Selective Toxicity of Secondary Metabolites from the Entomopathogenic Bacterium Photorhabdus luminescens sonorensis against Selected Plant Parasitic Nematodes of the Tylenchina Suborder

ABSTRACT Entomopathogenic Photorhabdus bacteria (Enterobacteriaceae: Gamma-proteobacteria), the natural symbionts of Heterorhabditis nematodes, are a rich source for the discovery of biologically active secondary metabolites (SMs). This study describes the isolation of three nematicidal SMs from in vitro culture supernatants of the Arizona-native Photorhabdus luminescens sonorensis strain Caborca by bioactivity-guided fractionation. Nuclear magnetic resonance spectroscopy and comparison to authentic synthetic standards identified these bioactive metabolites as trans-cinnamic acid (t-CA), (4E)-5-phenylpent-4-enoic acid (PPA), and indole. PPA and t-CA displayed potent, concentration-dependent nematicidal activities against the root-knot nematode (Meloidogyne incognita) and the citrus nematode (Tylenchulus semipenetrans), two economically and globally important plant parasitic nematodes (PPNs) that are ubiquitous in the United States. Southwest. Indole showed potent, concentration-dependent nematistatic activity by inducing the temporary rigid paralysis of the same targeted nematodes. While paralysis was persistent in the presence of indole, the nematodes recovered upon removal of the compound. All three SMs were found to be selective against the tested PPNs, exerting little effects on non-target species such as the bacteria-feeding nematode Caenorhabditis elegans or the entomopathogenic nematodes Steinernema carpocapsae, Heterorhabditis bacteriophora, and Hymenocallis sonorensis. Moreover, none of these SMs showed cytotoxicity against normal or neoplastic human cells. The combination of t-CA + PPA + indole had a synergistic nematicidal effect on both targeted PPNs. Two-component mixtures prepared from these SMs revealed complex, compound-, and nematode species-dependent interactions. These results justify further investigations into the chemical ecology of Photorhabdus SMs, and recommend t-CA, PPA and indole, alone or in combinations, as lead compounds for the development of selective and environmentally benign nematicides against the tested PPNs. IMPORTANCE Two phenylpropanoid and one alkaloid secondary metabolites were isolated and identified from culture filtrates of Photorhabdus l. sonorensis strain Caborca. The three identified metabolites showed selective nematicidal and/or nematistatic activities against two important plant parasitic nematodes, the root-knot nematode (Meloidogyne incognita) and the citrus nematode (Tylenchulus semipenetrans). The mixture of all three metabolites had a synergistic nematicidal effect on both targeted nematodes, while other combinations showed compound- and nematode-dependent interactions.

unbiased approach, and used bioassay-guided fractionation to isolate, purify, and structurally characterize bioactive SMs responsible for targeted bioactivities (31,32). For some SMs with a known structure, total chemical synthesis may replace the isolation of metabolites from the microorganisms. Such "synthetic SMs" may facilitate bioactivity evaluations, especially when the productivity of the microbial cultures is insufficient. Moreover, synthetic SMs may also offer a more facile source for structure-activity relationship studies, and a low-cost alternative for the commercial production of active ingredients for pharmaceuticals and agrochemicals, including nematicides.
Until now, only a few studies explored the targeted isolation of Photorhabdus SMs with activities against PPNs (33)(34)(35). Thus, 3,5-dihydroxy-4-isopropylstilbene from P. luminescens MD displayed nematicidal activity against the targeted plant-or fungal-feeding nematodes such as Bursaphelenchus and Aphelenchoides spp. However, it was also toxic at similar concentrations (up to 100 mg/ml) to Caenorhabditis elegans, a non-target bacterivorous nematode. More importantly, 3,5-dihydroxy-4-isopropylstilbene showed no activity against the most economically significant PPNs such as the root-knot nematode, Meloidogyne incognita (33,34). In contrast, indole, isolated from the same P. luminescens strain, provoked a transient paralysis on the tested nematode species, causing a nematicidal effect only at high concentrations. No results were reported for the effects of indole on C. elegans or entomopathogenic non-target nematodes (33,34). These pioneering studies also did not formally determine the LC 50 (lethal concentration causing 50% nematode death) or the EC 50 values of the compounds (effective concentration causing 50% nematode reversible paralysis) against the infective juvenile stages of the tested nematodes.
Previously, our team assessed the bioactivities of crude extracts (complex mixtures of SMs) of two Arizona-native Photorhabdus luminescens strains (Caborca and CH35) to reveal selective nematicidal activities against the infective juveniles of the targeted PPN species, the root knot nematode M. incognita (23). These same extracts displayed low to no nematicidal activity against non-target soil nematode species, such as C. elegans and the entomopathogenic nematodes Steinernema spp. In addition, the LC-MS fingerprints of these crude extracts indicated the presence of at least nine unique compounds not previously detected in other Photorhabdus species. Based on these results, we hypothesized that these Photorhabdus strains may produce novel SMs with selective nematicidal activity against economically important PPNs, and that certain combinations of these SMs may even result in synergistic effects.
The present study describes the isolation and structure elucidation of three SMs with nematicidal activities from the Arizona-native Photorhabdus spp. Caborca strain. Specifically, the nematicidal and/or nematistatic activities of the individual compounds and SM combinations were assessed against two targeted PPN species, M. incognita and T. semipenetrans, both ubiquitous in the southwestern U.S. The effects of the metabolites were also evaluated against a panel of non-target nematodes, including the free-living stages of bacteria-feeding and entomopathogenic species. The influence of these compounds on human cells and the assessment of the synergistic, additive, and antagonistic effects of the SM mixtures in vitro are also reported.
trans-cinnamic acid shows selective nematicidal activity against targeted PPNs. t-CA displayed a potent, concentration-dependent nematicidal activity against the targeted PPNs ( Fig. 2A). At 24 h postexposure, the LC 50 of t-CA was 67 mg/mL against M. incognita and 76 mg/ml against T. semipenetrans (Table 1). These LC 50 values were not statistically different, as shown by the non-overlapping 95% confidence intervals of the LC 50 values (Table 1) and corroborated by unpaired two-samples Wilcoxon tests (Table S4A).
Upon extended exposure to t-CA (24 h versus 48 h or 72 h), we observed a trend indicating that slightly lower t-CA concentrations may be required to kill 10%, 25%, 50%, and 90% of the two targeted nematode populations ( Fig. 2B and C; Fig. S1), including a 34% reduction of the LC 90 against M. incognita at 72 h versus 24 h (Table 2). However, these differences for the LC values at different time points did not reach statistical significance in most cases (Tables 2 and 3). A linear regression analysis and a Wilcoxon test further validated the observed non-significant effects of exposure times on t-CA potencies against the targeted PPNs ( Fig. S1; Tables S5 and S6), although a linear decrease of the LC 10 values was seen against M. incognita with a regression coefficient (b) of -.14 (P = 0.02; Fig. S1B) as confirmed by the Wilcoxon test (Table S5). In addition, no differences were observed in the nematicidal potencies of t-CA between the targeted PPNs upon 48 h and 72 h of in vitro exposure (Tables S4B and S4C).
t-CA exerted weak nematicidal activity against the selected non-target nematodes after 24 h of exposure at concentrations above 200 mg/mL ( Fig. 2A and Table 1). Mortality with one of the non-target entomopathogenic nematodes, S. carpocapsae, did not reach 50% even at the highest TCA concentration tested (500 mg/mL), so an LC 50 could not even be established ( Table 1). The LC 50 values against the other non-target entomopathogenic nematodes (H. bacteriophora and H. sonorensis) and the bacteria-feeding nematodes (C. elegans and Rhabditis blumi) ranged from 308 to 421 mg/mL, with the highest LC 50 value measured for H. sonorensis, the natural host of the t-CA producer P. l. sonorensis Caborca (Table 1). Thus, the LC 50 values of t-CA against the targeted PPNs at 24 h of exposure were significantly lower than those observed against the non-target nematodes (Table S4A), with this compound displaying a selectivity  index of $ 5-6 (LC 50 for the non-target nematodes over those for the targeted PPNs). Notably, the 48-h and 72-h LC 50 values of t-CA against the targeted PPNs were also significantly lower than the 24-h LC 50 of t-CA against the selected non-targeted nematodes (Tables S4B and S4C).  (4E)-5-phenylpent-4-enoic acid (PPA) shows selective nematicidal activity against targeted PPNs. PPA also exhibited potent, concentration-dependent nematicidal activity against the targeted PPNs (Fig. 3A). The LC 50 value of PPA for M. incognita (44 mg/mL) was significantly lower (a 33% reduction) than that observed for T. semipenetrans (66 mg/mL) at 24 h postexposure (Table 1; Table S4A). When considering extended exposure to this metabolite, a trend was perceivable in most cases for lower concentrations of PPA that were required to kill 10%, 25%, 50%, and 90% of the two target nematode populations after 48 h and 72 h of exposure ( Fig. 3B and C; Fig. S2). However, these trends did not reach statistical significance against either of the two targeted PPNs (Tables 2 and 3). A linear regression analysis and a Wilcoxon test led to the same conclusion when considering M. incognita as the target (Fig. S2A and  Table S5), while the LC 10 (Table S6). After being exposed to PPA from 48 h to 72 h, no significant differences in the nematicidal potencies of PPA were detected between the two targeted PPNs (Tables S4B and S4C). PPA showed weak nematicidal activity against the non-target, bacteria-feeding nematode species at 24 h of exposure at concentrations above 200 mg/mL ( Fig. 3A and Table 1). Thus, the LC 50 value of PPA against C. elegans was found to be 278 mg/mL, while that for R. blumi was significantly higher (346 mg/mL). In contrast, LC 50 values could not even be established against the selected entomopathogenic nematodes, because mortality did not reach 50% even at the highest PPA concentrations tested (400 mg/mL; Table 1). Thus, the 24-h LC 50 values of PPA against the targeted PPNs were significantly lower than those observed against the non-target nematodes (Table S4A), with this compound displaying a selectivity index of $ 6 to 8. Also, the 48-h and 72-h LC 50 values of PPA against the targeted PPNs were significantly lower than the 24-h LC 50 against the selected non-targeted nematodes (Tables S4B and S4C).
When considering extended indole exposure (up to 72 h), a trend was observed for increasing concentrations of indole to be necessary to sustain paralysis with increasing time. Judged by the non-overlapping 95% confidence intervals of the LC 50 values, these trends were not statistically significant for M. incognita (Table 2), while for T. semipenetrans, both the EC 50 and the EC 90 were significantly higher (84% and 79% increase, respectively) at 72 h compared with that observed at 24 h (Table 3). However, a linear regression analysis and a    Wilcoxon test suggest that these increasing trends are in fact significant for both of the targeted PPNs (Fig. S3 and Tables S5 and S6), with the exception of the EC 90 against M. incognita (b = 0.16; P = 0.13; Fig. S3A). Indeed, the use of non-overlapping 95% confidence intervals to determine the significance of trends is deemed potentially too conservative in some cases (42). After 48 h and 72 h of the indole exposure, the nematistatic potency of this compound did not vary between the targeted PPNs (Tables S4B and S4C).
Indole also induced temporary paralysis on the selected non-target, bacteria-feeding nematodes (C. elegans and R. blumi; Fig. 4A). The 24-h EC 50 of 37 mg/mL against R. blumi was significantly lower (a 77% reduction) than that observed against C. elegans (159 mg/mL) ( Table 1; Table S4A). In contrast, the selected non-target entomopathogenic nematode species were resistant to the nematistatic effects of indole in concentrations up to 400 mg/mL.
The 24-h EC 50 values of indole against the targeted PPNs were significantly lower than those observed against the non-target nematodes (Table S4A). There was also a significant reduction in the 24-h EC 50 values against the targeted PPNs when compared to those against the entomopathogenic nematodes alone (Table S4A), but the difference did not reach statistical significance against the bacteria-feeding nematodes (Table S4A). Indole displayed a selectivity index of 5-8 when compared to the selected non-target nematode species (7-11 to the entomopathogenic nematodes alone and 2-3 to the bacteria-feeding nematodes alone). These observations indicate that the nematistatic activity of indole is highly selective when contrasting the targeted PPNs with the selected entomopathogenic nematodes, but this selectivity is much more limited with respect to the bacteriovores used in this study. Noticeably, the 48-h and 72-h EC 50 values of indole against the targeted PPNs were still significantly lower than the 24-h EC 50 against the selected non-target nematodes or against the selected entomopathogenic nematodes alone (Tables S4B and S4C).
The effects of indole could also be escalated to a nematicidal outcome with much higher concentrations of the compound. Thus, indole displayed an LC 50 of 307 mg/ml against M. incognita at 24 h exposure, while the LC 50 against T. semipenetrans was significantly higher than that observed against M. incognita (388 mg/ml; Tables 5 and 6). Indole also showed nematicidal activity against C. elegans and R. blumi at concentrations of $ 300 mg/ml (data not shown).
Mixtures of t-CA, PPA and indole exhibit varied interactions against target PPNs. Different combinations of the three purified P. l. sonorensis SMs had different effects on the two targeted PPN species, ranging from antagonistic to synergistic (Tables 5 and 6). In M. incognita, the LC 50 of the t-CA 1 PPA combination was 22.9 mg/ml, with no statistically significant change upon the further addition of indole (LC 50 of 19.9 mg/mL for t-CA 1 PPA 1 indole) or upon the replacement of t-CA with indole (LC 50 of 40.0 mg/mL for PPA 1 indole; Table 5). The potencies of the TCA 1 PPA two-compound mixture and the t-CA 1 PPA 1 indole three-compound combination were significantly higher than those of indole or t-CA used as individual SMs ( Table 5; orthogonal contrast comparisons shown in Table S7). These two sets of compound mixtures also indicated synergistic effects against M. incognita, with an additive index (AI) of 0.17 (t-CA 1 PPA) and 0.24 (t-CA 1 PPA 1 indole) as calculated with the equation used for mixture toxicity evaluations (43). The PPA 1 indole mixture showed an additive effect (AI,-0.03), while the t-CA 1 indole combination was antagonistic (AI, 20.31; Table 5) against M. incognita.
In T. semipenetrans, the LC 50 of the t-CA 1 PPA 1 indole three-part combination (27.3 mg/mL) and that of the t-CA 1 PPA two-part combination (40.6 mg/mL) were significantly lower than those observed for each of the constituent metabolites ( Table 6). The threecompound mixture was also significantly more potent than the t-CA 1 indole or the PPA 1 indole two-part combinations ( Table 6; Table S8). Based on the equation used for the mixture toxicity evaluation (43), this three-component SM mixture had a synergistic effect against T. semipenetrans, with an AI of 0.19. The t-CA 1 PPA and the PPA 1 indole two-component mixtures exhibited weakly antagonistic effects (AI, 20.15), while the combination of t-CA 1 indole was strongly antagonistic (AI, 20.45; Table 6).
t-CA, PPA, and indole display no in vitro toxicity against human cells. None of the three isolated SMs from P. l. sonorensis Caborca displayed in vitro cytotoxicity against non-neoplastic human cells (HFF, foreskin cells), and against three human cancer cell lines (NCI-H460, non-small cell lung cancer; SF-268, central nervous system glioma; and MCF-7, breast cancer),

DISCUSSION
In this study, three secondary metabolites (SMs) with nematicidal activities were isolated from in vitro cultures of P. luminescens sonorensis strain Caborca, a bacterial symbiont of the Arizona-native entomopathogenic nematode Heterorhabditis sonorensis Caborca. The identities of the isolated metabolites as trans-cinnamic acid (t-CA), (4E)-5-phenylpent-4-enoic acid (PPA), and indole, respectively, were confirmed by 1H-NMR and 13 C-NMR spectroscopy (Tables S1 to S3) and comparison to authentic synthetic standards. While these SMs have previously been detected in cultures of other P. luminescens strains (31,(33)(34)(35)41), detailed investigations of the concentration-dependent nematicidal and/or nematistatic activities of these compounds and their mixtures against plant parasitic nematodes (PPNs) have not been reported.
We found that both t-CA and PPA display potent nematicidal activities against the J2 of two economically important PPN species, M. incognita and T. semipenetrans, with the 24-h LC 50 (Tables 1 to 3). The LC 50 values against the selected PPNs were significantly lower than those observed against the selected non-target nematode species (Tables 1 to 3; Tables S4A to S4C). While the nematicidal potency of t-CA did not differ significantly between the two targeted PPNs across observation times, PPA was 33% more potent against M. incognita than against T. semipenetrans at 24 h of in vitro exposure. However, when the incubation period was prolonged from 24 h to 48 or 72 h, the differences in the nematicidal potency of PPA against the target nematodes were no longer significant. Our investigation of the time dependence of the nematicidal effects of t-CA and PPA show that most mortality occurs within the first 24 h of exposure of T. semipenetrans and M. incognita (Tables 2 and 3; Fig. S1 and S2; Tables S5 and S6). This may indicate that both SMs are concentration-dependent nematicides that cause sudden-onset acute toxicity, although the mechanism of action of these SMs needs to be determined in further studies. Importantly, neither t-CA nor PPA had any effect, at concentrations corresponding to their LC 50 values in M. incognita or T. semipenetrans, against a panel of non-target nematodes including the bacteria-feeding species C. elegans and R. blumi, and the entomopathogenic nematodes H. bacteriophora, H. sonorensis, and S. carpocapsae ( Fig. 2A and 3A). t-CA and PPA elicited mortality only at high concentrations against these non-target nematodes (LC 50 values in the range of 278-421 mg/mL; Table 1). For PPA, LC 50 values could not be established for the entomopathogenic nematode species even at concentrations close to the water solubility limit of the compound (400 mg/mL). Moreover, neither t-CA nor PPA displayed in vitro cytotoxicity against human cells.
Isolation and structure elucidation of PPA as a constituent of the fermentation extracts of a Korean-native Photorhabdus luminescence [sic] strain was reported earlier (41). However, to the best of our knowledge no previous studies have investigated the activity of PPA against economically important PPNs. In contrast, t-CA has been shown to display nematistatic activity against the J2s of M. incognita and the potato cyst nematode Globodera pallida (44). Our results extend and refine these previous observations on t-CA and PPA by focusing on nematode mortality as the endpoint, and also reveal their selective nematicidal activity against both M. incognita and T. semipenetrans compared with a panel of non-target nematodes. Taken together, the potent and selective nematicidal activity of t-CA and PPA raises the possibility that these SMs may be suitable lead compounds to develop selective nematicides against these PPNs that cause multibillion dollar crop losses annually (3)(4)(5), and threaten food security worldwide (6,7).
With respect to indole, previous studies have noted nematicidal effects at high concentrations, and paralysis at lower concentrations against tested PPNs (34,45). In addition, Hu et al. (34) reported that the selected PPNs, Bursaphelenchus xylophilus and M. incognita, were more susceptible to indole than the non-target entomopathogenic nematodes such as Heterorhabditis spp. Here, we show that indole provokes nematistatic effects on the targeted PPNs T. semipenetrans and M. incognita, with EC 50 of 37 or 56 mg/mL, respectively (Table 1). While R. blumi was similarly sensitive (EC 50 of 37 mg/mL) to the nematistatic effects of indole, C. elegans, the other non-target bacteria-feeding nematode in our panel was much more resistant (EC 50 of 159 mg/mL; Table 1; Table S4). With concentration escalation, nematicidal effects could eventually be elicited against the targeted PPNs and the selected bacterivores, but only at very high concentrations (LC 50 of $ 300 mg/mL). Notably, indole displayed no significant nematistatic or nematicidal activities against any of the tested entomopathogenic nematodes even at concentrations near its limit of water solubility (Table 1), nor did it exert in vitro toxicity against human cells.
The onset of paralysis upon indole exposure was rapid, and this nematistatic effect was persistent against the studied PPNs up to 72 h, with only a very modest reduction in potency observed upon prolonged incubation with the compound (Tables 2 and 3; Fig. S3; Tables S5 and S6), attributable to degradation (45). Even at extended exposures, significantly lower EC 50 values were found against the targeted nematodes, compared to the 24-h EC 50 against the selected non-targeted nematodes. The increase of the EC 50 values against the targeted PPNs upon increased exposure periods, indicates that indole is predominantly a concentration-dependent nematistatic agent, where paralysis can only be sustained with increased indole concentrations. The temporary nature of the paralysis caused by indole was also noted for B. xylophilus and M. incognita (34). Correspondingly, our experiments showed that M. incognita recovered from indole-induced temporary paralysis when the metabolite was removed, with over 90% of M. incognita regaining baseline mobility during a recovery period of 24 h, regardless of the length of prior exposure (24 h to 72 h) to the metabolite ( Table 4).
The three nematicidal and/or nematistatic SMs investigated in this study are all readily produced by P. l. sonorensis Caborca under identical in vitro culture conditions. This raised the intriguing possibility that these SMs may also display synergistic interactions against the targeted species when they are used in combination. Here, we report the effects of two-or three-component mixtures of pure SMs on the targeted, economically important PPNs. Our findings reveal complex interactions that lead to different treatment outcomes based on the nematode species and/or the specific combinations of the metabolites used. Most importantly from a crop protection perspective, the three-component mixture (t-CA 1 PPA 1 indole) displayed a synergistic effect on both PPN species tested (Tables 5 and 6). However, most two-part compound combinations showed additive or weakly antagonistic effects except the t-CA1PPA mixture against M. incognita, with the t-CA 1 indole combination showing the highest rate of antagonism against the two PPNs tested (Tables 5 and 6). These results indicate that there is an idiosyncratic interplay between the nematodes under study and the complex biochemistry of the tested SMs. Future studies should focus on the mechanisms of action of these Photorhabdus metabolites and their interactions on multiple nematode species. These interactions also raise interesting questions on the chemical ecology of P. l. sonorensis SMs in multitrophic interactions during the symbiotic life cycle of the bacterium. Additional investigations should also address the in vivo performance, stability, and toxicity of these SMs in greenhouse, and eventually, field studies.
Conclusions. Developing more effective, safe, and ecologically sustainable chemical control agents against nematode pests is a priority area for crop protection science. PPA, t-CA and indole show potent, concentration-dependent nematicidal, and/or nematistatic activities against two highly destructive PPNs; display promising selectivity in relation to model non-target species such as selected bacteria-feeding and entomopathogenic nematodes; and elicit negligible in vitro toxicity against human cells. Together with their accessibility via chemical synthesis, these properties warrant further investigations into these Photorhabdus SMs, formulated separately or as their synergistic mixtures, as lead compounds for nematicide development.

MATERIALS AND METHODS
Bacterial strains and cultivation conditions. Photorhabdus luminescens sonorensis strain Caborca (46) was cultured on nutrient agar plates supplemented with 0.025% (wt/vol) bromothymol blue and 0.004% (wt/vol) 2,3,5-triphenyl tetrazolium chloride at 28°C in the dark (47,48). A single blue-green, 40 h to 44 h old bacterial colony was inoculated into 15 mL Luria-Bertani (LB) medium in a 50-mL flask and cultivated with shaking at 200 rpm overnight at 28°C in the dark. One-milliliter aliquots of the resulting preculture were transferred into fresh LB media (100 mL LB in 500-mL Erlenmeyer flasks), and the cultures were incubated with shaking at 200 rpm for 96 h at 28°C in the dark.
SM isolation and structure elucidation. P. l. sonorensis cultures (10 L total volume) were acidified with 6 M HCl to pH 4.0 and extracted twice with equal volumes of ethyl acetate with shaking at 200 rpm for 30 min each at 28°C in the dark. The mixtures were centrifuged for 10 min at 5,000 rpm at 4°C, the organic phases were collected and combined, dried with 1% (vol/vol) anhydrous sodium sulfate, filtered through filter papers, and concentrated in a rotary evaporator at 37 to 40°C. The resulting thick oily, dark brown residues (crude extracts) were transferred to glass vials with lids and stored at 4°C in the dark.
Six fractions that showed nematicidal activity were further purified by preparative HPLC (Delta Prep 4000 system with a PDA 996 detector, Waters Corporation; equipped with a Kromasil 100 C 18 reversedphase column, 250 mm Â 10 mm Â 5 mm, Sigma-Aldrich). This led to the isolation of three metabolites with nematicidal activities (t-CA and PPA were present in more than one silica gel fractions). The isolated metabolites were identified as trans-cinnamic acid (t-CA; Fig. 1 and Table S1), (4E)-5-phenylpent-4-enoic acid (PPA; Fig. 1 and Table S2), and indole ( Fig. 1 and Table S3). The chemical structures and purities of the isolated metabolites were determined by nuclear magnetic resonance (NMR) spectroscopy (one dimensional 1H-NMR and 13 C-NMR spectra), and liquid chromatography-mass spectrometry (LC-MS) to determine the molecular weight and the ion fragmentation patterns. Structure assignments were further confirmed by comparing the isolated SMs with chemically synthesized authentic standards from commercial sources ("synthetic SMs": t-CA and indole from Alfa Aesar, and PPA from Enamine). In addition, the nematicidal activities of the isolated SMs and the commercially available synthetic SMs were also compared and validated to be equipotent in causing mortality.
Nematode rearing. Only soil inhabiting (free-living or infective) stages of target and non-target nematode species were considered in the present study. All nematode cultures (except T. semipenetrans) were maintained in the Stock laboratory (University of Arizona). M. incognita was reared in planta using susceptible tomato plants (cv. Roma VF; seedlings with five or six true leaves; courtesy of Dr. J. Brown, University of Arizona). Eggs of M. incognita were extracted from infected tomato plant roots, and J2 were collected after hatching, following the procedures of Atamian et al. (50). T. semipenetrans was isolated from infested lemon orchards at the Yuma Agricultural Center, University of Arizona (Yuma, AZ, USA). Eggs of T. semipenetrans were extracted from infected lemon tree roots using the procedures of El-Borai et al. (51), and J2 were collected using the same procedure as that described for M. incognita.
The non-target species, C. elegans wild-type strain N2 Bristol (courtesy of Dr. G.L. Sutphin, University of Arizona) and Rhabditis blumi DF5010 (Caenorhabditis Genetics Center, University of MN, USA) were maintained in vitro in nematode growth medium (NGM) agar plates seeded with Escherichia coli (Migula) Castellani and Chalmers, strain OP-50, as described previously (52). Mixed juveniles and hermaphrodites (5 to 6 days old) were collected from the culture plates prior to the bioassays. Infective third-stage juveniles (IJs) of three nontarget entomopathogenic nematode species (H. sonorensis Caborca strain, H. bacteriophora TT01 strain, and Steinernema carpocapsae All strain) were also maintained by rearing them in vivo using wax moth larvae (Galleria mellonella) as the host, according to procedures described previously (53), and collected for bioassays within 2 to 4 days of their emergence from nematode-infected cadavers. Nematode population density (number of nematodes per ml) was standardized prior to setting up well-plate bioassays, by adjusting each nematode suspension to a concentration of 100 nematodes in 20 mL of distilled water.
Nematicidal assays. Twelve-well tissue culture plates were used as the experimental arena. For each of the tested nematode species, an inoculum of 100 nematodes, suspended in 20 mL distilled water, was added to each well. For t-CA (Alfa Aesar), 14 incremental concentrations were considered, ranging from 10 to 500 mg/ mL in distilled water, with a final volume of 1 mL/well (DMSO # 1%; vol/vol). For PPA (Enamine) and indole (Alfa Aesar), 13 incremental concentrations were tested ranging from 10 to 400 mg/mL in 1 mL of distilled water (DMSO # 1%; vol/vol). Due to the relative insolubility of PPA and indole in water, higher concentrations 400 mg/ mL than could not be considered. Distilled water with 1% DMSO was used as the negative control. To measure the combined nematicidal effects of the three SMs, equal concentrations of two or three SMs were mixed in 1:1 or 1:1:1 mixture. For example, a 20 mg/mL solution of a two-SM mixture contained 20 mg from metabolite A and 20 mg metabolite B, dissolved in 1 mL distilled water. Individual or combined SMs from 10 mg/mL to 200 mg/mL (in 20 mg/ml increments) were prepared in distilled water (DMSO # 1%; vol/vol) and used for the bioassays.
The well plates were covered with their lids and incubated at 25°C in the dark. For the targeted nematode species, nematicidal activity was recorded at 24 h, 48 h, and 72 h after initial exposure to the individual SMs. For the non-target species, data were only recorded at 24 h after initial exposure to the individual SMs. For SM mixtures, mortality was recorded 24 h after initial exposure of the targeted nematode species. Nematodes were probed with a needle to assess if they were alive, dead, or paralyzed. Typically, dead nematodes were straight and had a clear appearance due to cell necrosis, while paralyzed nematodes had a shrunken, wavy, curved, or rounded appearance. Those nematodes that resumed activity upon probing were considered to suffer from temporary paralysis (impaired motility). Paralyzed nematodes were rinsed at least three times in distilled water and held for an additional 24 h to check for their recovery from temporarily paralysis. To ascertain the LC 50 of indole, dilutions from 260 mg/mL to 400 mg/mL (in 20 mg/mL increments) were used (DMSO # 1%; vol/vol). Experiments were repeated at least three times per treatment for each concentration tested.
Cytotoxicity assays. The non-cancerous HFF cells (human foreskin cells) and three human cancer cell lines, NCI-H460 (non-small cell lung cancer), SF-268 (central nervous system glioma), and MCF-7 (breast cancer) were used to evaluate the in vitro cytotoxicity of the isolated Photorhabdus SMs. The MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) colorimetric method (54) was used with doxorubicin as the positive control, and DMSO as the negative control. Photorhabdus SMs were tested at 50, 100, and 200 mM concentrations, and the assays were repeated three times.
Data analysis. For the toxicity data visualization, the relationship between percent mortality or temporary paralysis and metabolite concentration were described by a sigmoidal curve, which was fitted by the logistical function using the statistical software package SigmaPlot for Windows, version 14.0 (Systat Software Inc.). The inflection point of the sigmoidal curve corresponds to the concentration that kills or immobilizes 50% of the test population (lethal concentration, LC 50 ; or effective concentration of temporary paralysis, EC 50 ). Concentration-response data of single metabolites or SM mixtures were pooled for each concentration and subjected to probit regression analysis using the Polo Plus program (LeOra Software). The LC or EC values, along with the corresponding 95% confidence intervals, were estimated for each single SM or SM mixture at each exposure time, depending on the experimental setting.
The selectivity index of a given SM was calculated by dividing the averages of the LC 50 or EC 50 values obtained against each of the selected non-target nematodes by the LC 50 or the EC 50 values observed for each of the targeted PPN tested.
For the analysis of SM combination effects, the statistical method described by DeLorenzo and Serrano (43) was considered: Where: S = sum of biological activity; A m = LC 50 for compound A in the mixture; A i = LC 50 for compound A measured alone (individual effect); B m = LC 50 for compound B in the mixture; B i = LC 50 for the individual effect of compound B. S values were used to calculate an additive index (AI). If S # 1.0, then AI = (1/S) 2 1.0. If S $ 1.0, then AI = S(-l) 1 1.0. An AI value with confidence intervals in the negative range indicates antagonistic toxicity; an AI value with confidence intervals greater than zero indicates synergistic toxicity; and an AI value with confidence intervals overlapping zero indicates additive toxicity. Statistically significant differences in LC or EC values among SMs at a given exposure time, within or among nematode species, were first interpreted based on non-overlapping 95% confidence intervals. In addition, averages of LC or EC values of each replicate from each single SM and mixtures were used to perform unpaired two-samples Wilcoxon tests via the NPAR1WAY procedure of SAS to compare and assess selective effects of each SM between the target and non-target nematodes species, within the target nematodes, or within the non-target nematode species. For the nematicidal or nematistatic time course studies with single SMs, data from the 24 h, 48 h, and 72 h exposures were used for a linear regression analysis with the REG procedure and unpaired two-samples Wilcoxon tests with the NPR1WAY procedure in SAS. To further evaluate the interaction effects among the seven different SM applications (single compounds and two-part or three-part compound mixtures), the LC 50 values were compared using one-way ANOVA with the GLM procedure of SAS, followed by an orthogonal contrasts test. All analyses were performed using SAS for Windows version 9.4 (SAS Institute Inc.).
Data availability. The data sets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.