Several factors can restrict the production yield in agricultural species and among them, weeds are a tough competitor, which usually can be controlled using different methods, such as mechanical, physical, biological and chemical (Silva and Silva, 2012). Combining different methods that reduce weed development and its effects on agricultural species is seen as a promising strategy (Nunes et al., 2010). However, weed management in recent decades has been accomplished mostly through chemical control (Oliveira et al., 2011), leading to the selection of resistant biotypes (Christoffoleti et al., 1997).

The high number of weed species resistant to herbicides calls for new modes of action and herbicide classes. However, an herbicide with a new mode of action has not been launched on the market in the last 20 years (Dayan and Duke, 2014). In this context, biological molecules are considered important sources to be explored for the development of new products for weed management (Duke et al., 2002). According to Duke et al. (2000) there are many toxins of natural origin still unexplored and which have herbicide action with different modes of action from those currently known. Fungi are one option and often are found in nature colonizing and causing severe damage in some species of weeds (Barreto and Evans, 1998). This has led to an increased number of studies using these microorganisms and their metabolites as weed control agents, as well as the discovery of new sites of action and bioactive molecules (Dayan and Duke, 2014). Several studies have shown herbicidal activity of metabolites produced by microorganisms on weeds. Varejão and Demuner (2013) identified compounds produced by Alternaria euphorbiicola showing herbicidal activity on Euphorbia heterophylla. Khattak et al. (2014) found metabolites of Aspergillus spp. and Penicillium spp. negatively interfering in the growth of Lemna minor and in the seed germination of Silybum marianum.

In a previous study, our research group collected thirty-nine phytopathogenic fungi from infected tissues of weeds and isolated them in the laboratory looking for the production of metabolites with herbicidal action. Among them, a strain belonging to the Diaporthe sp. genus was screened due to herbicidal effects of its secondary metabolites on target plants (Souza et al., 2016a). In a second study, the production of the bioherbicide in a bench-scale bioreactor with working volume of 3 L was assessed (Souza et al., 2015b). The bioherbicide efficiency rate reached 100% in pre-emergence, since there was no seed germination after its application. Diaporthe has been reported often as producing secondary metabolites for biological control of weeds (Ash et al., 2010; Andolfi et al., 2015). Phomentrioloxin B caused small necrotic spots on a number of plant species, whereas gulypyrone A caused leaf necrosis in Helianthus annuus plantlets (Andolfi et al., 2015). Cimmino et al. (2013) tested several compounds produced in liquid culture by Phomopsis sp. (teleomorph: Diaporthe gulyae) for the control of the annual weed Carthamus lanatus.

The main objective this study was to evaluate the application of fermented broth containing secondary metabolites of Diaporthe sp. to control Lolium multiflorum, Conyza sp. and Echinochloa sp. The herbicidal activity was evaluated in post-emergence and pre-emergence of these weed species and also on Glycine max, Triticum aestivum and Oryza sativa.

Weeds and cultures used in the bioassays

Two bioassays were carried out to evaluate the herbicidal effect of secondary metabolites of Diaporthe sp. obtained by submerged fermentation in the pre-emergence of soybean (Glycine max), wheat (Triticum aestivum), rice (O. sativa), rye grass (L. multiflorum) and sorghum (Sorghum halepense) and in the post-emergence of soybean (Glycine max), wheat (T. aestivum), rice (Oryza sativa), rye grass (L. multiflorum), horseweed (Conyza sp.) and rice grass (Echinochloa sp.).

Location of bioassays

Pre-emergence assays were carried out in an incubator with controlled humidity (85%) and temperature (28 °C), while the post-emergence evaluation was carried out in a greenhouse, located at the coordinates 29° 43’39.79’’S and 53°33’40.06’’W, Santa Maria – Brazil.

Production of metabolites by submerged fermentation

The strain of Diaporthe sp. was isolated from plants of the weed Solanum americanum Mill in the Pampa biome and was used in this study for bioherbicide production by submerged fermentation. The strain of Diaporthe sp. was kept in a potato dextrose agar (PDA) at 4°C and subcultured every 15 days. Cell production for pre-inoculum was incubated in a Petri dish containing PDA for 8 days at 28ºC (Souza et al., 2016a). Afterwards, a 6 mm disc of fungal mycelium was transferred to a 250 mL Erlenmeyer flask, which contained 100 mL of fermentation medium at 28°C at 120 rpm for 7 days (Inova 44R, NewBrunswick) for inoculum. The fermentations were carried out in a batch bioreactor (BIOTEC-C, Tecnal, Brazil), containing 3.0 L of the optimized culture medium, using a 10% (v/v) inoculum at an initial pH of 6.0, incubated at 28°C for 7 days at 40 rpm. The culture medium was composed of corn steep liquor 33% (v/v), sucrose 18 g.L^-1, (NH₄)₂SO₄ 2 g.L^-1, MgSO₄.7H₂O 0.5 g.L^-1, FeSO₄.7H₂0 1 g.L^-1 and MnSO₄.H₂O 1 g.L^-1. After fermentation, the cell mass was separated from the fermentation broth by centrifugation at 4000 rpm for 10 minutes and the supernatant used in the bioassays.

Herbicidal activity in pre-emergence

The activity of the bioherbicide obtained from Diaporthe sp. was evaluated in the pre-emergence of dicotyledon seeds (G. max) and monocotyledon seeds (T. aestivum, O. sativa, L. multiflorum and S. halepense). For this purpose, 10 mL of fermented broth without the cells were sprayed on a germitest paper containing 100 seeds of each culture and maintained at 28°C (POL-EKO, model KK) 350. A control test was carried out by replacing the fermented broth with distilled water. Seed germination was evaluated during 7 days, according to the Brazilian rules for seed analysis (Brasil, 2009). Afterwards, the germinated and non-germinated seeds were counted. All seeds that presented primary root protrusion higher than 2 mm were considered to be germinated.

Herbicidal activity in post-emergence

For the evaluation of herbicidal activity in post-emergence, seven different concentrations of fermented broth were sprayed on the upper part of G. max, T. aestivum, O. sativa, L. multiflorum, Conyza sp. and Echinochloa sp. plants. The concentrations were 0D (water), 1/4D, 1/2D, 1D, 2D, 4D and 8D, where D corresponds to undiluted fermented broth. A mineral oil 0.5% (v/v) was added to all treatments.

The experimental design was completely randomized with five replicates. The units were 500 ml polyethylene pots, filled with commercial substrate Macplant^®. Four seeds of each species evaluated were manually sown in each pot. After plant emergence only one plant was maintained per pot. For the fermented broth sprays, a backpack sprayer was used, which was pressurized by CO₂, provided with a bar pattern with four tips, model Teejet XR 110.02, with a pressure of 40 Ibf and spacing between tips of 0.5 m. The walking speed during spray was 1 m s^-1 and a volume of 200 L ha^-1 was used. At spray time, the plants had between 2-3 leaves and the temperature and relative air humidity were 31.2°C and 62%, respectively.

In each treatment, shoot dry matter, weed control and injury according to Frans et al. (1986) were evaluated. The injury for the species G. max, T. aestivum and O. sativa was determined at 7 and 15 days after the fermented broth spray (DAA). The control of L. multiflorum, Conyza sp. and Echinochloa sp. was evaluated at 15 DAA. After the determination of control and injury at 15 DAA, the aerial part of plants was sectioned, packed in paper bags and dried at 70°C during 72 h to obtain shoot dry matter.

Statistical analysis

All data were submitted to analysis of variance and Tukey’s test was used to verify significant differences between sprays (p<0.05) using the software Assistat^® 7.7 beta.

Herbicidal activity in pre-emergence

Table 1 presents the herbicidal activity in pre-emergence of fermented broth. Seeds of G. max, T. aestivum, O. sativa, L. multiflorum and S. halepense did not present root protrusion. The germination inhibition reached 100% for all species, whereas in the control test, germination varied by species, but reached values higher than 60% (Figures 1 and 2).

Herbicidal activity in post-emergence

The effect of fermented broth on shoot dry matter mainly for species of Conyza sp. and Echinochloa sp was exhibited in Figure 3. The best results for Conyza sp. were obtained when 4D and 8D were applied, however, plant death did not occur.

The effect of the bioherbicide on injury of crop species shown in Table 2. Only G. max presented injury symptoms, which was verified at 7 and 14 days after application (DAA) for concentrations higher than 2D and most severely for 8D. Plants showed symptoms, such as yellowing and blight on youngest leaves, whereas the oldest leaves showed partial necrosis at the tips, as presented in Figure 4.

Herbicidal activity in pre-emergence

These results confirm that metabolites of Diaporthe sp. have herbicidal activity in pre-emergence of some weed species, providing an important alternative for the management of weeds resistant to chemical herbicides. The pre-emergent control has the advantage of reducing competition from weeds with the crop especially in the initial phase of its establishment as a result of lower germination of weeds, besides enabling a reduction in the seed bank.

The metabolites of Diaporthe sp. showed a broad spectrum of action in pre-emergence, since activity was verified in both dicotyledonous (G. max) and monocotyledonous seeds (T. aestivum, O. sativa, L. multiflorum and S. halepense). Souza et al. (2015b) obtained an efficiency of 100% in the pre-emergence of Cumumis sativus (dicotyledonous) and Sorghum sp. (monocotyledonous) seeds when fermented broth of Diaporthe sp. was used. The control of different plant families, as well as different species, is considered beneficial from an agricultural viewpoint, since simultaneous growth of different species of weeds is common in crops and an herbicide should have the broadest action possible.

The inhibition of germination by the action of secondary metabolites of Diaporthe sp. may be related to molecules from the terpenoid groups which are produced by a wide range of fungi. Castro et al. (2001) reported that terpenoids have hormonal functions, inhibiting the growth and acting in active-transport through the membrane. However, the confirmation of this hypothesis should be based on more in-depth studies.

Herbicidal activity in post-emergence

In practice, the increased concentrations increased the amount of molecules with herbicide activity, which are very diluted in the fermentation medium (Varejão and Demuner, 2013). As a consequence, a suppressive effect was verified instead of control, as shown in Figure 5 for Conyza sp. For this weed, 4D and 8D resulted in about 50% reduction in weight of shoot dry matter, when compared with the control treatment. Even though significantly less, it can be seen that the fermented broth caused a reduction in the weight of shoot dry matter for Echinochloa sp. at all doses used. The 8D concentration of fermented broth also resulted in a significant reduction of weight of shoot dry matter for the G. max species. This confirms that phytotoxins are highly diluted in the fermentation broth, so that an effect only occurs when it is used at higher concentrations.

In addition, the plants presented significant shoot dry matter reduction as a result of injury caused by the application of fermentation broth. The observed symptoms suggest that phytotoxins present systemic action, as both G. max and Conyza sp. plants presented a retardation of plant growth. One hypothesis related to the retarded development of plants and consequent reduction of shoot dry matter may be related to interference of the phytotoxin in the biosynthesis of amino acids that are essential in plant growth, such as valine, leucine and isoleucine.

In this study, the application of fermented broth of Diaporthe sp. as a bio-herbicide in pre- and post-emergence of different crop species (O. sativa, G. max and T. aestivum) and weeds (Echinochloa sp., Conyza sp. and L. multiflorum) was investigated. In pre-emergence, the fermented broth showed an efficiency of 100%, since no seeds of any species germinated. In post-emergence, the most promising results were obtained for Conyza sp. and Echinochloa sp. species, because plant growth was suppressed. G. max showed injury to metabolites when concentrations were higher than 2D at 7 and 15 DAA. The biomolecules produced by Diaporthe sp. are promising for as a natural herbicide.

The authors have not declared any conflict of interests.

The authors thank CAPES and CNPq for the scholarships and the State Department of Development and Investment Promotion (SDPI-RS) for the financial support of this work.