Microbial Prospection for Bioherbicide Production and Evaluation of Methodologies for Maximizing Phytotoxic Activity

: The occurrence of several weeds in crops of high economic value directly affects grain yield. This scenario led to the mass investigation of chemical products that circumvent these adversities and provide control potential. Nonetheless, the excessive application of chemical herbicides has generated signiﬁcant concerns about the environment. Accordingly, the adoption of alternative practices, such as the application of microbial metabolites, emerges as strategic control actions, having a sustain-ability bias and allowing the reduction of risks of human and animal contamination. Appropriately, this study proposed to conduct a microbial prospection of microorganisms capable of producing secondary metabolites to inhibit growth and generate phytotoxicity in weeds. Furthermore, to increase the herbicidal activity, different strategies were evaluated involving microbial co-cultivation and ultrasound-assisted extraction after fermentation. Accordingly, 63 microorganisms were isolated from weeds with disease symptoms and submitted separately to submerged fermentation. Initially, the bioherbicidal activity was evaluated in Cucumis sativus plants, and the most promising were applied in Amaranthus hybridus and Echinochloa crusgalli . Treatments with fermented broth obtained from co-cultivation and ultrasound-assisted extraction after the fermentation process indicated an inhibition of plant development. The most promising strains were A14.2 and B22.2 ( Nigrospora sphaerica ), B14 ( Bacillus velezensis ), and γ ( Aspergillus ﬂavus ).


Introduction
Weeds are one of the main problems in agriculture [1,2]. They have a high capacity to compete for resources such as micronutrients, light, and chemical and biological resources [3]. The weed species of the genera, Amaranthus sp., Conyza sp., Echinochloa sp., and Bidens sp., have been reported to cause damage to different important crops worldwide [4]. Bidens pilosa and Amaranthus sp. affect rice, corn, and soybean crops; Conyza sp. affects mainly soybean, cotton, and corn plantations [5]; and Echinochloa sp. is commonly found in rice plantations [6]. These four weeds mentioned have resistance already registered to chemical herbicides [5], especially Conyza sp., which is known as one of the most resistant to chemical herbicides [7][8][9]. The plant collection was performed from November 2019 to March 2020, and the samples were kept at 4 • C at the Bioprocess Laboratory of the Federal University of Santa Maria, Santa Maria, Brazil, for the isolation of microorganisms. The plants were identified based on morphological characteristics and the separation of each sample was performed with the identification of the plant species and its location. Each injured tissue was transferred to a Petri dish containing PDA culture medium and incubated in a Biochemical Oxygen Demand (B.O.D.) (ELETROlab ® , São Paulo, Brazil)) incubator with a photoperiod of 12 h for 7 days at 25 • C [45]. Successive subcultures were performed until a pure culture was obtained. After the isolation of all prospected microorganisms, different culture media were used for microorganisms of fungal or bacterial morphology. To fungi, the subcultures were made with a 6mm disk extracted from the cultured mycelium, and inoculated in a new PDA Petri dish, previously sterilized and prepared [7]. To bacteria, the isolations were performed by a simple streaking technique in a new Petri dish with 1.0 g L −1 of meat extract, 5.0 g L −1 of peptone, 2.0 g L −1 of yeast extract, 5.0 g L −1 of NaCl, and 15.0 g L −1 of agar, previously sterilized and prepared [46,47].

Submerged Fermentation
The microorganisms were separated according to their morphology, according to visual characteristics during growth in Petri dishes, and were submitted to different culture media for submerged fermentation. The liquid culture medium for microorganisms with  6.0 [26]. The liquid culture medium for the microorganisms that presented bacterial morphology (LCM-B) was prepared with 1.0 g L −1 of meat extract, 5.0 g L −1 of peptone, 2.0 g L −1 of yeast extract, 5.0 g L −1 of NaCl, and pH 7.4 [46]. Inoculation was performed by transferring two 6 mm discs to a 250 mL Erlenmeyer flask containing 125 mL of sterile culture medium. The Erlenmeyer flasks were incubated in an orbital shaker (New Brunswick, model Innova 44) for 7 days for microorganisms with fungal morphology, and 3 days for microorganisms with bacterial morphology, at 28 • C and 120 rpm. After fermentation, the cells were filtered (Whatman ® , Merck KGaA, Darmstadt, Germany) and separated by centrifugation at 4000 rpm for 20 min (Eppendorf, model 5804R), and the supernatant was subjected to membrane filtration (Merck Millipore TM), with a pore size of 0.45 µm and diameter of 47 mm [7]. The filtered supernatant was used in the bioassays. The supernatant was incubated to verify microbial growth, with negative results. The 63 microorganisms were considered distinct treatments for the tests, and the bioherbicidal activity was evaluated by a series of analyses.

Co-Cultivation Medium
The selection of microorganisms was performed initially by 10 combinations of pairs of microorganisms cultivated in Petri dishes to evaluate the viability of co-cultivation, and to verify possible unwanted inhibitions between strains. Each combination was composed of two microorganisms, and the number of combinations was determined through combinatorial analysis for simple combinations. The combinations were performed with the inoculation of a 6 mm disk extracted from the Petri dish culture of each microorganism in a new PDA Petri dish, previously sterilized and prepared. After selecting the best combinations, fermentation was carried out, and a specific culture medium was used for each morphology. For combinations between microorganisms that presented fungal morphology, the LCM-F was utilized [26]. Considering the combinations of microorganisms with bacterial morphology, the LCM-B was utilized [46]. Finally, for the combinations with fungi and bacteria, the addition of all the components of both media (LCM-F plus LCM-B) was performed [26,46]. The culture media were incubated and maintained under conditions as described above.

Post-Fermentation Ultrasound-Assisted Extraction (PF-UAE)
After 7 days of fermentation, the broth and biomass were directly subjected to Ultrasound-Assisted Extraction (UAE). An ultrasonic processor, UP400S (Hielscher Ultrasonics GmbH), Sonotrode H22 (Hielscher Ultrasonics GmbH) with a diameter of 22 mm was used. Sonication was performed directly by coupling a sonication cell with a cooling bath at 10 • C, and inside the cell, it was filled with the fermentation broth with the biomass. The process conditions were 90% amplitude and 0.5 s cycle for 40 min [48]. Afterward, the cells were separated from the medium, centrifuged at 4000 rpm for 20 min (Eppendorf, model 5804R), and filtered by membrane (Merck Millipore TM) with a pore size of 0.45 µm and a diameter of 47 mm to obtain a cell-free supernatant [7].

Co-Cultivation with PF-UAE
Assays were performed associating co-cultivation with the use of PF-UAE. The cocultivation was conducted as described in Section 2.4.1, and after the fermentation period, the broth and biomass were submitted to UAE as described in Section 2.4.2. (experimental unit). Sowing was performed in trays; after emergence, the seedlings were replanted in 1 L pots containing 50% soil and 50% Mecplant ® commercial substrate. The seeds were obtained from the local market, without any pre-sowing treatment. In this first stage, each application was performed with a manual sprayer, with the application of 1 mL per plant, directly on the shoot. For the application in the control plants, the broth was replaced with distilled water. The application was performed 9 days after sowing. The visual assessment of phytotoxicity was performed 21 days after application (DAA), based on the scale of Frans and Crowley (1986) [49].

Bioherbicidal Activity on Weeds
The best results obtained in the assays with C. sativus were selected, and the bioherbicidal activity in Amaranthus sp., Conyza sp., B. pilosa, and E. crusgalli were evaluated. The plants were sown and replanted into 1 L pots containing commercial Mecplant ® substrate and kept in a greenhouse. Applications were made with a manual sprayer.
To evaluate the bioherbicidal activity of the crude broth, 1 mL of supernatant per plant was applied, in 4 repetitions, where each plant was considered a repetition. For the control plants, the broth was replaced with distilled water [50]. The crude broth assay and control plants were kept under the same conditions, and the evaluations were performed at 21 DAA. The evaluation assays of the processes with PF-UAE, co-cultivation, and a triple application of broth, consisted only of the application in A. hybridus and E. crusgalli. One seedling of A. hybridus and 2 to 3 seedlings of E. crusgalli were replanted in pots; each pot was considered a plant, and each plant a repetition. For one of the methodologies, the triple crude broth was applied; that is, 3 mL per plant, and the others followed as previously described [25]. The applications were conducted in 7 repetitions, and the evaluations were carried out in 14 DAA. In all weed applications, plant height (PH), root length (RL), shoot fresh mass (SFM), root fresh mass (RFM), shoot dry mass (SDM), and root dry mass (RDM) were evaluated [51].

Identification of Microorganisms
The molecular identification of promising microorganisms was performed by the Biological Institute of São Paulo, São Paulo, Brazil. The DNA of the isolates was extracted according to the method described by [52] from the microorganism grown in a culture medium. The extracted genomic DNA sample was subjected to a polymerase chain reaction (PCR) for amplification of the ITS (internal transcribed spacer) region of the rDNA (isolates A14.2 and B22.2), beta-tubulin gene segment (γ isolate), and gyrB gene segment (B14 isolate). The oligonucleotide primers for the ITS region were SR6R (5 -AAGWAAAAGTCGTAACAAGG-3 ) and LR1 (5 -GGTTGGTTTCTTTTCCT-3 ) (Vilgalys and Hester, 1990); for the beta-tubulin gene, they were TUB2Fd (5 -GTBCACCTYCARACC GGYCARTG-3 ) and TUB4Rd (5 -CCRGAYTGRCCRAARACRAAGTTGTC-3 ) [53]; and for the gyrB gene, they were gyrB-F (5 -GTNYAYCGTGAYGGNAAAATYCA-3 ) and gyrB-R (5 -GCAGARTCWCCCTCTACRATATA-3 ) (developed by R. Harakava, Biological Institute of São Paulo). The PCR mix consisted of 1 µL of DNA, 1 µL of each primer at 10 µM, 10 µL of 5X PCR buffer, 1 µL of 10 mM dNTPs, 0.2 µL of GoTaq DNA polymerase 5U/µL (Promega), and 35.8 µL autoclaved MilliQ H 2 O, to a final volume of 50 µL. The amplification program consisted of initial denaturation at 94 • C/2 min followed by 40 cycles of denaturation at 94 • C/10 s, annealing at 54 • C/30 s, extension at 72 • C/45 s, and final extension at 72 • C/4 min. The verification of the amplified products was performed by electrophoresis in a 0.8% agarose gel stained with ethidium bromide. The amplified products were purified by precipitation with polyethylene glycol, directed to sequencing reaction by the chain termination method using the Big Dye 3.1 reagent (Applied Biosystems), and analyzed in a 3500 xL automatic capillary sequencer (Applied Biosystems) [54]. For the construction of the phylogenetic trees, the Neighbor-Joining method was used with a bootstrap of 1000 repetitions in the MEGA 6.0 program [55].

Statistical Performance
The means were compared by Student's t-test in the Statistica ® software, version 8.0 [56], using a probability of error lower than 5% (p ≤ 0.05) as a significant difference criterion.

Initial Bioprospecting
Considering the 63 different strains of microorganisms isolated, 21 microbial strains exhibited visual phytotoxic effects on C. sativus plants (Figure 1), such as necrosis, yellowing, and leaf spots. The others showed no visual effects on the plants. The effects were already observed in the first week of evaluation.

Statistical Performance
The means were compared by Student's t-test in the Statistica ® softwa [56], using a probability of error lower than 5% (p ≤ 0.05) as a significant di rion.

Initial Bioprospecting
Considering the 63 different strains of microorganisms isolated, 21 mi exhibited visual phytotoxic effects on C. sativus plants (Figure 1), such as ne ing, and leaf spots. The others showed no visual effects on the plants. Th already observed in the first week of evaluation. Five microorganisms presented significant results concerning phytot B22.2, strain γ, strain A14.2, strain B13, and strain B14, as they presented m least one of the plants in the stand, and phytotoxicity equal to or higher t other microorganisms showed less intense effects and plant recovery throu evaluation period. Table 2 presents the quantitative evaluation of six different varia namely: plant height (PH), root length (RL), shoot fresh mass (SFM), ro (RFM), shoot dry mass (SDM), and root dry mass (RDM).

PH (cm)
RL (cm) SFM (g) SDM (g) RFM (g) Five microorganisms presented significant results concerning phytotoxicity: strain B22.2, strain γ, strain A14.2, strain B13, and strain B14, as they presented mortality of at least one of the plants in the stand, and phytotoxicity equal to or higher than 80%. The other microorganisms showed less intense effects and plant recovery throughout the trial evaluation period. Table 2 presents the quantitative evaluation of six different variables studied, namely: plant height (PH), root length (RL), shoot fresh mass (SFM), root fresh mass (RFM), shoot dry mass (SDM), and root dry mass (RDM). In plants of Conyza sp., no significant inhibitory effects were observed. Similar results were found for the species, B. pilosa, which indicated a reduction in plant height only in A14.2 (33%) and in B22.2 (16%). The other variables showed no inhibition effects. The E. crusgalli and A. hybridus plants presented higher sensitivity to the treatments applied. In E. crusgalli plants, all treatments indicated inhibitory effects of fresh and dry mass, both in shoots and in roots. The treatment, B14, showed the best inhibition rates in PH (19%), and had variations between 52 and 67% of inhibition in the variables, SFM, RFM, SDM, and RDM. The results observed in A. hybridus plants demonstrated the ability of the B13 treatment to reduce PH, SFM, and SDM, with reduction percentages between 22 and 73%. Additionally, no visible phytotoxicity lesions were observed in any of the treatments and plants.

Strategies to Increase the Herbicidal Activity of the Fermented Broth
Considering the results obtained in the tests of application of the crude broth, the two species that suffered significant effects (E. crusgalli and A. hybridus) were selected. Tests were performed with different methods after fermentation to potentiate the inhibitory effects observed in the previous stage of the study. Tables 3 and 4 indicate the results of all the methods proposed and applied to E. crusgalli and A. hybridus plants.  Table 4. Effects of metabolites from the crude broth, triple dose, and PF-UAE, as well as co-cultivation, and co-cultivation with PF-UAE on A. hybridus parameters at 14 days after application (DAA).

Triple Dose Application
In E. crusgalli plants (Table 4), the best treatments with the application of triple dosage were B14 and γ, which only did not change the PH. B14 remained with the same amount of inhibition as presented in the crude broth with a single dose. The treatments, B13 and B14, indicated some wilting symptoms in E. crusgalli plants after 3 DAA, indicating possible effects. Nevertheless, the plants recovered until the finalization of the experimental step.
Moreover, the treatments with B13 and γ demonstrated that the application of more doses influences the inhibition, doubling the number of inhibited variables. The opposite was observed in the treatment with B22.2, where better results were not observed with a higher volume of broth applied.

Post-Fermentation Ultrasound-Assisted Extraction (PF-UAE) of the Raw Broth
The treatments of strains, B13, B14, and B22.2, did not show significant inhibition effects on A. hybridus plants. The results were better with crude broth application than when compared with the PF-UAE.
The best effect on A. hybridus was when the γ treatment was applied, which indicated significant results in RL (33%), SFM (67%), SDM (52%), RFM (73%), and RDM (70%). Compared with the application of the broth without the PF-UAE, an improvement in the bioherbicidal activity was observed, evidencing the potential of this microorganism.
Inhibitions were higher in E. crusgalli plants, where the treatments, B22.2 and A14.2, presented the best inhibitory effects, influencing the development of all variables. Inhibitions ranged from 15% to 60% for B22.2, and 17% to 60% for A14.2. Compared with the application of the raw broth, it was observed that the PF-UAE improved the efficiency of most treatments applied to E. crusgalli.

Co-Cultivation Medium
The tests were performed with the addition of microorganisms from the row and column from the matrix presented in Table 5. The data are expressed as the first symbol referring to the row and the second referring to the column. The − sign indicates an inhibition effect and a + indicates a growth effect of the microorganism.
The selected combinations were A14.2/B22.2, B13/B14, B14/γ, and B22.2/γ. In this methodology, the broth of the co-cultivation of the microorganisms, B14 and γ, presented better performance than the other treatments in the application in plants of A. hybridus, besides obtaining better performance than those evaluated alone.
The treatment A14.2/B22.2 had no effects on A. hybridus, but the opposite was observed in E. crusgalli plants, where this treatment was the only one that presented the inhibition of variables. The inhibitions observed were in the fresh and dry mass, both in the shoots and roots.

Post-Fermentation Ultrasound-Assisted Extraction (PF-UAE) of the Co-Cultivation Broth
Evaluating the plants of E. crusgalli with the treatment derived from PF-UAE over cocultivation, a significant improvement was observed against the treatment of co-cultivation broth without UAE, also compared to the raw broth. In E. crusgalli, A14.2/B22.2 presented a higher percentage of reduction of variables compared to the treatment of co-culture without UAE, also showing a reduction of CR (12%). Metabolites from the co-cultures, B14/γ and B22.2/γ, showed inhibition of five variables in E. crusgalli both with and without the use of PF-UAE over the broth.
In A. hybridus, growth effects are systematically observed in all methodologies explored, with suspicion of troubles in the growth of the control plants, since the masses of the control plants are systematically lower than the average masses of the other tested plants. It is important to note the positive results achieved in the first assays carried with raw broth (Table 2).

Initial Bioprospecting
The results presented in the applications of C. sativus indicated that the broths from these microorganisms are promising, and the bioherbicide potential in weeds can be evaluated. The other microorganisms manifested less intense effects and recovery of the plants throughout the trial evaluation period.
Accordingly, the initial application in C. sativus gives considerable information to the next steps. C. sativus is commonly explored to perform bioassays, since it is a plant with high sensitivity to phytotoxic compounds, and is easy to grow [26,57]. Yellowing and leaf spots were also reported in studies, which led to the selection and identification of the fungus, Diaporthe sp. [57]. The fungus, Fusarium fujikuroi, was also selected by prospecting for microorganisms tested in C. sativus, showing dark spots and leaf chlorosis [26].

Inhibition Potential on the Studied Weeds
The four weeds initially evaluated have an important commercial interest, since they are known for affecting the production of important crops. Conyza sp. affects the crops of soy, cotton, and corn [5,58]; B. pilosa and A. hybridus affect the crops of rice, corn, and soy [4,5]; and E. crusgalli affects mainly rice crops [5,59]. Moreover, all these four weeds have been reported to exhibit resistance to commercial herbicides [5].
In these weeds, the six evaluated variables (PH, RL, SFM, RFM, SDM, RDM) were chosen for being meaningful in the impairing potential of the weeds over the crops. The reduction in weed height decreases the intensity of available light, due to the shading caused by the crop of interest, affecting its photosynthesis. Therefore, the reduction in root length/mass reduces the consumption of nutrients, such as water absorption, nitrogen compounds, and minerals, reducing weed development, as well as increasing the availability of these nutrients for the crop of interest [60][61][62].
Promissory results were observed in E. crusgalli and A. hybridus. Some reports in the literature already present studies of microorganisms, such as Lasiodiplodia pseudotheobromae, Pseudomonas aeruginosa, and Carum carvi, promissory for these species [27,63]. Studies have reported a reduction in height and shoot and root fresh and dry mass of E. crusgalli and B. pilosa from metabolites of the fungus, Diaphorte schini [51]. Furthermore, few studies present the biological behavior of the plant when submitted to secondary metabolite phytotoxins [20]. The reduction of weed fresh mass was also observed with the application of the secondary metabolites of Phoma dimorpha in Senna obtusifolia plants [64]. The delay and/or inhibition of plant growth may be associated with phytotoxin interference in the biosynthesis of amino acids, such as valine, leucine, and isoleucine, which are essential for plant growth and better performance [51,65].
The unsatisfactory results for Conyza sp. may be related to the fact that the genus, Conyza, has records of resistance to several chemical herbicides, such as Paraquat, glyphosate, and 2,4-dichlorophenoxyacetic acid (2,4-D) [5], besides being known to be one of the most resistant to chemical herbicides [7][8][9]. The species, B. pilosa, also gives insufficient results, which only presented a significant reduction in plant height in A14.2 (33%) and B22.2 (16%). The other variables showed no inhibition effects. The low effects may be associated with the low concentration of metabolites in the supernatant [66].
Due to the lack of significant results from the application tests of the crude broth on the Conyza sp. and B. pilosa, the next steps were carried out with the two species that were susceptible to the treatments (E. crusgalli and A. hybridus).

Strategies to Increase the Herbicidal Activity of the Fermented Broth
The low concentration of metabolites may be associated with the lack of phytotoxic results, as well as the rapid volatilization of the product, causing the broth not to be absorbed by plant tissues [25,[67][68][69]. Moreover, some bioherbicides of microbiological origin require post-fermentation steps, as well as the use of adjuvants, in search of methods that improve product efficiency [50,68]. Aiming to potentialize the inhibitory effects over these weeds, tests were performed with four different methods: dose augmentation, cocultivation of microorganisms, post-fermentation UAE (PF-UAE) of raw broth, and PF-UAE over co-cultivation broth. Higher dosage aims to directly increase the amounts of metabolites available to interact with the plant tissues, whereas co-cultivation and PF-UAE can give rise to synergetic effects between metabolites, or even the emergence of new ones.

Triple Dose Application
The increase in the dose of broth applied aims to simulate the equivalent effect of the application of broths with a higher concentration of metabolites, and it is commonly used in agriculture, as it aims to increase the effectiveness of the applied product [70]. Wilt symptoms were also observed with the application of secondary metabolites produced by Phoma sp. [25]. The doubling in the number of inhibited variables for treatments with B13 and γ with the increase of application dose indicates that these microorganisms have the potential for bioherbicidal activity in E. crusgalli. Studies indicated that increasing the applied dose potentiates the expected effects on plants, as it consequently increases the number of metabolites disposed of in the plant [25,71].

Post-Fermentation Ultrasound-Assisted Extraction (PF-UAE) of the Raw Broth
The worst results obtained with the application of PF-UAE in bacteria (strains B13 and B14) may be associated with the metabolites being sensitive to the effects of the process. Bacteria have a few complex structures, such as protein compartments that have some enzymes necessary for cellular metabolism, and there are not many intracellular products to be extracted [72,73]. Compared to the application of the raw broth, it was observed that the PF-UAE improved the efficiency of most treatments applied to E. crusgalli. Some metabolites can injure and damage some weed species more than others [74]. The γ treatment showed significant results in A. hybridus and not in E. crusgalli. The application of this methodology indicated that the possible metabolites from the intracellular medium have more pronounced effects on grasses. The more pronounced improvements were in microorganisms with fungal morphology, indicating that the higher cellular complexity of the fungi possibly plays a role in the achieved results. Fungal growth can be filamentous, with dispersed hyphae, or in the form of spherical pellets, consisting of compact hyphae [75,76]. These microorganisms grew in the form of pellets, which may interfere with the transport of substrates into the cell, which may have significant effects on cellular metabolism and the synthesis of products and defense mechanisms [77].
As fungi are eukaryotic organisms, the complexity of the intracellular environment is higher when compared to the prokaryotic cells of bacteria [77]. Fungi have a range of organelles, including fungal vacuoles that are responsible for producing and storing digestive enzymes, among other functions [78,79]. Considering these differences between fungal and bacterial cells, the PF-UAE over the intracellular components from fungi tends to be more promising compared to bacteria, given the higher diversity of metabolites, mainly digestive enzymes found in fungal cells [79].

Co-Cultivation Medium
This technique promotes advantages in improving quality; it has potential in the replacement of substrates, but mainly stands out for the increase in product yield [80]. Different metabolites may have been produced by co-cultivation, which improved the performance of the treatment with the γ microorganism, and reinforced the bioherbicidal potential of the B14 microorganism, which had already shown efficiency in other methodologies. This is due to competition between microorganisms, which, when inoculated together, are exposed to stress-inducing factors that stimulate gene complexes that have not yet been expressed under normal culture conditions and can induce the production of new compounds [32,81]. For the other treatments, co-cultivation was not more efficient than the crude broth in A. hybridus. Additionally, to the production of new compounds, the loss of some metabolites present in the culture may also occur when submitted to co-cultivation [81].

Post-Fermentation Ultrasound-Assisted Extraction (PF-UAE) of the Co-Cultivation Broth
The combination of PF-UAE and co-cultivation techniques promotes a few significant improvements compared with these two techniques employed isolated, mainly in A14.2/B22.2 cultivation over E. crusgalli. This is possibly due to the synergy in the characteristics of these two techniques, as already discussed in Sections 4.5 and 4.6.
Studies claimed that Amaranthus is among the most difficult species to control globally, and in field experiments, A. hybridus was indicated to be more resistant to treatment applications than E. crusgalli [82].
Growth effects are systematically observed in all methodologies explored (PF-UAE and co-cultivation medium) against A. hybridus, where it was observed that the averages of all parameters of the control were lower than most treatments. This observation may be directly related to the inadequate development of the control plants, as some factors influence plant evolution, such as water, temperature, and light, which is the most important variable that affects photosynthesis and plant growth and development [65]. The hypothesis of problems with the growth of control is reinforced by the positive results achieved in the first assays carried with raw broth, presented in Sections 3.2 and 4.2.

Microbial Identification
The properties and applications of Bacillus velezensis have been studied both in agriculture and in the fermentation industry. This microorganism is easy to grow and has a high ability to resist stress. The literature presents reports of applications as a pesticide to combat classified airborne fungi or soil diseases [83]. Furthermore, some reports indicated Nigrospora sphaerica as an important species against the water hyacinth [84]. Additionally, in Table 4, the results show A. hybridus presents higher resistance compared to E. crusgalli plants. The significant differences mostly refer to growth. There was only inhibition in the plant height with B14 and γ treatment. Growth effects are systematically observed in all methodologies used (UAE and co-cultivation medium) against A. hybridus, where it was verified that the averages of all parameters of the controls were lower than most treatments. The microorganisms presented by the bioprospecting performed indicated a high potential to produce secondary metabolites for weed control.

Conclusions
The purpose addressed in this study was promising in obtaining a bioherbicide for weed control through the production of metabolites by submerged fermentation. The highest growth-inhibiting effects on E. crusgalli plants were obtained when the A14.2 and B22.2 (N. sphaerica) broth with biomass was submitted to PF-UAE for cell disruption. The metabolites from the co-cultivation of B14 (B. velezensis) with γ (A. flavus) were indicated to be significantly efficient in inhibiting the development of A. hybridus. The exploration of methodologies such as co-cultivation and ultrasound-assisted extraction proved to be excellent tools for obtaining new compounds, which may be promising in the production of a bioherbicide. The results obtained encourage the development of more studies of concentration and formulation for the potentiation of the bioherbicidal activity of the metabolites coming from the microorganisms explored in the bioprospecting performed in this study.