virus was not detected at any time, and such samples were not included in the graphs. Symptoms of phytotoxicity were not observed in any of the treated plants in this and the following sections, with all EOs and HSs.

Treatments Applied after Inoculation
In Figure 1, panel B shows the evolution of the ZYMV relative titer over time when treatments were applied after inoculation. An early decrease in relative virus titer was observed 7 d.p.i. in treatments with M. suaveolens extracts of the orders of magnitude of 10 −2 with EO and 10 −1 with HS. With F. vulgare EO, the virus titer was around one half of the infected untreated control, while no changes were recorded with F. vulgare HS treatment. Two weeks after inoculation, M. suaveolens extracts maintained the same levels of virus titer observed in the first sampling instance, while F. vulgare extracts showed a decrease, reaching one-tenth of the control with EO and one-third with HS. Only M. suaveolens EO-treated plants showed a relative virus titer below the control 21 d.p.i., while all other treatments were 2-3-fold higher. In the final sampling instance, 28 d.p.i., both EOs displayed a virus titer decrease of the order of magnitude of 10 −1 , while in HS treatments, the virus was at levels above the control. The positive-control treatment with ribavirin had the same trend as most experimental treatments, showing a peak decrease in virus titer 14 d.p.i.

Treatments Applied after Inoculation
In Figure 1, panel B shows the evolution of the ZYMV relative titer over time when treatments were applied after inoculation. An early decrease in relative virus titer was observed 7 d.p.i. in treatments with M. suaveolens extracts of the orders of magnitude of 10 −2 with EO and 10 −1 with HS. With F. vulgare EO, the virus titer was around one half of the infected untreated control, while no changes were recorded with F. vulgare HS treatment. Two weeks after inoculation, M. suaveolens extracts maintained the same levels of virus titer observed in the first sampling instance, while F. vulgare extracts showed a decrease, reaching one-tenth of the control with EO and one-third with HS. Only M. suaveolens EO-treated plants showed a relative virus titer below the control 21 d.p.i., while all other treatments were 2-3-fold higher. In the final sampling instance, 28 d.p.i., both EOs displayed a virus titer decrease of the order of magnitude of 10 −1 , while in HS treatments, the virus was at levels above the control. The positive-control treatment with ribavirin had the same trend as most experimental treatments, showing a peak decrease in virus titer 14 d.p.i.

Treatments Applied at the Same time as Inoculation
The evolution of PAL expression (i.e., 2 −ΔΔCt ) in plants when treatments were applied at the same time as inoculation is shown in Figure 2

Treatments Applied after Inoculation
PAL relative expression in plants treated after inoculation is shown in Figure 2, panel B. Overall, no significant changes in the expression levels of the gene were observed with any experimental treatment at any sampling time. Ribavirin-treated plants had a peak PAL expression 14 d.p.i.

Effect of Plant Extracts on Virus Vector Choice and Survival
Both EOs and HSs obtained from M. suaveolens and F. vulgare were tested to determine their potential effects on the settling of A. gossypii adults on treated C. pepo leaves, and adult survival and fecundity. All the bioassays were performed using 300 µg/mL EO or 1:2 v/v HS concentration, which were the same used for plant treatments (Sections 4.4.1 and 4.4.2).

Repellency
Choice tests between EO-or HS-treated and water-treated leaves showed that a significantly higher proportion of aphids preferred to settle onto the control leaves rather than the treated leaves with all the tested compounds ( Figure 3). This repellency effect started 1 h after treatment and persisted for 24 h. The repellency ascribable to M. suaveolens was significantly stronger with EO than with HS in all the time intervals. F. vulgare EO and HS had similar effects on aphid settlement 1 h and 2 h after treatment; then, HS repellency significantly decreased at 4 h and 24 h. In general, M. suaveolens HS was less efficient than the other compounds in inhibiting A. gossypii adults from settling on treated leaves. PAL expression 14 d.p.i.

Effect of Plant Extracts on Virus Vector Choice and Survival
Both EOs and HSs obtained from M. suaveolens and F. vulgare were tested to determine their potential effects on the settling of A. gossypii adults on treated C. pepo leaves, and adult survival and fecundity. All the bioassays were performed using 300 μg/mL EO or 1:2 v/v HS concentration, which were the same used for plant treatments (Sections 4.4.1 and 4.4.2).

Repellency
Choice tests between EO-or HS-treated and water-treated leaves showed that a significantly higher proportion of aphids preferred to settle onto the control leaves rather than the treated leaves with all the tested compounds ( Figure 3). This repellency effect started 1 h after treatment and persisted for 24 h. The repellency ascribable to M. suaveolens was significantly stronger with EO than with HS in all the time intervals. F. vulgare EO and HS had similar effects on aphid settlement 1 h and 2 h after treatment; then, HS repellency significantly decreased at 4 h and 24 h. In general, M. suaveolens HS was less efficient than the other compounds in inhibiting A. gossypii adults from settling on treated leaves.

Toxicity and Fecundity
Both the M. suaveolens and F. vulgare HS treatments significantly increased the mortality rates of A. gossypii adults compared with the control water treatment in most of

Toxicity and Fecundity
Both the M. suaveolens and F. vulgare HS treatments significantly increased the mortality rates of A. gossypii adults compared with the control water treatment in most of the observed time intervals (Table 2). M. suaveolens HS was responsible for a highly significant toxic effect early after treatment, causing 58.6 and 62.1% of mortality after 24 and 48 h, respectively. The survival of F. vulgare HS-treated aphids was significantly reduced 24, 72 and 96 h after treatment compared with the control, but this effect was not as sharp as the M. suaveolens HS effect. Treatments with M. suaveolens and F. vulgare EOs showed lower levels of aphid mortality, which often did not significantly differ from the water control, especially in the earliest daily data collection instances.
M. suaveolens HS treatment was also responsible for significantly reducing offspring production compared with the water control in the first three daily data collection instances (Table 3). No significant differences in aphid fecundity were found between the other treatments and the control. Table 2. Results of toxicity assay on M. suaveolens and F. vulgare essential oils and hydrosols against A. gossypii adults. Twenty replicates were carried out, and 10 adults were used per replicate (N = 200). MS EO = M. suaveolens essential oil; MS HS = M. suaveolens hydrosol; FV EO = F. vulgare essential oil; FV HS = F. vulgare hydrosol. Toxicity is expressed as adult mortality percentage for each treatment; Chisquare values and significance of differences reported are based on Chi-square tests for each treatment vs. untreated control. Significance is expressed as ** = p < 0.01; *** = p < 0.001; **** = p < 0.0001.

Discussion
The use of natural resources from plant species to control plant viral diseases and their insect vectors has been broadly investigated in recent years. In particular, EOs and HSs have gained interest due to the important role they play in nature in the protection of plants as antibacterials, antivirals, antifungals, insecticides and repellents of undesirable insects. A treatment based on a natural compound with safety characteristics favorable for human health and the environment, and effective in controlling plant viruses and repelling their insect vectors is highly necessary for modern agricultural systems worldwide. EOs and HSs have been tested as antiphytovirals, insecticides and repellents, reporting promising results [31,[48][49][50]. During our previous study, we investigated the potential biological activity of essential oils and hydrosols from O. vulgare, T. vulgaris and R. officinalis against ZYMV and tomato leaf curl New Delhi virus (ToLCNDV) in C. pepo [20]. The results of our study supported the feasibility of using such compounds to control ToLCNDV, whereas poor biological activity was observed against ZYMV. The established experimental protocol, involving inoculation, treatment, sampling, and measurement of virus titer and plant gene expression, was considered reliable and reproducible for further use in similar studies. Hence, in this study, we used the same system to test more EOs and HSs from two other plant species, M. suaveolens and F. vulgare; we also report their repellent activity against the aphid vector of ZYMV, A. gossypii, to potentially combine control of virus and vector with the same treatment. The choice of the two plant sources of extracts was based on the repellency activity against A. gossypii exerted by M. suaveolens and F. vulgare extracts, which was ascertained in a preliminary broader screening involving different extracts from five plant species. Given the above-mentioned activity, we hypothesized that M. suaveolens and F. vulgare EOs and HSs could also effectively control ZYMV in C. pepo; the extracts were first characterized using GC-MS and then tested as treatments on ZYMV-infected C. pepo plants in vivo. The verification of our hypothesis was accomplished by means of (i) the measurement of the relative virus titer in new leaves of systemically infected, treated plants and (ii) the evaluation of the relative expression of the PAL gene, to investigate a potential mechanism of action of the extracts. In fact, the biological activity of natural compounds is often ascribed to an indirect mode; rather than directly damaging the target organism, they are assumed to stimulate plant defense response, reducing pathogen load and symptomatology, making the host more tolerant to the pathogen, and ultimately remaining productive [32].
A comprehensive evaluation of treatments with natural compounds necessarily implies the chemical characterization of extracts, which can present broad variability of composition due to many factors related to the plant material (harvesting season, plant chemotype and cultivar, type of tissue and age of the plant) and the distillation method [51]. Hence, it is crucial to characterize the actual composition of the extract used in treatments, to know the compounds applied to the plant and possibly to identify biologically active molecules in the mixture. Given the variability of the extract composition, the knowledge of active molecules is, therefore, of great importance; under certain restraints, the actual performance of an extract can be assessed by verifying the presence of active molecules, whatever the residual composition.
EO and HS dissolved organic compounds were chemically characterized using GC-MS in terms of qualitative and quantitative relative composition. M. suaveolens EOs from different sources and geographical origins were already studied in the literature [52,53], and overall, all showed high percentages of oxides, which is confirmed by our results. Decanenitrile, the main compound of M. suaveolens HS, was not reported in high concentration in previous literature studies on this extract; the significant presence of hydrophilic oxygenated compounds previously observed, which is a typical feature of HSs, is instead in accordance with our results. Regarding F. vulgare extracts, the fairly high concentrations of fenchone and anethole observed in EO were already reported in the literature [54], whereas pentanedioic acid (p-t-butylphenyl)ester was found in high concentration in our experiments but was not previously reported in fennel extracts. As expected, F. vulgare HS ether extract was enriched in fenchone, an oxygenated compound already present in EO and accumulating in HS due to its polar moiety.
The characterization of both M. suaveolens HS and F. vulgare EO thus evidenced the presence in high concentrations of a compound not previously reported in similar extracts. This is quite common in the study of plant extracts, whose composition is strongly affected by the numerous factors mentioned above and may vary significantly.
The experiments in which treatments were mixed with virus inoculum and then applied to the plants (described in Section 4.4.1) were aimed to first assess the potential effect of EOs or HSs on the vitality and infectivity of ZYMV in vitro. This procedure, well reported in the literature as "inhibition activity" assay, is often performed as a preliminary experiment to test the possible interaction of EOs or HSs and the virus under the simplest conditions, involving in vitro contact between virus and treatment; then, the effect on a plant inoculated with such mixture is observed [15,16,18,19]. This procedure is necessary, because viruses are obligate parasites, so the mere in vitro inhibition test is not feasible. In this situation, the mechanism of action of the treatment could be retrieved via interference with coat proteins or the inhibition of the formation of capsid proteins, which are necessary for adsorption or entry into the host plant.
In this assay, a late but effective response was observed in terms of a decrease in viral titer. Fold change was substantially insignificant until 21 d.p.i. with all treatments, whereas 28 d.p.i., a huge decrease was observed, outperforming ribavirin in the case of both M. suaveolens HS and F. vulgare extracts. All these extracts have been previously described for their biological activity [36,[55][56][57], but they have never been tested for the control of phytopathological viruses. Menthene, carvone, limonene and eucalyptol have been described to be foremost responsible for most biological activities and were also detected in our extracts.
The experiments in which treatments were applied after virus inoculation were performed to assess the potential of treatments to reduce the damage of an established infection, i.e., "curative" activity, as it is referred to in the literature [16,18]. The evolution of the ZYMV titer with time was very different when treatments were applied after inoculation; this timing, even though in the frame of an experimental trial under controlled conditions, is probably more adherent to the actual situation of infection/treatment in crop management. In this experiment, the response to treatment was recorded earlier but to a smaller extent. The most effective treatment was M. suaveolens EO, which maintained its activity at all sampling times.
The evaluation of PAL expression levels in treated plants was performed to investigate the likely mode of action of treatments involving plant defense response; in fact, PAL is a key enzyme in the phenylpropanoid biosynthetic pathway, and such compounds play a role in plant defense against many pathogens [58]. PAL is upregulated upon virus infection [47,59] and is involved in resistance development [46]. In this work, the PAL gene was chosen to first assess the hypothesis of an indirect mode of action of treatments, i.e., stimulating plant response rather than directly damaging the target pathogen. The overexpression of the PAL gene upon treatment indicates a plant response activating metabolic pathways to produce defense compounds.
In treatments applied concurrently with inoculation, the upregulation of PAL was observed to have a bimodal trend with most treatments, with two peaks 14 and 28 d.p.i., whereas ribavirin had a consistently increasing trend resembling an exponential curve. The peak 14 d.p.i. was also confirmed when treatments were applied after inoculation for most extracts, including ribavirin. In our previous work, the expression of PAL was also measured in a similar experiment with EOs and HSs from other plant species; in such a trial, HSs generally displayed better performance than EOs, while for M. suaveolens and F. vulgare, a similar trend was not confirmed. Based on these results, we can speculate that the activity of PAL regulation is probably more ascribable to the chemical composition of single extracts than to the nature of the extract (EO or HS).
PAL upregulation was already observed upon treatment with EOs; when these extracts were applied in the post-harvest treatment of fruits and vegetables to control molds, overexpression of PAL occurred in the treated material [60].
Another option for assessing the activity of plant treatments is the application before virus inoculation, i.e., "protective" activity, as referred to in the literature [16,19]; in such assays, the ability of treatments to prevent the establishment of systemic infection is evaluated. We did not perform this experiment, but to date, this application mode has been successfully tested on different pathosystems.
Finally, EO and HS treatments also showed promising repellent activity against the vector of ZYMV, A. gossypii. Many EOs are currently used as repellents against harmful insects, and they are consistently considered good alternatives to synthetic molecules; some of them are registered by environmental protection agencies for such use. This field is still extensively studied [31] and has been recently extended to the use of hydrosols for the same purposes [34,61]. Extracts from several species of the genus Mentha, such as EOs from Mentha piperita and Mentha longifolia, and HSs from Mentha pulegium, are known to actively repel aphids [34,62]. Our data showed that M. suaveolens extracts could also play an important role in inhibiting A. gossypii settlement onto treated plants, and especially, EOs repelled around 90% of the tested adults for at least 24 h after treatment. Regarding F. vulgare extracts, few data are currently available on their potential repellent activity [63]. Our results show that both HSs and especially EOs were able to efficiently reduce the settling of A. gossypii on treated zucchini leaves compared with the untreated control. As expected, the F. vulgare extracts contained high concentrations of fenchone and anethol. These compounds have already been targeted as bioinsecticides against aphids [40,64,65] and are now also strongly suggested for their potential role as repellents.
M. suaveolens and F. vulgare EOs did not show significant toxic effects against A. gossypii adults at the applied concentrations. HSs induced higher mortality levels than the control at different time points, but only M. suaveolens HS showed to be highly effective within a short time after treatment. Indeed, this compound halved the aphid population in just one day and significantly reduced the offspring production of the remaining adult specimens. The promptness of the toxic effect is an essential requisite of aphicide compounds, which should reduce the target population before the aphids start to actively suck the host's sap. In this way, the chance of virus acquisition/transmission, which naturally occurs through rapid sucking punctures, would decrease. The early toxic effect of M. suaveolens HS has already been assessed against another aphid species, Toxoptera aurantii, confirming the potential of this hydrosol in aphid control [35].

Aromatic Plants
The source plants were obtained through an ongoing project aimed to investigate how EO production and associated chemical composition can be influenced by different cropping techniques; in the frame of this project, M. suaveolens and F. vulgare plants were harvested and subjected to EO distillation and HS separation. M. suaveolens Ehrh and F. vulgare Mill. plants were grown at Stazione di Base del Centro Appenninico del Terminillo "Carlo Jucci" in Rieti (Italy). The initial transplant was performed in September 2016. The plants were harvested in summer 2018; then, they were dried, sealed and stored in a closet.

Essential Oils
EOs extracted in the frame of the project were obtained in low amounts to be used for all activities; hence, additional amounts were purchased from Farmalabor srl (Assago, Italy), and their chemical composition was analyzed using GC-MS (see Section 4.2.3).

Hydrosol Preparation
M. suaveolens and F. vulgare plants were subjected to 2 h HD extraction of aerial parts using a Clevenger-type apparatus as previously described [66]. EOs were separated from HSs and stored in tightly closed dark vials at −18 • C until further utilization in other studies. Extractions were performed according to the protocol of European Pharmacopeia. Fresh leaves (2 kg) of aerial parts from each plant species were used for distillation. HSs were separated from EOs using decantation, avoiding the carryover of EOs. The HS organic part was extracted twice with diethyl ether (Sigma-Aldrich, Milan, Italy) in a separation funnel to eliminate water and was stored at 4 • C in brown glass vials in the dark until further analysis or testing. The EO/diethyl ether phase was dried over anhydrous sodium sulfate (Sigma-Aldrich, Milan, Italy), and diethyl ether was then evaporated.

Gas Chromatography-Mass Spectrometry Analysis
GC-MS analyses of EOs and HSs were performed using an Agilent Technologies gas chromatograph (GC 7890A) coupled with a single quadrupole mass spectrometer (5975C Inert XL MSD) and an autosampler (CTC analytics PAL system). A 5MS (30 m × 0.25 mm × 0.25 µm + 10 m of pre-column; MEGA srl, Milan, Italy) column was used for sample chromatography, and ultra-pure helium (6.0 BIP; SAPIO srl, Monza, Italy) at a flow rate of 1 mL/min was used as the gas carrier. Injector, source and transfer line were settled at 280 • C, 280 • C and 250 • C, respectively.
Before injection, EOs were diluted to 1:10 v/v with n-hexane (97% purity; Sigma Aldrich, Milan, Italy), while HSs were previously liquid-liquid-extracted in n-hexane at 1:1 v/v. The organic fraction was separated, dried with anhydrous sodium sulfate, concentrated and directly injected. EOs were injected in split mode (split ratio of 1:50) at 50 mL/min split flow, whereas HSs were injected in splitless mode.
Sample separation was achieved using the following temperature ramp: 5 min at 60 • C; from 60 • C to 220 • C at a rate of 4 • C/min; from 220 • C to 280 • C at a rate of 11 • C/min; isocratic for 15 min 280 • C; then, from 280 • C to 300 • C at a rate of 11 • C/min.
Mass spectra were recorded in electronic impact (EI) mode at 70 eV. The analysis was conducted in full scan mode, from 30 to 450 m/z, with a solvent delay of 5 min. The retention index (RI) was calculated using a separately injected n-alkane standard mixture (C8-C30; Sigma Aldrich, Milan, Italy).
The samples were injected three times, and the obtained chromatograms were aligned and deconvoluted using the open-source software MS-DIAL 4.8 [67]. The area of each compound was extracted and mediated. Peak annotation was achieved using the RI and spectral similarity matching with a cosine score cut-off of 70% using an in-house EI spectral library [68], following Metabolomics Standards Initiative of the International Metabolomics Society. In particular, as suggested by [69], the annotations were considered at level 2 (putative annotation based on spectral library similarity) or level 3 (putatively characterized compound class based on spectral similarity to known compounds of a chemical class). Moreover, the relative quantitation of these compounds was also calculated as the mean of the relative percentage for each peak (peak area/total ion current (TIC) area) over the three replicate injections (Table 1 and Table S1). The ZYMV inoculum was mixed with EO or HS solution in 0.1 M pH 7.4 phosphate buffer to obtain final concentrations of 1:10 v/v virus inoculum, 300 µg/mL EO or 1:2 v/v HS and incubated on ice for 1 h. Then, host plants in the developmental stage of fully expanded cotyledons were mechanically inoculated with the above-described solution of inoculum + treatment, and 20 µL was smeared on each cotyledon. Each treatment was applied on three biological replicates (i.e., plants), and the following controls were also included: (i) ZYMV-infected treatment with ribavirin as positive control (final concentrations: 1:10 v/v virus inoculum and 300 µg/mL ribavirin); (ii) ZYMV-infected non-treatment as negative control (final concentration: 10:10 v/v virus inoculum); and (iii) healthy control (mock-inoculated with phosphate buffer). Treatments were only applied once. When the first leaf was expanded, 7 d.p.i., plants were sampled by removing a disk from the above-mentioned leaf. More sampling was performed with the same procedure 14, 21 and 28 d.p.i. on the second, third and fourth expanded leaves, respectively.
Biological replicates (3 plants per treatment) were pooled, and the pools were analyzed as a single sample. Quantification of virus titer and expression of the PAL gene using real-time RT-PCR were performed as described below. All the experimental procedures reported in the present and the following sections are depicted as a flow diagram in Supplementary Figure S1.

Treatments Applied after Inoculation
Host plants were inoculated under the same conditions described in the previous section but with no treatment added to the inoculum (final concentration: 1:10 v/v virus inoculum). After 5 h, treatments of EOs or HSs were applied to the inoculated cotyledons under the same conditions described in the previous section (final concentration: 300 µg/mL EO or 1:2 v/v HS). Application was performed by smearing inoculated leaves with EO or HS solution; treatments were only applied once. Biological replicates, control treatments and sampling procedures were the same as those described above.

RNA Extraction and Real-Time RT-PCR
Total RNA was extracted from collected samples using RNeasy Plant Mini Kit (Qiagen, Milan, Italy) according to the manufacturer's instructions. Extracts were checked for purity and concentration with a NanoDrop™ spectrophotometer (ThermoFisher Scientific, Milan, Italy). Up to 10 µg of RNA was exposed to 2 U TURBO DNase™ (TURBO DNA-free™ kit; Life Technologies, Milan, Italy) in 10X TURBO DNase™ Buffer (total reaction volume of 50 µL) at 37 • C for 25 min. Then, 5 µL of DNase Inactivation Reagent from the kit was added; the mix was incubated at room temperature for 5 min to stop the reaction and centrifuged at 10,000 g for 90 s; then, the supernatant was recovered as DNA-depleted RNA for downstream analyses.
TaqMan ® real-time RT-PCR assay was used for the relative quantification of ZYMV using primers and probe targeting the ZYMV coat protein (CP) gene [70]. The amplification reaction had a final volume of 20 µL, containing 2X TaqMan ® RT-PCR Master Mix, 40X TaqMan ® RT Enzyme Mix (TaqMan ® RNA-to-C T 1-Step Kit; Life Technologies), 300 nM of each primer, 50 nM probe, and 1 µL of DNA-depleted RNA.
The relative expression of PAL was analyzed with SYBR Green ® real-time PCR assay; first, RNA was reverse-transcribed to cDNA to convert plant transcriptome into DNA substrate for real-time PCR. The reaction had a final volume of 20 µL, containing 5X firststrand buffer (Invitrogen, Milan, Italy), 5 µM random hexamers (Promega, San Diego, CA, USA), 10 µM dNTPs (Promega), 100 U M-MLV (Promega) and 2 µL of RNA. The reaction was incubated for 45 min at 42 • C and 3 min at 94 • C in CFX96 Touch PCR System (Bio-Rad, Milan, Italy); then, 1 µL of RNase cocktail mix (Life Technologies) was added to remove traces of unreacted RNA. SYBR Green ® real-time PCR assay for PAL expression was performed using primers designed by Zhang et al. [71]. The reaction had a final volume of 10 µL, containing 2X SsoAdvanced™ Universal SYBR Green ® Supermix (Bio-Rad, Milan, Italy), 150 nM of each primer and 1 µL of cDNA template. All real-time (RT-)PCR assays were performed using CFX96 Touch RT-PCR System (Bio-Rad, Milan, Italy) with primers and probes synthesized by Eurofins Genomics (Ebersberg, Germany). The instrument automatically set the threshold. The virus relative titer and PAL expression levels were calculated using the method of ∆∆Ct [72]. ZYMV-infected plants not exposed to treatments as described in Section 4.4.1 were considered the control group; the housekeeping gene was the elongation factor EF-1α gene of C. pepo, targeted with previously published primers (for SYBR Green ® assay) and probe (for TaqMan ® assay) [73]. All samples represented a pool of three biological replicates (i.e., plants subjected to the same treatment in the same pot) and were assayed in two technical replicates. Relative virus titers and PAL expression levels, defined as 2 −∆∆Ct , were calculated with CFX Maestro Software ver. 2.2 (Bio-Rad, Milan, Italy), and the results were expressed as means ± standard error (SE).
Choice tests to investigate potential EO and HS repellency effects were carried out in 15 cm Ø Petri dishes, filled with moistened filter paper. In each dish, a zucchini leaf sprayed with EO or HS using a hand sprayer and a leaf of the same age and size sprayed with deionized water were placed onto filter paper with their lower surface facing upwards. The two leaves were separated with wet cotton to avoid possible interference between them, and the left/right position of the two leaves within the dish was inverted among the experimental replicates. An area of the dish equally far from the two leaves was kept free to host the aphids. Ten apterous adults of A. gossypii were transferred from the rearing cage to the free area of each dish using a fine brush and were allowed to spread and start sucking on the leaves. Petri dishes were then sealed with parafilm. The number of aphids that settled on each leaf was recorded 1, 2, 4 and 24 h after treatment. Twenty replicates were performed for each EO or HS treatment. Repellency was calculated based on the counts of adults on untreated and treated leaves using the following equation [62]: Toxicity and fecundity assays were carried out in 9 cm Ø Petri dishes filled with moistened filter paper. A C. pepo leaf was placed on filter paper with the lower surface facing upwards, and ten apterous adults of A. gossypii were transferred onto it. The leaf was sprayed with EO or HS before and after aphid transfer. Leaves of the same age and size sprayed with deionized water provided the control treatment. The filter paper was moistened from time to time to maintain leaf turgidity, and Petri dishes were sealed with parafilm. The mortality rate of the aphids was recorded 24, 48, 72 and 96 h after treatment. Aphids were considered dead when they did not respond to gentle prodding with a brush. During each data collection instance, aphid offspring were also counted and removed. Twenty replicates were performed per EO or HS treatment and water treatment. Data from test choices, toxicity and fecundity assays were analyzed with Chi-square test to assess statistically significant differences between each treatment and the untreated control. For the test choice (repellency) assay, pairwise Chi-square test comparisons among treatments were also performed.

Conclusions
The composition of EOs was very complex, due to the presence of tens of molecules belonging to monoterpene hydrocarbons and oxygenated monoterpenes. On the other hand, in the HS organic phase, a main compound was recognizable.
The results of experiments involving both plant treatments and insect assays indicated good performance in reducing virus titer and repelling the aphid vector. M. suaveolens HS also showed toxicity and offspring inhibition characteristics against A. gossypii. The measured PAL expression levels in treated plants also suggested a mechanism of action based on the stimulation of plant defense response through the phenylpropanoid pathway. In the frame of envisaged integrated pest management to reduce the use of synthetic pesticides, these treatments represent potential biopesticides for the concurrent control of ZYMV and its vector in C. pepo crops with a single substance. Further studies are necessary to achieve a formulation applicable for practical use in the greenhouse and possibly in the field.