Feeding deterrence towards Helicoverpa armigera by Tithonia diversifolia tagitinin C-enriched extract

2.1 General experimental procedures and reagents

The Liquid Chromatography system consisted of a three-pump Shimadzu chromatograph model LC-20AR Prominence equipped with a SIL-10AF auto sampler, a CTO- 20A column oven, a CBM-20A communications bus module, a DGU-20A3R in-line degasser, and an SPD-M20A photodiode array detector. The Lab solution® software was used to process the data; the analyses were conducted in a Shim-pack VP-ODS column (250 × 4.6 mm i.d., 5 μm; Shimadzu). Acetonitrile (chromatographic grade) was supplied by Mallinkrodt Baker Inc. (Phillipsburg, NJ, USA), and water was purified with a Milli-Q-plus filter system (Millipore, Bedford, MA, USA). Classic and vacuum liquid chromatography (respectively CC and VLC; glass columns of 450 X 25 mm and 50–100 mm i.d.) were used to purify tagitinin C (Fig. 1A) by using silica gel 60 (Merck, art. 9385) and silica gel 60H (Merck, art. 7736). Commercial hexanes and ethyl acetate (EtOAc) were purified by distillation in our facilities. The NMR spectra were run on a Bruker DPX 400 spectrometer (400 MHz for ¹H and 100 MHz for ¹³C). The samples were dissolved in CDCl₃, and the spectra were calibrated with the solvent signals at 7.26 (¹H) and 77.0 (¹³C). We used HPLC grade p-coumaric acid (Fig. 1B, CA), a natural phenylpropanoid that is not found in T. diversifolia leaves, as Internal Standard, acquired from Sigma-Aldrich (≥98% purity, Darmstadt, Germany).

Close

Fig. 1

Molecular structures of tagitinin C (A) and p-cumaric acid (B).

Export to PPT

2.2 Plant material

Tithonia diversifolia leaves were collected in Patrocínio (Minas Gerais state, Brazil; S18° 56′ 38″ W46° 59′ 34″) by Gabriel da Costa Inacio. Plant identification was carried out by Rosângela de Oliveira Araújo from the University Center of the Cerrado-Patrocínio (UNICERP), Patrocínio, Minas Gerais, Brazil, and the voucher specimen was deposited in the Herbarium Uberlandense (HUFU) under the number HUFU 73.901.

2.3 Obtaining of tagitinin C-rich leaf rinse extract (TagCE) and isolation of tagitinin C

Initially, 8.0 kg of T. diversifolia leaves were rinsed with dichloromethane (6 L) for 3 min to obtain 23,290 g of TagCE after filtering and solvent evaporation (Ambrosio et al., 2008). About 9.0 g of the extract were then chromatographed over 600.0 g of silica gel 60H (Merck, art. 7736), using vacuum liquid chromatography (VLC) (Pelletier et al., 1986) with increasing amounts of ethyl acetate in n-hexane as eluent (2 L for each fraction). This procedure delivered seven fractions that were analysed through TLC, using silica gel PF254 (Merck art. 9385; 1 mm thickness) and isocratic n-hexane: ethyl acetate 1:1 as mobile phase after solvent evaporation. The second and third fractions (0.511 and 2.309 g, respectively) were combined and then submitted to another fractionation using VLC (150.0 g of silica gel 60H; increasing amounts of n-hexane: ethyl acetate – 1 L for each fraction). After TLC analysis using the conditions previously described, the tenth fraction showed a single spot and was identified as tagitinin C after ¹H and ¹³C NMR analyses and comparison with literature data (Baruah et al. 1979). The NMR data also allowed us to conclude that the purity of the isolated compound was above 99%.

2.4 Development and validation of the analytical method

Chromatographic analyses of TagCE were performed through RP-HPLC-PDA using a Shim-pack VP-ODS column (250 mm × 4.6 mm i.d., 5 μm; Shimadzu) and an isocratic system consisting of acetonitrile and water with 0.1% acetic acid (45:55 v/v) as mobile phase, which was established through initial experiments using the gradient scouting run evaluation as previously described (Snyder and Dolan, 1996). The temperature of the column was set at 40 °C, the flow rate was 1.0 mL.min⁻¹, the injected volume was 20.0 µL and the detection wavelength was set at 254 nm. The developed and validated analytical method complies with the requirements established by the National Agency for Sanitary Vigilance in Brazil (ANVISA, 2003).

2.5 Preparation of sample solutions, construction of linear analytical curves, determination of limits of detection (LOD) and quantification (LOQ)

The analytical solutions were prepared by dissolving 1.0 mg of the sample plus 0.5 mg of p-cumaric acid (CA), used as internal standard in 5 mL of acetonitrile, followed by filtration through a UNIFLO 25/0.2 PTFE syringe filter (Whatman/Schleicher & Schuell, Maidstone, UK).

A dilution series of metabolite tagitinin (400.0, 300.0, 200.0, 100.0, 80.0, 40.0, and 20.0 µg·mL⁻¹) with internal standard (CA; 100.0 µg·mL⁻¹) was prepared in acetonitrile. Then, 20.0 µL aliquots of these solutions were injected into the HPLC equipment in triplicate. Linear analytical curves were obtained by plotting the area ratio of the individual chromatographic standards to the internal standard CA. The correlation coefficient was calculated using Microsoft Excel®.

The LOD and LOQ values were determined on the basis of the standard deviation of the response (σ) and of the slope of the analytical curve (S), using the expressions LOD = 3.3σ/S and LOQ = 10σ/S.

2.6 Selectivity

The selectivity of the method was evaluated by comparing DAD spectral data in the ascending, upper and descending regions of the respective peaks of tagitinin C in the linear analytical curve and in the sample (Souza et al., 2012). All the spectra matched, confirming that no other metabolites co-eluted with the target compound.

2.7 Precision and accuracy

Intra-day precision was evaluated by analysing six samples, and inter-day precision was assayed on two consecutive days. The concentration and the relative standard deviation (RSD) of the analyte were determined in TagCE. For the inter-day precision, the concentration values of tagitinin C were compared by applying the t test (p < 0.05), using GraphPad Prism®.

The accuracy of the method was determined by spiking known contents of tagitinin C into TagCE at low, medium and high levels (respectively 25, 50 and 100% of the compound content determined in the precision experiment) in triplicate. The spiked samples were analysed, and the recoveries were calculated by comparing the measured to the spiked concentrations (Moreira et al., 2013, 2016).

2.8 Robustness

The robustness of the chromatographic method was assessed by following the experimental design proposed by Plackett-Burman for seven factors and eight experiments, as previously described (Vander Heyden et al., 2001; Moreira et al., 2016). Briefly, the selected operational factors related to the chromatographic method were as follows: mobile phase flow rate (Flow), column oven temperature (Temp.), percentage of organic solvent in the mobile phase (%B), detection wavelength (λ) and volume of injected samples (Inj.). Two dummy factors had to be included to reach a saturated design and variation levels for each of these operational factors (Table 1). Robustness values were expressed as RSD (%) of the responses.

2.9 Feeding deterrence on artificial diet

The artificial diet used in the bioassay was prepared according to Greene et al. (1976). The extract was sprayed when the diet temperature decreased and reached 25 °C, to avoid the potential degradation of compounds present in the extract. Dichloromethane extract was tested at concentrations of 100, 500 and 1000 mg mL⁻¹. In the control treatment, only dimethyl sulfoxide (DMSO) solvent was used. The extract was dissolved in DMSO at a ratio of 1% (w/w) extract on the dry weight of the artificial diet. A 0.1-mL aliquot of the diluted extract was applied to the diet with a micropipette, and the diet remained at room temperature for 30 min for drying. The procedure was repeated for each tested extract concentration, for the control (DMSO) and for the positive control (flubendiamide insecticide used at the manufacturer's recommended dose of 0.35 µL mL⁻¹). After diet drying, 10-day-old third instar larvae were added; each repetition consisted of a Petri dish (90 mm diameter) containing an artificial diet and one larva. Evaluations were performed every 24 h until pupae formation and subsequent emergence of adults. The pupae were transferred to 1000-mL plastic cups where they remained until the emergence of the adults. The experiment was conducted in a BOD chamber adjusted to a temperature of 25 ± 2 °C, a relative humidity of 70 ± 10% and a photoperiod of 14 h light and 10 h dark. The experimental design was completely randomised, with 10 replications for each treatment. The mortality of the larval and pupal insects was evaluated.

2.10 Feeding deterrence on plants

The effect of feeding deterrence, besides being tested on the artificial diet, was also confirmed in host plants. The plant species used in the experiments (cotton, beans, corn, soybeans, sorghum and tomato) were selected due to the frequent occurrence of H. armigera in these crops in Brazil (Czepak et al., 2013; Specht et al., 2013; Pratissoli et al., 2015). The experiment was conducted in a BOD chamber adjusted to a temperature of 25 ± 2 °C, a relative humidity of 70 ± 10% and a photoperiod of 14 h light and 10 h dark. The arena was made up of a 10-cm diameter acrylic Petri dish. Each experimental unit consisted of an arena containing six discs of 1-cm diameter leaf, interspersed with discs treated and untreated with the extract. In each arena, there were discs of only one host plant. Prior to the experiment, the 10-day-old third instar larvae were kept 24 h without food. One larva was inserted into the centre of each arena, and after 3 h, the leaf discs were removed, and the leaf area was measured to record the consumption in a Lee Tools device, with an accuracy of 0.01 cm². After measuring the leaf area, the discs were oven-dried at 60 °C for 48 h and weighed on a precision scale.

The procedure was repeated for each tested extract concentration (100, 500 and 1000 mg mL⁻¹) and each host plant (cotton, beans, corn, soybeans, sorghum and tomato). The control used was DMSO, and for the positive control, we used flubendiamide insecticide at the manufacturer’s recommended dose of 0.35 µL mL⁻¹. The experimental design was completely randomised, with 10 replications for each treatment.

The feeding deterrence index was calculated using the following equation (Simmonds et al., 1989): $$\text{I}=\left(\text{Ac}-\text{At}\right)/\left(\text{Ac}+\text{At}\right);$$

• Ac = control disc consumed area (mm²)

• At = treated disc consumed area (mm²).

Feeding deterrence values above 0.50 and 0.90 are considered significant and highly significant, respectively (Simmonds et al., 1989).

2.11 Data analysis

Data on the feeding deterrence on artificial diets and on plants for H. armigera were submitted to the Kolmogorov and Bartlett tests to verify the normality of their residuals and the homogeneity of their variances, respectively. The data met the assumptions and were then submitted to analysis of variance (ANOVA). Treatment means were compared using Tukeýs test (P < 0.05). All analyses were conducted using the SAS software (SAS Institute, 2015).

3.1 Chemical analysis

After obtaining TagCE through the leaf rinse process, its tagitinin C content was determined using the developed and validated reverse-phase HPLC analytical method, whose validation parameters are presented herein. In this sense, the linear calibration curve obtained by plotting the area ratio of tagitinin C to that of the internal standard (CA) against the standard concentrations, and a correlation coefficient of 0.9998 was achieved. The linear regression equation (y = 0.0070x + 0.0029) and the values of LOD (2.8670 µg·mL⁻¹) and LOQ (8.6879 µg·mL⁻¹) indicate that the developed method is highly sensitive and adequate to determine the tagitinin C content in TagCE. Fig. 2 displays a comparison between typical HPLC-DAD chromatograms obtained for the sample (TagCE plus CA) and for the analytical standard (tagitinin C plus CA).

Close

Fig. 2

HPLC-DAD chromatograms showing (A) TagCE and p-coumaric acid (CA; internal standard) and (B) the analytical standard (tagitinin C) and p-coumaric acid (CA; internal standard) peaks, obtained in the established analytical conditions (isocratic mobile phase system of 45% acetonitrile in Milli Q water containing 0.1% acetic acid, mobile phase flow rate of 1 mL·min⁻¹, column temperature of 40 °C, injected sample volume of 20 μL and detection at 254 nm).

Export to PPT

Intra-day precision was determined by analysis of tagitinin C in six samples of TagCE, and inter-day precision was assayed on two consecutive days. The obtained concentrations were respectively 98.724 µg·mL⁻¹ and 101.763 µg·mL⁻¹, which corresponded to about 11.36% of tagitinin C content in TagCE, thus confirming the previously reported high contents of this compound in T. diversifolia and leading us to propose this compound as the chemical marker of TagCE (Ambrosio et al., 2008; Baruah et al. 1979; Orsomando et al., 2016; Sut et al., 2018). The RSD values for tagitinin C concentrations were bellow 5%, and the means obtained for different days did not differ statistically (p < 0.05). To determine the accuracy of the method and after the addition of known contents of tagitinin C (25, 50 and 100%), the RSD values for all concentrations obtained, in triplicate, were also below 5%, indicating that the results were adequate for the analytical purpose of the developed method. Moreover, the robustness of the method was assessed as previously described by Moreira et al. (2013) and satisfactorily met the requirements of the main Brazilian regulatory agency (Anvisa, 2003).

3.2 Feeding deterrence

In the feeding deterrence bioassay, spraying the extract on the artificial diet caused larval mortality of 50.0% and pupal mortality of 70.0 and 80.0% at concentrations of 500 and 1000 mg mL⁻¹, respectively (Table 2). Positive control using flubendiamide insecticide showed 100.0% mortality of individuals tested within 24 h.

The T. diversifolia extract showed feeding deterrence activity for H. armigera larvae in the six host plants studied. The concentrations of 500 and 1000 mg/mL presented higher levels of feeding deterrence, ranging from 0.5 to 1.0, and higher leaf dry mass due to the lower leaf consumption at these concentrations (Table 3).