Screening of Panamanian Plant Extracts for Pesticidal Properties and HPLC-Based Identification of Active Compounds

A library of 600 taxonomically diverse Panamanian plant extracts was screened for fungicidal, insecticidal, and herbicidal activities. A total of 19 active extracts were submitted to HPLC-based activity profiling, and extracts of Bocconia frutescens, Miconia affinis, Myrcia splendens, Combretum aff. laxum, and Erythroxylum macrophyllum were selected for the isolation of compounds. Chelerythrine (2), macarpine (3), dihydrosanguinarine (5), and arjunolic acid (8) showed moderate-to-good fungicidal activity. Myricetin-3-O-(6’’-O-galloyl)-β-galactopyranoside (13) showed moderate insecticidal activity, but no compound with herbicidal activity was identified.


Introduction
Plants and their extracts have been used for a long time for crop protection. They are a promising source for pesticides due to the fact that many plants produce secondary

Results and Discussion
A library of 600 extracts prepared from Panamanian plants was screened for fungicidal, insecticidal, and herbicidal activity. A total of 19 extracts fulfilled previously defined activity criteria, which were: a ratio of ≥ 0.75 for fungicidal, ≥ 50% activity against larvae and adult insects for insecticidal, and ≥ 50% (Agrostis stolonifera and Poa annua) or ≥ 80% (Matricaria inodora) for herbicidal activity (Tab. 1S, Supporting Information). A flow chart illustrating the further progression of samples is shown in Fig. 1. Active extracts were submitted to HPLC-based activity profiling [12,13], and collected micro-fractions were submitted to screening in the respective assays. Based on the above activity criteria, 12 extracts were prioritized. With the aid of chromatographic and activity profiles, five extracts were then selected for a detailed investigation. Among these, two extracts were chosen for their fungicidal (Fig. 2), one extract for its insecticidal (Fig. 3), and two extracts for herbicidal activity (Fig. 4).

Fig. 1.
Workflow for the discovery of agrochemicals from Panamanian plant extracts The methanolic extract of Bocconia frutescens (Papaveraceae) showed fungicidal activity against Magnaporte oryzae in time windows corresponding to major UV-absorbing peaks ( Fig. 2A). Two of the active fractions and one additional fraction also showed activity against other fungal strains (Fig 1S, Supporting Information). The two early-eluting main peaks were identified as sanguinarine (1) [14] and chelerythrine (2) [14] (Fig. 5). Compound 2 showed moderate activity against Botryotinia fuckeliana, M. oryzae, Phytophtora infestans, and Septoria tritici. The late-eluting active peaks were identified as oxysanguinarine (4) [15] and dihydrosanguinarine (5) [14]. Compound 4 showed no fungicidal activity, while 5 was active against M. oryzae, P. infestans, and S. tritici. Macarpine (3) [16] was in a microfraction active against P. infestans (Fig 1S B, Supporting Information). The purified compound showed good fungicidal activity against P. infestans and M. oryzae. With the exception of 4, the compounds had been previously reported from B. frutescens [17,18].  [19], and were not further pursued. Arjunolic acid (8) [21] was purified by normal phase flash chromatography, and its presence in the active fraction was confirmed by HPLC-DAD-ELSD. Compound 8 was active against M. oryzae and S. tritici. In previous studies [22,23], the fungicidal activity of arjunolic acid (8) in a mixture with asiatic acid was reported, while in the current study the activity of purified 8 was confirmed. Two additional compounds outside of the active time window were also isolated and identified as β-hydroxypropiovanillone (6) [24] and 3'-O-methyl-3,4-O,O-methyleneellagic acid (7) [25]. All compounds are reported for the first time from M. affinis, since no phytochemical studies have been conducted on this species before.
A broad hump in the chromatogram of the methanolic extract of Myrcia splendens (Myrtaceae) indicated the presence of tannins (Fig. 3). However, two distinct windows of insecticidal activity against Ceratitis capitata were seen between t R 7-10 min. After largescale extraction, peaks a and b depleted, and c even disappeared, while peak 15 was extremely enriched in the crude extract. Prior to HPLC purification, the extract was separated over polyamide yielding five tannin-depleted fractions ( Fig 2S, Supporting Information). From the first active time-window, compound 13 was isolated and identified as myricetin-3-O-(6''-O-galloyl)-β-galactopyranoside [26] (Fig. 5). The compound showed weak activity against C. capitata at 2500 ppm. From the second active time window, inactive myricitrin (15) [27] and quercitrin (16) [28] were isolated. Additional compounds isolated from fractions outside of the active time windows were gallic acid (12), myricetin-3-O-β-galactopyranoside (14) [29], and myricetin (17) [30]. Compound 15 had been previously reported from M. splendens [31], while the other compounds were new for the species.  (18) was obtained from PA5 and confirmed by spiking with a commercial sample. In addition, inactive compounds 12, 15, 16 were isolated, together with mearnsitrin (19) [36]. All compounds were new for C. aff. laxum, since no phytochemical data have been previously reported on this species.
The extract of Erythroxylum macrophyllum (Erythroxylaceae) showed distinct activity against post-emergent M. inodora, even though the broad hump in the HPLC chromatogram was indicative of tannins ( Fig 4B). In time windows t R 3-5 min and t R 15-16 min, the activity could not be correlated to a peak in the UV or MS traces. The extract was filtered over polyamide, and five tannin-depleted fractions were obtained ( Fig 4S, Supporting Information). Compounds in the active time windows were purified by HPLC, and identified as neochlorogenic acid (24) [37,38], protocatechuic acid (25) [39], [41], and rutin (28) [42]. However, none of the flavonoids showed activity in the herbicidal assay when tested as pure compounds. In addition, 16 and ombuin-3-O-βrutinoside (29) [41] were isolated. All compounds are reported here for the first time from E. macrophyllum. In total, four fungicidal and one weakly insecticidal natural product were discovered by means of HPLC-based activity profiling. In contrast, none of the compounds purified from active time windows of C. aff. laxum and E. macrophyllum showed herbicidal activity. The activity in these time windows may have been, at least in part, due to the presence of tannins. This might have been confirmed by a retest for activity of tannin-depleted extracts. The example of fungicidal compounds showed that the profiling approach could be efficiently used for discovery of bioactive compounds of possible agrochemical interest.

Fungicidal Assay
The activity against phytopathogenic fungi (Botryotinia fuckeliana, Magnaporthe oryzae, Phytophtora infestans, and Septoria tritici) could be demonstrated by the treatment of fungal spore suspensions and analysis of the growth in microplates using a robot system.
The tests were done in 96-well microtiter plates. Compounds were transferred as DMSO solutions into empty plates, followed by the addition of a spore suspension of the fungus of interest in a nutrient solution. Compounds were tested either in a single concentration, or as serial dilutions at 10 concentrations. Each plate contained eight solvent control wells and eight reference wells containing a known fungicide. The plates were incubated at 23°C and 90% relative humidity. Fungal growth was assessed by measuring the optical density at 620 nm, immediately after treatment, and 10 times in intervals of 15 hours. In order to calculate the activity of a compound at a given concentration, the optical density values of each measurement of a compound was compared with those of the control and the reference, giving results from 0 to 1, whereby higher values indicated higher activity. ED 50 values were calculated with the aid of the dilution series. A compound having an activity ratio ≥ 0.75, or an ED 50 ≤ 10 mg/l was considered as active.

Insecticidal Assay
Tested insect species were Anthonomus grandis, Heliothis virescens, Ceratitis capitata, Megoura viciae, and Myzus persicae. Insecticidal activity, either as a contact or systemic insecticide, against piercing/sucking insects (adults and offspring) was assessed in a test unit consisting of 24-well microtiter plates containing broad bean leaf disks. The compounds were formulated using a solution containing 75% v/v water and 25% v/v DMSO. Different concentrations of formulated compounds were sprayed onto the leaf disks at 2.5 µl, using a custom-built micro-atomizer. Two replicates were prepared. After application, leaf disks were air-dried, and 5-8 adult insects were placed onto the leaf disks placed into wells of a microtiter plate. Insects were then allowed to suck on the treated leaf disks, and were incubated at about 23 ± 1°C and about 50 ± 5% relative humidity for 5 days. Mortality was visually assessed.
Activity against biting insects (larvae) was evaluated in a test unit consisting of 24-well microtiter plates containing an insect diet and 20-30 insect eggs. Test compounds were formulated using a solution containing 75% v/v water and 25% v/v DMSO. Aliquots (20 µl) of different concentrations of formulated compounds were sprayed onto the insect diet using a custom-built micro-atomizer. Two replicates were used. After application, microtiter plates were incubated for 5 days at 23 ± 1°C and 50 ± 5% relative humidity. Egg and larval mortality was then visually assessed. Compounds with ≥ 50% mortality in adult insects and larvae were considered as active.

Herbicidal Assay
Herbicidal activity was assessed on pre-and post-emergent Matricaria inodora, Agrostis stolonifera, and Poa annua. The culture containers used were plastic 96-well plates containing peat substrate. For the post-emergence treatment, the test plants, once they reached a height of 1-3 cm (depending on the plant species), were sprayed via a spray nozzle with the test compounds in 1,000 ppm DMSO solution. The application rate corresponded to 2 kg/ha, with an application volume of 200 L/ha. Plants were kept at 20-35°C. The test period extended over 7 days. During this time, the plants were tended, and their response to the individual treatments was evaluated visually. The cutoff for herbicidal activity was ≥ 50% inhibition of growth (or 80% in the case of Matricaria inodora) of the treated weed, either pre-or post-emergence.