Plant materials

Leaves of M. disticha and M. magellanica were collected in the Conguillio National Park, while leaves, bark and seed of M. boaria were collected at the Maquehue experimental station belonging to Universidad de La Frontera. Both sites are located in the Araucanía Region, Chile. Taxonomic identification was done in the Faculty of Farming, Livestock and Forestry Sciences, Universidad de La Frontera, Araucanía Region.

Preparation of the ethanolic extract

Extracts were prepared in the Laboratory of Ecological Chemistry, Universidad de La Frontera. The plant material was macerated in ethanol p. a. at 100 % in a proportion of 3:1, and removed after 72 hours; this process was repeated three times. It was then roto-evaporated and lyophilised.

To determine the secondary metabolites present in the extract, simple and fast tests were performed. The following tests were performed to determine the phytochemical screening of the ethanolic extracts: Alkaloids: Precipitation and coloring reactions, with the reagents of Drangerdoff, Mayer and Wagner; Tannins: Reaction with Ferric Chloride; Flavonoids: Precipitation and color reaction reactions by Shinoda test; Anthraquinones: Bornträger reaction; Coumarin: Precipitation reaction and staining with NaOH; Saponins: Foam test; Terpenos / Steroids: Lieberman-Burchard reaction and Salkowski reaction.

Insects

Adults of H. convergens were collected from fennel (Foeniculum vulgare Mill.) and thistle (Cirsium vulgare (Savi) Ten.) plants on farms next to Angol (37°48’S, 72°43’W). In this species only the mortality was evaluated. Following the method used by ^{Rebolledo et al. (2012)} to evaluate mortality, four specimens were deposited in ½ litre plastic containers, with 10 repetitions. They were fed on maize flakes sprayed with the treatments, including the control with no application, using distilled water as the solvent. The treatments correspond to: 3; 6; 9; 12 and 15 % w/v. The mortality was evaluated after 24, 48, 72, 96 and 120 hours, the food being renewed each time.

Starting in November 2013 adult of A. superciliosus were collected from commercial blueberry fields belonging to San José Farm, owned by Exportadora Patagonia Food S.A., located in Gorbea (39°06’S, 72°41’W) in the central valley of the Araucanía Region. A second sample of insects was collected from November 2014. Following the method used by ^{Rebolledo et al. (2012)} the extracts were applied on leaves of R. ulmifolius which were put in ½ litre plastic containers with one insect adult of A. superciliosus, with 10 repetitions. The treatments correspond to: 2, 5, 10, 20 and 30 % w/v. The mortality was evaluated after 24, 48, 72, 96 and 120 hours, the food being renewed each time. Additionally, the method of ^{Zapata et al. (2006)} was used to determine food preference through two experiments:

No-choice assay (NC): Only treated or untreated food with extracts was supplied in the container.

Choice assay (C): Treated (extract) and untreated (no application) food was supplied in the container.

The food substrate given to the insects was bramble leaves (Rubus ulmifolius Schott (Rosales: Rosaceae)). Plant extracts were applied uniformly with an atomizer at the start of the assay and when the food was renewed at each evaluation. A control with no application was used. Leaf area consumption was evaluated after 24, 48, 72, 96 and 120 hours. The leaf area of the leaves used in the assay, and the difference consumed, were calculated using the digital image processing programme ImageJ 1.47v developed in the National Institute of Health, which is in the public domain.

Experimental design and statistical analysis

The experimental design used for all the assays was completely randomized, with 10 repetitions. Abbot’s formula was used to correct the natural mortality found in the control. The effectiveness of the treatments was evaluated by analysis of variance (ANOVA), with significance level α = 0.05, after the data was transformed using the square root of the arcsine of the percentage (√x %^-1) in order to ensure homogeneity of the variances. Statistical analysis of the experimental data was performed using SAS® (SAS Institute, 2009) to find out the (Log probit analysis (at 95 % confidence level), regression equations and mean percentage mortality. The means were separated using Tukey’s test (α = 0.05) (^{Silva et al. 2003}).

Bioassays on H. convergens

The highest mortalities were recorded for MBS (82 %) and MML (54 %) extracts. None of the other ethanolic extracts evaluated presented a mortality higher than 50 % (MBL: 37 %; MDL: 27 %; and MBB: 25 %) (Fig. 1).

Figure 1 Mortality means ± S. E. of Hippodamia convergens for each extract. Different letters in the columns indicate significant differences after analysis by ANOVA followed by the Tukey’s test at P < 0.05. MBS: M. boaria seed; MBL: M. boaria leaf; MDL: M. disticha leaf; MML: M. magellanica leaf; MBB: M. boaria bark. 

For the MBS ethanolic extract there were significant differences between the concentrations, being the mortality dependent on the dose. The median lethal concentration for this extract was 3.2 % w/v (32 mg/ml). For all concentrations evaluated, the mortality was higher than 50 %. For other extracts no significant differences were observed between the concentrations. The mortality per extract, treatment and comparison of means are given in Table 2.

For all extracts the highest mortality was recorded at 24 hours of exposure (Table 3). The mortality means by extract, treatment and evaluation time are given in Table 4.

Bioassays on A. superciliosus

Toxicity. In the choice assay, only the MBL and MBS extracts produced mortalities higher than 50 % (Fig. 2), while this was exceeded in the no-choice assays with MBL, MDL and MBS extracts (Fig. 3). The highest mortality was obtained with MBS extract in both assays. The values obtained in the choice assays were: MBS: 85 %; MBL: 61 %; MDL: 48 %; MML: 35 % and MBB: 34 %. In the no-choice assay the values were: MBS: 84 %; MBL: 15 %; MDL: 51 %; MML: 13 % and MBB: 15 %. According to ^{Broussalis et al. (1999)}, a mortality over 50 % is considered promising in pesticide studies with plant extracts.

Figure 2 Mortality means ± S. E. of Aegorhinus superciliosus for each extract in choice assay. 

Different letters in the columns indicate significant differences after analysis by ANOVA followed by the Tukey’s test at P < 0.05. MBS: M. boaria seed; MBL: M. boaria leaf; MDL: M. disticha leaf; MML: M. magellanica leaf; MBB: M. boaria bark.

Figure 3 Mortality means ± S. E. of Aegorhinus superciliosus for each extract in no-choice assay. 

Different letters in the columns indicate significant differences after analysis by ANOVA followed by the Tukey’s test at P < 0.05. MBS: M. boaria seed; MBL: M. boaria leaf; MDL: M. disticha leaf; MML: M. magellanica leaf; MBB: M. boaria bark.

The mortality percentages per extract, treatment and type of assay are given in Table 5.

For the MBS extract, in the choice assay there was a significant difference between the treatments and the control, however there was no difference between the concentrations of 10-20 and 30 % w/v with LC₅₀: 23 mg/ml. In the no-choice assays there was a significant difference between the concentrations evaluated and the control, but not between the different concentrations. The mortality increased with the evaluation time, reaching its highest at 120 hours for both choice and no-choice assays.

For MBL extract, there was a significant difference between treatments in the choice assay, with LC₅₀: 43 mg/ml; the highest mortality was recorded after 72 hours of evaluation. In the no-choice assay there was no significant difference between the concentrations and the control, with the highest mortality after 24 and 48 hours.

The mortality means by extract and evaluation time are given in Table 6. The mortality means by type of assay, extract, treatment and evaluation time are given in Table 7.

As ^{Pamo et al. (2004)} indicate, the mortality increases as the dose is strengthened and as the time increases (days) since botanical pesticides are slower-acting than synthetic products. The mortality percentages in the controls remained constant over the evaluation period.

^{Alarcón et al. (2012)}, using a methanolic extract of M. boaria seeds, obtained the compound 9-β-furoyloxy-1α, 6β, 8α-triacetoxy-dihydro-β-agarofurane which presented strong insecticidal action (70 %) on larvae of Drosophila melanogaster Meigen, 1830 (Diptera: Drosophilidae). The present study shows that ethanolic extract of M. boaria seed contains compounds with insecticidal action on H. convergens and A. superciliosus.

The lowest mortality was obtained with MBB extract in choice (C) and no-choice (NC) assays. There was no difference between the assessed concentrations and increased over time, in both assays. However, it has been shown that mortality with bark extracts of the genus Maytenus was concentration dependent. In the work of ^{Avilla et al. (2000)} from an unknown specie of Maytenus, bark of the plant roots was extracted with n-hexane: diethyl ether (1:1). The mortality of Cydia pomonella (Linnaeus, 1758) (Lepidoptera: Tortricidae) was concentration-dependent and increased over time. Of the isolated compounds, 20-R-hydroxytingenone was the most active compounds against codling moth larvae. The mortality caused by 20-R-hydroxytingenone reached 100 % five days after exposure, the LC₅₀ ranging from 13.0 mg/mL (5 days later) to 8.5 mg/mL (10 days later).

Percentage leaf area consumed. In the choice assay, there was no statistically significant difference in the percentage leaf area consumed with ethanolic extracts of MBS, MBL, MML and MDL, but there was a difference between these and MBB ethanolic extract. The lowest percentage leaf area consumed was recorded with MBS extract and the highest with MBB extract (Fig. 4). A diketone isolated from the stem bark extract of Maytenus elaeodendroides showed feeding deterrent activity on Sitophylus oryzae (L.) (Coleoptera: Curculionidae), a plague that causes significant damage to rice growing (Nogueiras et al. 2010).

Figure 4 Percentage leaf area consumed (mean % ± standard error) of Rubus ulmifolius in choice assay. Different letters in the columns indicate significant differences after analysis by ANOVA followed by the Tukey’s test at P < 0.05. MBS: M. boaria seed; MBL: M. boaria leaf; MDL: M. disticha leaf; MML: M. magellanica leaf; MBB: M. boaria bark. 

In the no-choice assay, the lowest percentages of leaf area consumed were found with MBS and MBL extracts, just as in the choice assay, with no significant differences between them. The highest percentage consumption was recorded with MML extract, higher than that obtained with MBB extract in the choice assay; there was no significant differences between these extracts and their respective controls (Fig. 5). On the other hand, ^{González et al. (1993)} reported that compounds isolated from leaves of M. magellanica showed antifeedant activity on Spodoptera littoralis (Lepidoptera: Noctuidae) in a choice test. This study shows that the lowest percentage of leaf area consumed by A. superciliosus with MML extract could be related to these compounds but their activity depends strongly on concentration.

Figure 5 Percentage leaf area consumed (mean % ± standard error) of Rubus ulmifolius in no-choice assay. Different letters in the columns indicate significant differences after analysis by ANOVA followed by the Tukey’s test at P < 0.05. MBS: M. boaria seed; MBL: M. boaria leaf; MDL: M. disticha leaf; MML: M. magellanica leaf; MBB: M. boaria bark. 

The lowest percentage of leaf area consumed was obtained with MBS extract in choice (C) and no-choice (NC) assays. González et al. (1997) demonstrated that ethanol extracts obtained from M. boaria had antifeedant activity on S. littoralis (Boisduval, 1833) (Lepidoptera: Noctuidae). However, they do not specify of what part of the plant were obtained the extracts. The finding in this study shows that ethanolic extracts of seed, bark and leaves of M. boaria do not show the same activity on A. superciliosus. In insects kept in the laboratory and fed on with M. boaria branches, leaves and seeds, was observed that they first consume the bark, then the leaves and finally the seeds.

Although ^{Aguilera (1988)}, ^{Arias (2000)} and ^{Zavala et al. (2011)} consider M. boaria to be the principal host species for A. superciliosus, in this study the lowest percentage of leaf area consumed was with M. boaria leaf (MBL) and seed (MBS) extracts. However, ^{Rodríguez et al. (2003)} affirm that feed inhibitors derived from plant extracts are secondary metabolites that according to their concentration interfere insects from responding positively to olfactory stimulants, affecting their ability to find and recognize their host. These metabolites are probably those obtained in this case with ethanol as solvent. ^{Zapata et al. (2006)} showed that the anti-feeding effect of M. boaria on the third larval stage of Spodoptera littoralis was attributable to the presence of dihydro-β-agarofurane sesquiterpenes, which are very abundant in this plant.

Table 8 give the leaf area percentages consumed per extract and treatment for choice (C) and no-choice (NC) assays respectively. These tables show that percentage leaf area consumed was lowest in the C and NC assays with ethanolic extract of seeds of M. boaria at higher concentrations. There were no significant differences in consumption by evaluation time, decreasing to 72 hours in the no-choice assay. The mortality values and leaf area consumed obtained for this extract are consistent with a high insecticide activity recorded by ^{Céspedes et al. (2001)} on S. frugiperda (Smith) (Lepidoptera: Noctuidae), who conclude that the effect seems to be a combination of antifeedant action and post-gastrointestinal toxicity. While with MML extract, in the C assay for all concentrations the consumption was lower than in the control. However, in the NC assay at lower concentration the consumption was higher than control.

Most works that antifeedant and insecticidal activity of plants belonging to the genus Maytenus reported have been tested against the insect S. littoralis such as: M. senegalensis (Lam.) Exell showed a high antifeedant and insecticidal activity on third larval instar of S. littoralis. It has effects on behavior and post ingestive toxicity on the developmental stages of S. littoralis (^{El-Aziz and El-Din 2007}). From M. canariensis (Loes.) G. Kunkel & Sund. five new sesquiterpenes with skeleton dihydro-β-agarofurano were isolated, which showed antifeedant activity against S. littoralis at concentration 10 µg/ml (^{González et al. 1993}); and from leaves of M. chiapensis Lundell wilfordine, euonine and alatusinine compounds, that showed strong antifeedant activity on S. littoralis, were isolated (^{Nuñez et al. 2004}).