Almost all groups at 0.8 mg mL− 1, presented higher inhibition percentage than that previously observed for the leaf crude extract (around 40% - Novaes et al. 2016), specially groups D, E, G and I. Although group G showed strong phytotoxicity on the elongation of wheat coleoptiles and germination rate of U. decumbens, no significant effect was observed on initial growth of seedlings. This species is an aggressive weed in Brazil, which besides its invasiveness in fields, also shows easy adaptation to the soils of the Brazilian savanna (Alvim et al. 1990) and allelopathic potential (Souza et al. 2003). Novaes et al. (2016) observed over 70% of inhibition of germination and also on shoot and root lengths when diaspores of U. decumbens were grown with crude leaf extract of Annona coriacea. The difference of results between the leaf crude extract and group G on the growth of the weed seedlings could be probably because some active compounds responsible for this effect may have been placed in other groups after fractionation of the extract.
The good germination results of Group G on U. decumbens justified its prospection for allelochemicals which could be used as natural origin pesticides in future. Since 2009, Brazil is the largest consumer of pesticides in the world (INCA, 2015). The indiscriminate and often misguided use of these products have been responsible for intoxication of human beings, especially workers who make their application in the field, and biological communities (INCA 2015). Thus, the search for natural pesticides, which are biodegradable and do not produce contaminants such as synthetic products, is of fundamental importance.
The group G afforded 11 glycosylated flavonols, as presented before (Novaes et al. 2018). The flavonol occurrence in Annonaceae species was already described by Santos & Salatino (2000). Compounds 5, 7 and 11 significantly affect the elongation of wheat coleoptiles. Novaes et al. (2013b), Nebo et al. (2014), Watanabe et al. (2014), and Marsni et al. (2015) described phytotoxic activity of many flavonoids on the elongation of wheat coleoptiles. Most of the compounds showed effects between 0–70% of inhibition, even in the highest concentrations, as 1 mM. Some aglycones, however, were very active. The flavone (2-phenyl-4H-1-benzopyran-4-one) showed almost 100% inhibition in concentrations between 0.1 mM and 1 mM (Nebo et al. 2014), while heliannone B showed 100% inhibition at 1mM, but had strong reduction of activity in concentrations lower than 0.3 mM (Marsni et al. 2015). Here, the three compounds with the highest activity were, respectively, a quercetin, a kaempferol and an isorhamnetin glycosides, and their sugars were a glucose (5 and 11), and a glucose connected to a rhamnose (7), what is also present in other compounds without activity. Therefore, at least in this work, there was no correlation between the structures and the activities.
Compounds 5 and 11 affected the elongation of wheat coleoptiles close to those exhibited by the herbicide. These compounds did not show important results on tomato and lettuce germination and seedling growth. Compounds 5 and 11 presented completely different effects on U. decumbens: compound 5 had almost no effect on it, and compound 11 significantly affected its germination, showing elevated standard-deviations. High standard-deviations are expected for results using U. decumbens since it is a wild species and present wide intrinsic variation. At first, germination should be less affected by phytotoxic compounds than plant growth because of the barrier provided by seed, especially in those of larger endosperm (Soltys et al. 2012). However, this was not our case. Reduction of weed germination is especially important to farmers interested in their control. Other authors have already reported the same effect on this weed (Novaes et al. 2016; Rial et al. 2014).
One possible explanation for this difference between the results of compounds 5 and 11 could be related to the distinct complexity of the biological systems in both assays. While coleoptile assay evolves the activity of a single tissue with less differentiated cells, germination and initial growth assays are developed on more complex systems, evolving the development of a new individual, which could also show complex defense responses against stressful environment (Gniazdowska et al. 2015), in our case, the presence of an allelochemical.
Phytotoxicity of the compounds 5 and 11 have been studied before on standard target species. Parvez et al. (2004) observed a decrease of 70% on the initial growth of Arabidopsis thaliana in the presence of 5. Contradicting our results, Almeida et al. (2008) observed 60% of inhibition of lettuce root elongation when exposed to compound 5. For these authors, flavonoids with a catechol group, like compound 5, are responsible for changes on cellular membrane permeability and modify the radicular lengthening needed for root protrusion. Compound 11, isolated from leaves of Melilotus neapolitana Ten. (Fabaceae) by Esposito et al. (2008) showed no effect on coexisting species of Mediterranean herbaceous plant community, Petrorhagia velutina (Guss.) P.W.Ball & Heywood (Caryophyllaceae), Dactylis hispanica Roth, and Phleum subulatum (Savi) Asch. & Graebn. (Poaceae).
Although some effect of compounds 5 and 11 were observed on germination/initial growth with the three target species used (tomato, lettuce and U. decumbens), they were much less expressive than those previously found with ethanolic crude leaf extract on Annona coriacea (Novaes et al. 2016). With the crude extract, germination inhibition was around 60–70% and the initial development of shoots and roots were also reduced for all the three target-species. These differences between the results could be due to the joint action of the compounds when they are together in the extract. Joint action is the set of interactions that can occur when compounds are mixed and they can show synergic, additive or antagonist activities (Inderjit et al. 2002). These phenomena have been studied in elongation of wheat coleoptiles and antioxidant assays (Garcia et al. 2015; Rial et al 2016; Galindo et al. 2017), but there is no data proven it in germination and growth bioassays. Garcia et al. (2015) showed that binary mixtures of polymethoxyflavones of citrus showed synergistic effects on the elongation of wheat coleoptiles and antioxidant assays, whilst Rial et al. (2016) showed that the joint action of sesquiterpene lactones isolated from Cynara cardunculus L. (Asteraceae) were, predominantly, additive of inhibitory activity on the coleoptiles. To be more certain of additive or synergic activities between 5 and 11, experiments of binary mixtures of the compounds, with different variations of concentrations, as performed by these authors, would be necessary, especially in germination and growth bioassays.
Concluding, the groups D, E, G and I of A. coriacea extract showed higher inhibitory activities on elongation of wheat coleoptiles than that observed for the crude extract. Eleven flavonols were isolated from group G and, as far as we know, this is the first report of wheat coleoptile elongation bioassay to all the compounds and it is also the first phytotoxicity results for five of them. There was no apparent correlation between the structure of the compounds and their activity in the present study. While isoquercitrin (5) and isorhamnetin-3-O-glucoside (11) showed elevated phytotoxicity in wheat coleoptiles elongation bioassay, they had no effect or were mainly stimulatory to germination and initial growth of lettuce and tomato. Only 11 was inhibitory to U. decumbens germination. The higher structural complexity of seeds, when compared to wheat coleoptile, is suggested as a possible explanation for distinct effects of isolated flavonoids on coleoptile elongation and germination/initial growth assays. There were also differences between the activity of the isolated compounds and the crude extract of A. coriacea on the target plants and it could have resulted from joint action of the compounds, which shall be tested in further experiments.