Salmea scandens shrubs were collected in San Rafael Toltepec Oaxaca, Mexico (15°88’ N and 96°45’ W, 341 masl) during July 2019. The bark was manually removed from the stem and dried under shade until a moisture content of 12% was reached, three days were necessary (AOAC, 2000). Afterwards, a total of 300 g of dry stem bark was ground in a mill (FRITSCH® model Pulviressette 19) and sieved through a 2-mm mesh. The samples were stored in vacuumed-sealed polyethylene bags at room temperature and kept in complete darkness until used.

Maceration in ethanol was used as the extraction method to obtain 3- and 10-day extracts. Because of the limited information regarding the maceration extraction time of alkylamides from S. scandens stem bark, a 3-day maceration time was implemented as a starting point. This time has been reported as the optimal for alkylamides extraction time for Echinacea purpurea using an ethanolic extract (Spelman et al., 2009). Additionally, this time was used because E. purpurea and S. scandens belong to the same family of Asteraceae and contain essentially the same type of alkylamides (Bauer & Remiger, 1989; Villa-Ruano et al., 2015). Regarding the 10-day maceration time, it was selected based on a previous study where the extraction process of alkylamides from S. scandens reached equilibrium (maximum concentration of alkylamides) at this time (Pacheco-Esteva, 2018). Powdered samples were placed in a flask with a 1:5 weight ratio of dried stem bark (g) and anhydrous ethanol Fermont® and left at room temperature (13°C during the night and 28°C as the highest temperature in the shadow during the day) for 3 and 10 days. The extracts were filtered and concentrated using a rotary vacuum evaporator (Buchi ®, R-100) under reduced pressure at 40°C. A total of 3.78 and 4.23 g of concentrate were obtained for 3 days and 10 days, respectively. Finally, the concentrate was stored at 4°C in a desiccator until the day of the test. The extraction process was made in triplicate and the percentage of extraction was determined according to Eq. (1) proposed by Prasad et al. (2009):

The relative concentrations of alkylamides of the 3- and 10-day extracts were determined by thin layer chromatography (TLC). First, the extracts were eluted in hexane/AcOEt in a 2:1 proportion; then, the areas of the spots corresponding to N-isobutyl-(2E,4E,8Z,10E/Z)-dodecatetraenamides (retention factor R_f = 0.5) for each extract were measured (corroborated as described below). Finally, the relative concentration of alkylamides was estimated by the ratio between the two areas. The areas were determined five times.

Since isobutyl amides are strongly associated to larvicidal activity (Villa-Ruano et al., 2015), the presence of alkylamides in S. scandens were verified prior to the bioactivity tests. Considering a R_f of 0.5 in hexane/AcOEt (2:1) reported for N-isobutyl-(2E,4E,8Z,10E/Z)-dodecatetraenamides (Bauer & Remiger, 1989), a semi-purified fraction was obtained from an ethanolic extract from the stem bark of S. scandens by using a preparative layer plate, silica gel coated with flourescent indicator F254, size 20 x 20 cm, 1.5 mm layer thickness (Merck®). The semi-purified fraction was dissolved in deuterated chloroform (CDCl₃) to be characterized by Proton and Carbon Nuclear Magnetic Resonance (¹H NMR, ¹³C NMR) in one and two dimensions using a JEOL ECA 500 MHz spectrometer. The semi-purified fraction was also characterized by Attenuated Total Reflectance-Fourier-Transform Infrared spectroscopy using a Nicolet 6700® FTIR-ATR spectrometer from 4000 to 650 cm^-1, with 32 scans accumulated for the spectrum.

The FAW larvae used for the tests were collected from local landrace maize plants ˈCriollo bolitaˈ, from San Antonio de la Cal and Zaachila (17°03’10” N and 96°69’80” W, 1500 masl), in Oaxaca, Mexico. According to the farmer, no insecticide was applied to the local landrace maize plants before the larvae were collected. We identified the larvae based on specific morphological characteristics such as the inverted Y on the head and the presence of four raised dots on body segments (FAO, 2018; CABI, 2020). For the in vitro testing, 500 larvae were carefully collected from the field and kept individually in plastic vessels with maize leaves and perforated lids. The larvae were taken to the laboratory where the following selection process was implemented. First, the larvae were acclimated in the laboratory for three days at room temperature (25 ± 2°C) and fed with fresh maize leaves (FAO, 2018). The lengths of 30 larvae were measured, which ranged between 21.10 and 34.40 mm; and the average weight of that larvae was of 0.252 ± 0.042 g. Therefore, the larvae used for this study corresponded to the 5^th and 6^th instars, which cause large crop damage and consume a considerable amount of corn foliage (Guzmán Prada et al., 2016). Finally, to select uniform larvae, it was decided to use only those close to the average weight; in consequence, the smallest and the largest were discarded. For the in situ testing, the presence of the larvae in the plant was corroborated and the larvae were visually selected for the tests in accordance with their morphological characteristics; as in the in vitro test, plants with small larvae were also discarded from the experiment.

The stock solution was prepared by dissolving 3.00 g of the ethanolic extract concentrate from S. scandens and 0.08 mL of Tween 20 as an emulsifier (corresponding to 0.8% w/w) in 10 mL of distilled water. The stock solution was used to prepare the different doses by diluting it with distilled water. A solution of distilled water with 0.8% Tween 20 was used as a control for all the tests. The details of the doses used for each test are described in the corresponding section.

The efficacy of the 3-day maceration extract was evaluated by topical application in an in vitro screening test. The efficacy of the 10-day maceration extract was evaluated also by topical application; however, an additional lower dose value was included based on the results from the previous screening test with 3-day maceration extract. Finally, the effectiveness of the 10-day maceration extract was evaluated in the in situ tests by both topical and spraying methods.

To evaluate the efficacy of the S. scandens extracts, tests were carried out following the method of topical application proposed by The American Entomological Society (Brazzel, 1970) with slight modifications. The modifications consisted in applying 50 µL of extract at specific concentration to the larvae instead of 20 µL because 5^th and 6^th instars larvae were evaluated in the present study instead of second instar larvae. The selected larvae were taken with three fingers covered by latex gloves, and the emulsified extract was applied on the prothorax with a micropipette (Fig. S1 [suppl]). After the application, the larvae were kept in an individual plastic vessel with a perforated lid at room temperature and with fresh maize leaves as food. The monitoring of mortality (Eq. 2) was carried out after 12, 24 and 48 h post application and larvae that did not react to a poking stimulus were considered dead:

where X=larvae living in the control after treatment, Y=larvae living after the extract application.

For the screening test, the 3-day extract was used at doses per larva of 0.10, 0.20 and 0.30 mg µL^-1 (see Appendix [suppl] for calculations). Eight larvae per dose and four replicates were considered for this test besides a control group.

For the 10-day extract, doses per larva were complemented to 0.05, 0.10, 0.15, 0.20 and 0.30 mg µL^-1 to evaluate the effect at lower doses. For this test, 10 larvae per dose, four replicates and a control were considered.

To validate the effectiveness of the alkylamides-rich extract of S. scandens obtained by 10-day maceration on the larvae in situ, two experiments were performed, one by topical application and the another by spraying. The experiments were carried out in a maize plantation in Zaachila, Oaxaca, Mexico (16°55´51.0” N; 96°43´21.8” W; 1550 masl), in late August 2019 when the temperature and solar radiation were high (mean temperature of 30°C, no clouds present). Since the testing was carried out in an uncontrolled environment, the mortality (Eq. 3) as the reduction of the larvae population after the extract treatment was determined according to Henderson & Tilton (1955):

where CA=larvae population of control after treatment, TA=larvae population after the extract application, CB=larvae population of control before treatment, and TB=larvae population before the extract application.

a) Field test by topical application

A completely randomized experimental design was used to evaluate the reduction of the FAW larvae population caused by S. scandens extract at doses per larva of 0.10, 0.20 and 0.30 mg µL^-1. Dried ethanolic extract (50 µL) emulsified in distilled water at the corresponding dilution was applied directly over the larvae localized in the plant. The experimental unit consisted of 10 larvae (one larva present in each plant) and 4 replicates made per dose. A control group receiving distilled water with 0.8% Tween 20 was included. The reduction of the larvae population produced by the S. scandens extract was evaluated 24 h post-application using Eq. (3).

b) Field test by spraying

In the maize plantation, an area of 20 × 20 m was delimited to apply two doses of emulsified S. scandens extract by spraying 0.50 ± 0.06 mL into each maize whorl using a hand sprayer; the equivalent doses per larva were 0.05 and 0.10 mg µL^-1. The application was carried out using a completely randomized experimental design with 40 plants per experimental unit and where the presence of one larva of FAW had been previously corroborated. In addition, distilled water with 0.8% Tween 20 was used as a control over 40 plants with larvae. The reduction of the larvae population produced by the S. scandens extract was evaluated 24 h post-application using Eq. (3).

Prior to the analysis of the data, the mortality data of S. frugiperda were corrected for control mortality by using Abbott’s correction formula (Abbott, 1925). The data of larvae survival probabilities were subjected to the survival analysis using the non-parametric Kaplan-Meier model as a function of time. The log-rank (Mantel-Cox) test was employed to evaluate the significance of the pair-wise comparisons of the mortality and survival curves. All tests were performed to a significance level of α =0.05. The IBM Statistical Package for Social Sciences (SPSS) version 25® was used for all the analyses.

The ethanolic extract was firstly analyzed by TLC (Fig. 1a), and the most prominent spot corresponded to alkylamides, as the spectroscopic characterization showed. As described in the methodology, the compounds assigned to alkylamides were fractionated to characterized them. The signals observed in the ¹H NMR spectrum from the semi-purified fraction of N-isobutyl-(2E,4E,8Z,10E/Z)-dodecatetraenamides (Fig. 1b) agree with all expected protons, and these were assigned based on the data reported by Bauer et al. (1988) and corroborating the coupling constant. Additionally, the corresponding signals of ¹³C were assigned by 2D HSQC and HMBC (see Figs. S2 and S3 [suppl], respectively) spectra and the signals observed in the ¹³C NMR spectrum are shown in the Fig. 1c and 1d. The details of the ¹H and ¹³C NMR signals are shown in Table 1.

The FTIR-ATR spectrum obtained from the semi-purified N-isobutyl-(2E,4E,8Z,10E/Z)-dodecatetraenamides from S. scandens (Fig. S4 [suppl]) showed the characteristic bands associated with alkylamides in accordance with Bauer et al. (1988). The characteristic bands identified were N-H stretching band at 3286 cm^-1, and C=C and C=O stretching bands at 1656 and 1544 cm^-1, respectively. The C-H stretching double bond was observed at 3020 cm^-1 and the C-H stretching and bending double bond at 1628 cm^-1 and 997 cm^-1, respectively.

The average percentages of the concentrates extracted at 3-day and 10-day macerations were 2.52 ± 0.13 and 2.82 ± 0.16, respectively. These results indicate that the yield significantly increased when the maceration time increased from 3 to 10 days (F=10.58, df=1; p=0.012). In addition, the relative concentrations of alkylamides increased from 2.49 to 3.52 when the maceration time increased from 3 to 10 days, respectively. A significant difference between relative concentrations of the two maceration times was also found (F=470.41; df=1, p<0.05).

a) The 3-day maceration extract (screening test)

The mortalities caused by the 3-day maceration extract from the in vitro screening test were recorded after 12, 24 and 48 h. The results analyzed as cumulative survival probability of S. frugiperda are shown in Fig. 2a. They indicate that the survival probability of S. frugiperda larvae decreased with the increase in the dose of S. scandens extracts and the monitoring time. The decrease in survival probability with the increase in dose followed the same pattern for the different monitoring times, which allows to preliminarily conclude that no apparent dose and monitoring time interactive effect occurs. It is worth noting that no larvae mortality was recorded for the control. The survival probability indicated significant differences between doses (p<0.05) at 48 h post application. The highest larval survival was recorded by 0.10 mg μL^-1 with 71.87 ± 2.09%, followed by 0.20 mg μL^-1 with 50 ± 2.81%, and the lowest by 0.30 mg μL^-1 with 40.62 ± 3.73%.

The Log Rank (Mantel-Cox) test results indicate that the survival distributions were significantly different (X² = 27.23; df = 3; p = 5×10^-6). Pair-wise comparisons detected that the survival of the control was significantly different than the applied doses (p<0.05); this means that 0.10 mg μL^-1 per larva caused a significant mortality against S. frugiperda larvae (p = 1×10^-6).

The mortality caused by the dose 0.10 mg µL^-1 per larva was significantly different than that caused by 0.20 mg μL^-1 (X² = 4.24; df = 1; p = 0.039) and by 0.30 mg μL^-1 (X² = 7.69; df = 1; p = 0.006). Finally, the mortality caused by the 0.20 mg μL^-1 per larva was similar to the mortality by 0.30 mg μL^-1 per larva (X² = 0.52; df = 1; p = 0.471). The highest mortality was 59.38% using the dose of 30 mg μL^-1 of ethanolic extract from S. scandens at 48 h.

b) The 10-day maceration extract

The mortality caused by the 10-day maceration extract from the in vitro test was recorded after 12, 24 and 48 h and analyzed in terms of the cumulative survival probability of S. frugiperda (Fig. 2b). The survival analysis results indicated a significant decrease in S. frugiperda larvae with increasing doses of S. scandens, as well as with the monitoring time. It was also observed that a similar cumulative survival probability can be reached with a low dose at a longer monitoring time, i.e. the survival at dose of 0.20 mg μL^-1 at 48 h was higher than that at 0.30 mg μL^-1 at 24 h. The highest larval survival was recorded at 0.05 mg μL^-1 with 72.50 ± 1.26%, followed by 0.10 mg μL^-1 with 40.0 ± 0.82%, and the lowest larval survival was caused at 0.30 mg μL^-1 with 20.0 ± 1.17%. Less than the 50% of larvae populations survived with a dose of 0.10 mg μL^-1 until the end of the experiment.

The Mantel-Cox test results indicated significantly different survival distributions between doses (X² = 29.42; df = 4; p = 6×10^-6). Pair-wise comparison results indicated that the survival of the control was significantly different from the rest of the doses (p < 0.05); this means that the application of the S. scandens extract, even at a low dose, caused a significant mortality of the S. frugiperda larvae.

The mortality caused by the 0.05 mg μL^-1 dose per larva was significantly lower than for the other studied doses: 0.10 mg μL^-1 (X² = 9.12; df = 1; p = 0.003), 0.15 mg μL^-1 (X² = 17.58; df = 1; p = 2.8×10^-5), 0.20 mg μL^-1 (X² = 19.89; df = 1; p = 8×10^-6) and 0.30 mg μL^-1 (X² = 24.60; df = 1; p = 7×10^-6). Furthermore, the mortality caused by 0.10 mg μL^-1 per larva was similar to that caused by the 0.15 (X² = 1.38; df = 1; p = 0.240) and by 0.20 mg μL^-1 (X² = 1.99; df = 1; p = 0.157) per larva doses, but lower than that at 0.30 mg μL^-1 (X² = 3.83; df = 1; p = 0.041) per larva. Mortality caused by the 0.15 mg μL^-1 per larva dose was also similar to those caused by the 0.20 (X² = 0.49; df = 1; p = 0.826) and 0.30 mg μL^-1 (X² = 0.62; df = 1; p = 0.432) per larva doses. Finally, the mortality caused by 0.20 mg μL^-1 per larva was similar to the 0.30 mg μL^-1 per larva doses (X² = 0.31; df = 1; p = 0.579).

The results of mortality, analyzed as survival probability, from the in situ test by topical and spraying applications of the 10-day maceration extract recorded after 24 h are summarized in Fig. 3. Fig. 3a shows the cumulative survival probability by topical application in larvae of S. frugiperda; the dose factor was significant; this means that the increase in dose caused a significant decrease in survival of larvae. According to survival probability, the highest larvae survival was recorded by 0.10 mg μL^-1 with 57.5 ± 1.71%, followed by 0.20 mg μL^-1 with 45.0 ± 1.17%, with the lowest larval survival caused by 0.30 mg μL^-1 with 37.5 ± 1.26%. Less than the 50% larvae populations survived with a dose of 0.20 mg μL^-1 at 24 hours. No larvae mortality was recorded for the control.

The Mantel-Cox test results indicate significantly different survival distributions (X² = 37.95; df = 3; p = 2.9×10^-8). Pair-wise comparisons between the control and different doses were significantly different: 0.10 mg μL^-1 (X² = 20.06; df = 1; p = 8×10^-6), 0.20 mg μL^-1 (X² = 29.97; df = 1; p = 4.4×10^-8) and 0.30 mg μL^-1 (X² = 35.91; df = 1; p = 2.1×10^-9). The comparison also indicates that the mortality caused by 0.10 mg μL^-1 per larva was similar to those caused by the 0.20 (X² = 1.25; df = 1; p = 0.264) and 0.30 mg μL^-1 (X² = 3.20; df = 1; p = 0.07) per larva.

Fig. 3b shows the cumulative survival probability by spraying application. Similar to the effects by topical application on the larval survival of FAW, the survival analysis results indicated significant differences between doses. The highest larval survival was recorded by 0.05 mg μL^-1 with 72.55 ± 2.12%, followed by 0.10 mg μL^-1 with 57. 60 ± 1.04% per larva and at the dose of 0.15 mg μL^-1 a 49% reduction of S. frugiperda was reached. More than 50% of the larvae populations survived with a dose of 0.10 mg μL^-1 at 24 hours.

The Mantel-Cox test results indicated significantly different survival distributions (X² = 31.32; df = 3; p = 2.75×10^-7). The control was significantly different than the rest of the doses; 0.05 mg μL^-1 (X² = 12.59; df = 1; p = 3.9×10^-4), 0.10 mg μL^-1 (X² = 16.76; df = 1; p = 4.2×10^-5), 0.15 mg μL^-1 (X² = 26.33; df = 1; p = 2.9×10^-7), which means that even the lowest dose caused a significant mortality of S. frugiperda larvae. The mortality caused by the dose 0.05 mg μL^-1 was similar to that induced by 0.10 mg μL^-1 (X² = 0.52; df = 1; p = 0.472) whereas it was significantly different to that induced by 0.15 mg μL^-1 (X² = 4.21; df = 1; p = 0.03). Also, the mortality induced by the 0.10 mg μL^-1 dose was similar to that induced by the 0.15 mg μL^-1 dose (X² = 1.82; df = 1; p = 0.178).