Chemoecological Study of Trophic Interaction Between Pseudosphinx Tetrio L. Larvae and Allamanda Cathartica L.

‒ The larval caterpillar Pseudosphinx


INTRODUCTION
Commonly known as "glutton caterpillar", Pseudosphinx tetrio L. (Linnaeus 1771) is a velvety black caterpillar (Lepidopterae) about 12 cm long, with bright yellow stripes along the body and a bright orange-red head (Haxaire and Herbin 2000;Haxaire and Rasplus 1986). P. tetrio has the ability to eat up to twice its weight in food during a day this fact explains the common name of "glutton caterpillar". According to aposematic hypothesis, it becomes toxic for its predators upon to absorb toxic compounds present in the latex of some plants (Warrell 2009). The larva of P. tetrio feeds preferentially on the leaves of plants as Allamanda cathartica L.
In the context of tropical chemical ecology, many different plant-herbivore interactions are highly specialized (Forister et al. 2015). In a self-defence process, plants can biosynthesize secondary metabolites to reduce herbivory. In turn, herbivores can react to these compounds in attempt to metabolize, excrete selectively or to use them for own defence (upon incorporation into the body) (Eisner and Meinwald 1995;Opitz and Muller 2009). Herein, we will explore the chemical interaction between P. tetrio and A. cathartica by several analytical methods, including an innovative electrochemical approach (named electrochemical ecology) and multivariate analysis in order to obtain better insights regard their specific interaction.

MATERIAL AND METHODS
Sample's preparation. Healthy and herbivored leaves of A. cathartica, as well as P. tetrio caterpillars have been collected in Guadeloupe Island (French West Indies), specifically in "Le Gosier" city (16°13'00.5"N 61°31'09.9"W). Leaves were cleaned (with distilled water) and lyophilized. The caterpillars have been kept in a cage for 24 hours, to collect their faeces, and starved for 48 hours to use their corpse in the evaluation of biomolecules incorporation to the caterpillar bodies excluding the digestive material. Finally, the dried leaves, the caterpillars' bodies and faeces were lyophilized and powdered.Afterwards, 50 g of each sample was extracted by maceration extraction for 48 hours. The maceration was carried out in a ternary mixture of dichloromethane/methanol/distilled water (Yu Lin et al 2007;Mushtaq et al 2014) (1:1:1 v/v, 200 mL). Upon the complete extraction, the solutions were filtered and extracted by liquid-liquid extraction and two phases were obtained for each manipulation (Table 1).Finally, these phases have been dried by rotary evaporation.

Table 1 Phases obtained for each manipulation after liquid-liquid extractions
Analysis and quantification. Analysis were performed using several techniques such as thin layer chromatography (TLC), nuclear magnetic resonance spectroscopy ( 1 H-NMR), high performance liquid chromatography coupled with mass spectrometry (HPLC-MS) and cyclic voltammetry (CV). Nuclear magnetic resonance spectroscopy ( 1 H-NMR). 15 mg of each extract was solubilized in 650 µL of DMSO and transferred to NMR tubes. 1D 1 H nuclear magnetic resonance spectra were recorded with a BRUKER Avance 300 MHz spectrometer equipped with a BBO probe and automatic tube changer. Chemical shifts (δ) are expressed in ppm relative to tetramethylsilane (TMS) taken as external reference and internal calibration is performed on the solvent signal. All spectra were processed using Topspin 2.1 software. The classical 1D proton with a 90° pulse width was performed. The spectra were acquired using 256 scans and 2 dummy scans of 32 K data points with a spectral width of 5411.255 Hz. Cyclic Voltammetry (CV). Cyclic voltammograms were obtained at 298±1 K in a conventional threeelectrode cell using a platinum wire auxiliary electrode and an Ag/AgCl (3M NaCl) reference electrode.
Measurements were carried out with CH I660 equipment using 0.10 M potassium phosphate buffer at pH 7.0 as a supporting electrolyte. The working electrode was prepared by evaporating 50 μL of an ethanol (EtOH) suspension of extract of interest, ground leaves orinsect samples under air on a glassy carbon electrode (GCE, BAS MF 2012, geometrical area 0.071 cm 2 ). In order to mimic the natural environment, no degasification of the electrolyte was performed.
Principal Component Analysis (PCA). The raw data from HPLC-MS was exported as netCDF files, using DataBridge software (Waters, USA), and pre-processed using XCMS Online (Mahieu et al 2016;Tautenhahn et al 2012) for feature detection, retention time correction and alignment of metabolites detected on HPLC-MS analysis. The dataset was created with 12 samples from each organic and aqueous extract fraction (3 samples from each manipulation (see Table 1)). Peak detection was performed using cent Wave peak detection (Δm/z = 10 ppm; minimum peak width, 5 s; maximum peak width, 20 s) and mzwid = 0.015, minfrac = 0.5, bw = 5 were used for retention time alignment. The processed data (csv file) was further exported to MetaboAnalyst 4.0 (Chong et al 2018). All data variables were scaled by pareto method prior to PCA. ).PCA is an unsupervised method commonly used to identify patterns between multivariate samples (Wold et al 1987;Jolliffe, 2005) and was recently employed on chemical variability of Allamanda cathartica extracts (Rodrigues et al 2020).

RESULTS AND DISCUSSION
TLC analysis. First, a TLC study has been conducted. The different results of this approach were exposed in the Table2. In our first hypothesis (Hyp. 1), we propose that the caterpillars ingest and store certain metabolites (for example the metabolite with a Rf of 0.16 on TLC 3 (orange color)). The second hypothesis proposed (Hyp. 2), also illustrated in Table 2, refers to a selective excretion of an unidentified molecule by the caterpillars (for example the metabolite with a Rf of 0.95 on TLCs 2 and 3 (blue color)). Finally, the third hypothesis (Hyp. 3), is related with the response of the healthy plant to the herbivory (for example the metabolite with a Rf of 0.28 on TLC 1 (pink color), as this metabolite is present in the healthy leaves and not in the herbivored ones. The opposite case can be observed with the metabolite with a Rf of 0.18 on TLC 4 (green color) that is only seen in healthy leaves and thus could be considered as a metabolic marker for herbivory. DMSO-d 6 as solvent. Solvents signals from deuterated DMSO and residual water at 2.5 and 3.3 ppm, respectively, were removed (Fig. 1). The analysis indicates minor differences between spectra of health and herbivored Allamanda leaves. This result is contradictory with regards our Hypothesis. 3 (were a reaction of the plant to aggression was attended).
The spectra derived from caterpillars' bodies and faeces present big differences between the samples, and the chemical profile of healthy Allamanda and caterpillar's faeces are very similar. This result confirms our Hypothesis 2, showing the similarity between the chemical profile of A. cathartica leaves and caterpillars' faeces when compared to the caterpillars' bodies. These similarities and dissimilarities confirm our hypothesis (Hyp. 3) of the selective excretion of toxic compounds.  The behaviour of films on glassy carbon electrodes of healthy and predated leaves of A. cathartica are showed in Fig.3. It was observed a set of three chemically irreversible oxidation processes A1, A2 and A3 (0.28 V, 0.65 V and an undefined shoulder near to 1.00 V respectively). Where as A1 and A2 peaks decrease upon cycling until a totally passivated and stable state is attained approximately at the second cycle, an increasing of A3 peak was observed. Thus, the variations in the ratio of peaks A1 and A3, can infer differences in the nature of the redox species in A. cathartica healthy and herbivored leaves. These oxidation processes represent a profile observed in several vegetable species and are attributed to the oxidation of polyphenolic organic compounds. In fact, the voltammogram of the predated leaves exhibits the same signals than healthy leaves, but the peak A1, is depleted while the signal A3 is enhanced. Finally, a reduction process approximately at -1,0 V can be observed when the potential sweep is reversed towards less positive potentials. Although these processes are slightly more negative for the healthy than for herbivore leaves, both correspond to the typical behaviour of the oxygen reduction in water. This result is basically opposed to our hypothesis 3 about the chemical response to the herbivory, but by this technique we can identify slight differences between healthy and herbivore leaves maybe related with polyphenols (reaction against oxidative stress) (Singh et al 2021). The cyclic voltammetry of films on glassy carbon electrodes of P. tetrio's bodies and its excrements after predation on leaves of P. tetrio are presented on Fig.4. The cyclic voltammogram does not show clear oxidation process in the initial anodic scan for the body of P. tetrio, but revealed three oxidation processes at 0.22, 0.45 and 0.78 V, for its faeces, namely A1, A2 and A3, respectively,which are similar to those found for A. cathartica analysis. Once again, a typical oxygen reduction process in water was observed incathodic region for both analysis. Further, a characteristic signal C1 appears in both samples suggesting the presence of oxidized metabolite of P. tetrio. This result confirms the hypothesis 2 in agreement with the other analytical approaches. A comparison between the voltammograms of films from the water extracts of healthy and predated leaves of A. cathartica in contact with air-saturated 0.10 M phosphate buffer at pH 7.0 was performed (Fig. 5). Here, the signal A1 appears clearly recorded in the healthy leaves and is accompanied by a well-defined cathodic peak at -1.25 V (C2). This signal C2 disappears in the voltammograms of the predated leaves while the signal A1 decreases and the signal A3 at 0.78 V becomes clearly marked. This result confirms the hypothesis 3, showing a reaction of the leaves against the herbivory, the disparition of signal C2 and the apparition of signal A3 as result of the herbivory.
This voltammetry is to some extent reproduced in the aqueous extracts of P. tetrio and its excrements, reproduced in Fig. 6, thus suggesting the existence of common electroactive compounds. PCA Analysis. A total of 24 chromatograms were acquired using HPLC-EI-MS, 12 from aqueous fraction and 12 from organic fraction (3 biological replicates for each species). To identify differences in the metabolites or even in their concentration in the samples, the set of 24 samples was analyzed obtained from XCMS and used for data analysis. Two spreadsheets were obtained after data processing of HPLC-MS analysis, with 168 and 412 variables from organic and aqueous fractions, respectively (retention time m/z). In the PCA analysis, two plots were generated, namely score and loading plots. The first one shows the samples grouping, whereas the second one indicates the contribution of each variable to these samples. The PCA analysis were focused on the detection of any inherent pattern within the data. As observed in Fig.7, samples of healthy and herbivored Allamanda leaves are close related, with small difference each other, when compared with the caterpillars' bodies and faeces. Further, when all samples are compared, the predated and not predated leaves from Allamanda are very similar to each other, and different from the other samples. In addition, the Caterpillars' bodies extracts present a great difference. In Fig. 8 (loadings plot) are presented the variables that influences the grouping observed.  In conclusion, we investigated the plant-herbivore interaction between caterpillar Pseudosphinx tetrio and the flowering plant Allamanda cathartica. In this approach, several techniques were employed in attempt to gain better insights regarding this trophic relationship, such as TLC, 1 H NMR, HPLC-MS (analyzed by a multivariate PCA) and an innovative approach by electrochemical methods (electrochemical ecology). The results point out the similar profile between the health and herbivored A. cathartica's leaves and the caterpillars' faeces. The similar analytical profile between the leaves of A. cathartica and the faeces of P. tetrio, as well as the difference with the caterpillar's bodies suggest a selective excretion of compounds by the caterpillar (proposed hypothesis 2). These compounds found selectively in the faeces (and not in the body) can explain that P. tetrio can feed this toxic Apocynaceae species.

DECLARATIONS.
Funding. The authors would like to acknowledge for financial support to the projet "AgroEcoDiv" from European Union Fund (FEDER) and "Région Guadeloupe". We would also like the Caribaea Initiative association for two month study grant from Matignon.

Conflicts of interest/Competing interests (include appropriate disclosures)
.The authors declare that they have no conflict of interest.
Availability of data and material. Not applicable.

Additional declarations for articles in life science journals that report the results of studies involving humans and/or animals ethics approval. Not applicable
Consent to participate (include appropriate statements). Not applicable.