Agricultural production has always been threatened by the presence of plant pathogens such as fungi, bacteria, and viruses (Kordali et al., 2016). Moreover, weeds are another major issue. In fact, they com­pete with crops for resources (water, nutrients, light) and cause huge economic losses, up to 34% in major crops (Araniti et al., 2015). Every year, approximately 2.5 million tons of pesticides are used on crops worldwide to fight plant diseases (Koul et al., 2008) with consequences on human health, soils and the environment (groundwater contamination and development of weed resistance). This intensive use has been recognized as one of the main drivers of biodiversity losses (Schütte et al., 2017).

In the last few decades, there has been growing interest in investigating eco-friendly alternatives, in particular essential oil (EO)-based methods in order to curtail pesticide use because pesticides cause extensi­ve damage to agricultural and natural systems (Ben Ghnaya et al., 2013). Moreover, the use of EOs obtai­ned through a cheap production process can reduce the frequent applications of certain synthetic pestici­des that have deleterious effects on the environment and human health (Pavela & Benelli, 2016). But EOs can also present a low health risk during application. One of the great challenges for further research is to design an efficient stabilization process so as to apply EOs in fields. In the same vein, several studies have pointed out that EOs may have not only antifungal activity (Koul et al., 2008; Tian et al., 2012; Ahluwalia et al., 2014; Hmiri et al., 2015; Boubaker et al., 2016) but also the ability to inhibit weed seedling growth (Uremis et al., 2009; Poonpaiboonpipat et al., 2013). In the Mediterranean region and especially in Tunisia, the most widespread botanical family is Lamiaceae, which has antimicrobial properties (Pintore et al., 2002). Among these aromatic plants, the most interesting species is Rosmarinus officinalis ( R. officinalis) which is known for its antifungal activity (Angioni et al., 2004; Giamperi et al., 2011; Hmiri et al., 2015) and its richness in EOs characterized by the predominance of monoterpenes - mostly 1,8 cineole -, camphor, and α-pinene (Zaouali et al., 2010). Hence, the main aims of this study were (1) to assess the antifungal activity of R. officinalis EO against three potential plant-pathogenic fungi, (2) to evaluate its herbicidal activity on three weed species for the first time, and then (3) to formulate a bioherbicide in order to enhance its efficiency and stability.

Our results show that R. officinalis EO is an inte­resting antifungal and herbicidal agent from which a more environment-friendly alternative to chemical herbicides might be derived. The antifungal and her­bicidal activities of EOs have been widely reported in recent years (Pintore et al., 2002; Salamci et al., 2007 Tian et al., 2012; Amri et al., 2012; Kaur et al., 2012; Ben Ghnaya et al., 2013; Ahluwalia et al., 2014; Bouabidi et al., 2015; Hmiri et al., 2015; Ali­­pour & Saharkhiz, 2016; Synowiec et al., 2017), but to our knowledge, only a few studies have focused on their effect on post emergence when sprayed on weeds (Hazrati et al., 2017). R. officinalis is largely used in traditional medicine (Pintore et al., 2002; Ben Jemia et al., 2015) and widely known for its antimicrobial and antioxidant activities (Bozin et al., 2007; Celiktas et al., 2007; Zaouali et al., 2010), but the present stu­­­dy unveils its herbicidal effect in pre-emergence and post-emergence for the first time. On the other hand, GC-MS analysis of our R. officinalis EO extracted from dried leaves and flowers identified 19 compounds dominated by oxygenated monoterpenes including 1,8 cineole, camphor, and borneol. These results are in agreement with Zaouali et al. (2010), who showed that these three major components are also predominant in the Tunisian R. officinalis EO. However, their per­centages varied between 26.0-51.2%, 4.9-29.7% and 3.3-10%, respectively. These differences in chemical composition could be related to environmental fac­­­tors (the climate, the season, the soil), the genetic diversity of the species, and the geographic conditi­ons (Ben Ghnaya et al., 2013). Interestingly, the monoterpenes identified as main constituents in our EO have been described as powerful inhibitors of the seed germination and growth of several plant species (De Martino et al., 2010; Barton et al., 2014). These compounds also showed antifungal activity (Ben Ghnaya et al., 2013; Ahluwalia et al., 2014; Marei & Abdegaleil, 2018).

In addition, EOs from plants of the Lamiaceae family, and among them R. officinalis, are known for their antimicrobial activity (Hendel et al., 2016). In our study, R. officinalis significantly inhibited the spore germination of P. italicum, F. oxysporum, and F. culmorum. F. oxysporum and F. culmorum have been widely documented as the most important plant pests; they cause substantial economic losses worldwide (Hollingsworth & Motteberg, 2008). R. officinalis EO from Greece caused a dose-dependent inhibition of the mycelial growth of five fungi ( Sclerotinia sclerotiorum, Phytophthora nicotianae, Sclerotium cepivorum, F. oxysporum, and Fusarium proliferatum) (Pitarokili et al., 2002). In addition, Sardinian R. officinalis EO (450 and 900 µL/mL) showed a weak activity against all tested fungi ( Botrytis cinerea, F. oxysporum lycopersici, Fusarium graminearum, F. culmorum, and Rhizoctonia solani). On the other hand, these EOs present multiple mechanisms of action due to a large number of active compounds that reduces the development of fungal resistance. For instance, a recent study confirmed that 1,8 cineole alone had a low antifungal power but showed an important synergistic effect with α-pinene (Hmiri et al., 2015). These two compounds were identified in our EO. In the same li­ne, other reports suggested that 1,8 cineole combined with terpinen-4-ol, also the major component of Me­laleuca alternifolia EO, had a significant synergistic effect on the hyphal morphology of B. cinerea and its ultrastructure as compared to the treatment using either component alone. In fact, 1,8 cineole can penetrate the cell and damage cellular organelles without affecting membrane permeability. On the other hand, terpinen-4-ol destroys membrane integrity and increases per­meability, resulting in ion leakage and membrane dys­functioning. Several studies reported that EOs could cause structural and functional damage by disrupting the membrane permeability and the osmotic balance of the cell (Yu et al., 2015). Other studies have shown that they can acidify the external medium and decrease ATPase and dehydrogenase activities in Aspergillus flavus cells (Tian et al., 2012). Furthermore, EO from seeds of Anethum graveolens showed fungicidal ac­tivity against Sclerotinia sclerotiorum by inhibiting mycelial growth and sclerotial germination. This effect is the consequence of the inhibition of ergosterol synthesis, malate dehydrogenase, and succinate dehy­drogenase (Ma et al., 2015).

In parallel, to our knowledge, no study had yet tackled the herbicidal activity of R. officinalis EO. In fact, our experiments highlighted the outstanding inhibition of three different weeds after treatment with our EO. This was seen on the percentage of ger­mination, root growth, and hypocotyl length. In fact, 100% inhibition of germination and seedling growth was observed with our EO at 5 mM. In this context, Poonpaiboonpipat et al. (2013) showed that at 1 µL and 2 µL/Petri dish of Cymbopogon citratus EO, there was no significant effect on shoot or root length, but seedling length was shorter at 4 and 8 µL/Petri dish. The strong phytotoxic activity was due to the pre­sence of oxygenated monoterpenes, which is quite similar to that of Tunisian Eucalyptus erthrocorys EO, renowned for its overwhelming phytotoxic effect (Ben Ghnaya et al., 2013). In this context, among 12 EOs tested on weeds, caraway, thyme, peppermint, and sage oils were classified as the most phytotoxic ones owing to the existence of oxygenated monoterpenes in a 64.1-93.3% range (Synowiec et al., 2017). In line with this, among six monoterpenes tested by Gouda et al. (2016), 1,8 cineole and (S)-limonene were showed to inhibit Echinochloa crus-galli shoot growth. The major components of EOs are very important for their biological activity, but even the minor ones could have significant synergistic effects (Synowiec et al., 2017). Many other individual compounds identified in R. officinalis, such as α-terpineol, citronellal, citronellol, and α-pinene, have been confirmed to have phytotoxic activity (Zhang et al., 2014). In contrast, among 25 EOs, only those containing volatile pheno­lic compounds such as thymol, carvacrol, eugenol, alcohols or ketones, showed strong phytotoxic effect on different weed seeds, even though the mode of action of all these compounds has not yet been detailed and a number of effects and hypotheses have been reported by many authors. Several authors assume that EOs act by causing biochemical and physiologi­cal changes in seedling growth (De Martino et al., 2010). For instance, Cymbopogon citrates EOs notably slowed down α-amylase activity in E. crus-galli seeds (Poonpaiboonpipat et al., 2013). Another clear exam­­­ple is Artemisia sp. EO: it induced reactive oxygen species production, which in turn caused damage re­sulting in lipid peroxidation, decreased membrane flu­idity, and finally increased membrane leakiness and inactivated receptors, enzymes and ion channels (Kaur et al., 2012). Moreover, 1,8 cineole inhibited root growth and stopped DNA synthesis through several steps (Koitabashi et al., 1997).

We applied R. officinalis EO not only in pre-emer­gence tests but also for the first time in post-emergence tests, by spraying it on weeds under greenhouse conditions. Based on the visual damage induced by this EO on weeds three days after spraying, herbicidal properties were noticed. Necrosis and wilting leaves were observed at a concentration of R. officinalis EO starting from 20 mM. Similar results showed that the spraying of Cymbopogon citratus EO from 1.25 mM to 10 mM on E. crus-galli leaves caused wilting, and the leaves exhibited desiccation symptoms. In addition, this EO decreased the chlorophyll a, b and carotenoid contents, and caused electrolyte leakage, indicating membrane disruption and loss of integrity (Poonpaiboonpipat et al., 2013). Monoterpenes, which are present at 80.19% in our R. officinalis EO, may affect plant photosynthesis, energy metabolism, and the biosynthesis of secondary metabolites such as phenolic compounds (Gouda et al., 2016). In addition, it has been confirmed that the penetration of monoterpenes through the cell wall and cell membrane can cause cellular potassium leakage that inhibits glucose-dependent respiration. A recent study showed that the spraying of a nano-emulsion of Satureja hortensis EO reduced the weed chlorophyll content, and increased electrolyte leakage and cell membrane disruption (Hazrati et al., 2017).

We investigated a formulation of R. officinalis EO as a bioherbicide for the first time, based on the following observations: (1) as R. officinalis EO is lipophilic, it does not dissolve well in water; (2) in the same line, the reported herbicidal effect of Satureja hortensis EO in the absence of tween 20 was lower on control weeds; and (3) EOs contain terpenoids that are volatile, thermolabile, and may be easily oxidized and hydrolyzed (Pavela et al., 2016). For these reasons, we used an emulsifier providing better stability, effi­­cacy and persistence for the for­mulation. An ionic surfactant reduced the effective concentration of eucalypt oil for a high herbicidal activity against P. minor (Batish et al., 2007). Based on that, a recent study showed that a formulation containing palm oil, tween 20 and span 80 improved the herbicidal activity of metabolites from Phoma sp. (Todero et al., 2018).

To our knowledge, this is the first report that links the chemical composition of Tunisian R. officinalis EO to its fungicidal and bio-herbicidal effects on plant pathogens and weeds, respectively. Moreover, the formulation of the bio-herbicide based on Tunisian R. officinalis EO was attempted in this work for the first time. Hence, this work opens new perspectives on the application of Tunisian R. officinalis EOs as a novel biocontrol strategy against harmful plant pathogens and weeds. It also paves the way for new strategies and pathways for the biopesticide industry to create alternative chemical pesticides designed to be less harmful to the environment and human health than current ones. For agronomic applications, we found that R. officinalis EO could be used as a biofungicide at low concentrations between 1 mM and 6 mM with­out any phytotoxic effect in post-emergence tests. At concentrations higher than 20 mM, this EO can be used as a post-emergence bioherbicide. According to our preliminary results, the use of EOs in the formu­lation of bioherbicides can offer new prospects for the sustainable production and practical use of EOs. To go further in the experiments, it could be really interesting to determine the modes of action of R. officinalis EO on weeds and fungi and try to improve the effectiveness and stability of the bioherbicidal R. officinalis EO formulation.