Screening of Tunisian plant extracts for herbicidal activity and formulation

Abstract The need to reduce the use of chemical pesticides currently fosters great interest in eco-friendly biological control agents. In addition, the isolation of plant allelopathic substances and the evaluation of their phytotoxic effects can lead to the discovery of new natural herbicides. In this context, our study aimed to assess the herbicidal activity of ten crude extracts obtained from aerial parts of Tunisian spontaneous plants against Trifolium incarnatum, Silybum marianum and Phalaris minor. It confirmed that the Cynara cardunculus methanolic extract best inhibited weed germination and seedling growth, and caused necrosis or chlorosis. Following a bioassay-guided fractionation, five main phenolic compounds, (1) syringic acid, (2) p-coumaric acid, (3) myricitrin, (4) quercetin and (5) naringenin were identified in the most active crude methanolic extract. Then, only 3 of the flavonoids contained in the most active fraction were tested on Trifolium incarnatum. The 3 compounds had a significant phytotoxic effect and therefore could be employed in a new composition of botanical herbicides to control crop weeds. Besides, a novel herbicide composition was designed to improve the post-emergence activity of the methanolic extract. The formulation containing the C. cardunculus crude methanolic extract showed the same herbicidal activity as the standard industrial bioherbicide containing pelargonic acid. These results make C. cardunculus a suitable source of natural compounds potentially usable as natural herbicides.


35
Weeds are registered as harmful plant pests posing a serious problem in agriculture 36 worldwide (Suksungworn et al., 2016). They cause huge economic losses which can rise up to 37 34% in major crops by affecting yields and competing with crops for nutrients, light, and water 38 (Araniti et al., 2015;Jabran et al., 2015). 39 The most efficient weed control methods currently include mechanical or hand weeding, 40 and application of chemical herbicides. The latter has been proven to have negative impacts on 41 environmental, animal, and human health (Böcker et al, 2019). Besides, it can increase weed 42 resistance to phytochemicals (Jabran et al., 2015;Ahmed et al., 2017). For these reasons, 43 scientists are working on the identification of a biological solution that can minimize the 44 impacts of synthetic herbicides in agricultural production (Morra et al., 2018;Sbai et al., 2016;45 Chengxu et al., 2011). This solution could offer a number of benefits such as increased target 46 specificity and rapid degradation of the active substance (Cordeau et al., 2016).  Ribeiro et al. 54 (2015), Araniti et al. (2015), and Lim et al. (2017) showed that these plant extracts inhibited 55 weed germination and seedling growth. Nevertheless, few studies have shown an herbicidal 56 effect of these compounds in post emergence by direct spraying on weeds. This inhibitory effect 57 has often been related to the presence of phenolic compounds (Omezzine et al., 2011;Sbai et 58 al., 2016;Jelassi et al., 2016). These are the most represented secondary metabolites implied in varied between 5.29 and 29.71 %. 84 Methanol filtrates were combined, concentrated under vacuum, and fractionated using 85 a reverse-phase silica gel (Sigma-Aldrich, a subsidiary of Merck KGaA) to remove sugars and 86 other polar compounds. The samples were loaded on a column containing 10 g of C18 resin per 87 gram of dry extract, followed by 80 ml of water. The phenolic fractions adsorbed onto the resin 88 were then eluted with 30 ml of increasing MeOH percentages (successively 20, 40, 60, 80  Seeds of T. incarnatum S. marianum and P. minor were sterilised using 0.5 % sodium 119 hypochlorite for 2 min. Each crude plant extract was first solubilised in Tween 1% and then 120 diluted with distilled water to the desired concentration. Filter paper was moistened with 2 ml 121 of Tween 1 % solution (which did not interfere with the assays) as a control, or with the crude 122 methanolic extract solution prepared from different Tunisian plants at 5 g/L for the treatments.

123
Ten seeds of T. incarnatum, S. marianum or P. minor were then immediately placed in Petri 124 dishes, and three replicates (3 Petri dishes) were prepared for each extract and for each plant 125 species. All Petri dishes were randomly placed in a growth chamber at 23 ± 1°C, in darkness.

126
The number of germinated seedlings was counted, and their hypocotyl and root lengths were 127 measured after 7 days. The inhibition rate of the root and hypocotyl lengths was calculated (1) Inhibition rate (%) = * 100

130
Where T is the length of the roots or hypocotyls of the treated seedlings, and C represents the 131 length of the roots or hypocotyls of the control seedlings. its strong resistance to pesticides (Mccurdy et al., 2013).

175
The phytotoxic effect of the plant extracts was higher on P. minor than on T. incarnatum 176 and S. marianum. In fact, the majority of plant extracts including C. cardunculus and A. herba-177 alba completely inhibited P. minor seedling growth.

Effect of the plant extracts on post-emerged weeds under greenhouse conditions
179 Different methanolic extracts were sprayed on T. incarnatum, S. marianum, and P.
180 minor at 7.5, 20, or 34 g/L (Table 3). Only the C. cardunculus extract at 7.5 g/L and 20 g/L 181 caused some necrosis and chlorosis on the leaves of all three weeds. Its herbicidal activity 182 reached up to 37 %. At 34 g/L, several plant extracts had a phytotoxic effect against these 183 weeds. C. cardunculus seemed to have again the best herbicidal activity, which reached 184 62.76%. That was why it was selected for the identification of bioactive compounds and the 185 formulation of crude extracts and of the active fraction.

267
The mode of action of allelochemicals, including the flavonoids highlighted in our 268 study, still remains unknown (Soltys et al., 2013;Vyvyan, 2002). In fact, these secondary activity, phytohormonal activity, and ion uptake (Yan et al., 2014;Ribeiro et al., 2015;Yan et 272 al., 2015;Ahmed et al., 2017). They modify the expression of one or more genes, which then 273 leads to plant death (Cordeau et al., 2016). Chalcone, an aromatic precursor of the synthesis of 274 these flavonoids, induced programmed cell death in Arabidopsis thaliana roots (Tielas et al., 2013). In the same line, phenolic acids act as phytotoxic agents by inducing overproduction of 276 reactive oxygen species (ROS) that disturb respiration and photosynthesis (Lim et al., 2017;277 Ladhari et al., 2014;Franco et al., 2015;Araniti et al., 2014).  (Sakihama et al., 2002;de Martino et al., 292 2012; Ribeiro et al., 2015). Moreover, on a morphological scale, we noticed that root inhibition 293 in response to different methanolic plant extracts was related to the development of thin 294 secondary roots. This can be caused by oxidative stress (Franco et al., 2015), potentially due to 295 the interaction of phenolic compounds with auxin, cytokinin, and gibberellin transport (Ribeiro 296 et al., 2015). In accordance with our study, Franco et al. (2015) found that exogenous 297 flavonoids, such as those identified in our C. cardunculus extract, could delay primary root 298 growth and improve the expansion of lateral roots. They modified the expression model of 299 specific genes involved in root tissue differentiation.

301
Our results show that C. cardunculus is a promising plant with pre-emergence but also post-     1 Natural herbicide based on pelargonic acid was used as a positive control (at the same concentration in the market). 2 replicates were performed for each treatment. Value in a column followed by the same letter are not significantly different at P <0.05, as established by Tukey's test.