Nematophin, an Antimicrobial Dipeptide Compound From Xenorhabdus nematophila YL001 as a Potent Biopesticide for Rhizoctonia solani Control

This study was conducted to purify and identify metabolites of antimicrobial activity against phytopathogens from Xenorhabdus nematophila YL001. Three dipeptide compounds were purified from its cell-free cultural broth and identified as (±)-nematophin, cyclo (L-Pro-Gly), and N, N′-dimethyl-cyclo (L-Phe-L-Leu). Nematophin demonstrated a wider antifungal spectrum than the other two compounds. It also exhibited strong inhibitory effects on mycelial growth of Rhizoctonia solani and Phytophthora infestans with EC50 values of 40.00 and 51.25 μg/ml, respectively. Its (S)-configuration structure [(+)-nematophin] was also synthesized and exhibited higher antimicrobial activity than the enantiomeric mixture. The detached leaf assay revealed that nematophin possessed significant preventive and curative efficacy against R. solani on broad bean leaves showing corresponding control efficacies of 93.01 and 94.93% at 1,000 μg/ml, comparable to those of a chemical fungicide (carbendazim) at 500 μg/ml. Additionally, the pot experiments indicated that nematophin could effectively inhibit the disease extension on rice and broad bean plants caused by R. solani. Nematophin also exerted some adverse influences on the sclerotial development of R. solani by dramatically suppressing their formation and maturation at 40.00 μg/ml, as well as their germination at 15.00 μg/ml. Morphological and ultrastructural observations showed that the hyphae of R. solani became twisted, shriveled, and deformed at the growing points after exposure to nematophin at 40.00 μg/ml, and that the subcellular fractions also became abnormal concurrently, especially the mitochondrial structure. These results indicate that nematophin has great potential to be used as a bio-pesticide in agricultural production.


INTRODUCTION
Xenorhabdus nematophila, a well-studied genus of entomopathogenic bacteria, lives in symbiosis with Steinernema nematodes. Previous research has revealed that X. nematophila is a potent producer of natural compounds with versatile biological activities including antifungal, antibacterial, and antimalarial activities (Crawford et al., 2012;Challinor and Bode, 2015;Engel et al., 2017;Dreyer et al., 2018). Many of these metabolites belong to peptides such as xenocoumacins, nematophin, xenorhabdins, rhabduscins, and xenorxides (Li et al., 1997a,b;Ji et al., 2004;Böszörményi et al., 2009;Helge, 2011;Yang et al., 2011). Such peptides are of diverse structures and resultant multiple bioactivities including antimicrobial activity against plant pathogens (Li et al., 1997a;Yang et al., 2011;Dreyer et al., 2018). Besides, whole genome programs have shown that approximately 7.5% of genomic genes encode proteins involving the secondary metabolism biosynthesis in the strain X. nematophila ATCC19061, most of which encoded molecules are cryptic (Chaston et al., 2011). Given that the culture supernatants of many Xenorhabdus species exhibited potent antimicrobial activity against many plant pathogens (Fang et al., 2014;Hazir et al., 2016;Sharma et al., 2016), it is feasible to discover novel secondary metabolites of antifungal activity from X. nematophila as bio-pesticides for agricultural production.
Plant diseases caused by fungi are destructive attacks on crop production, generating enormous economic losses worldwide. Rhizoctonia solani is a necrotrophic fungal pathogen that infects more than 200 plant species worldwide, resulting in severe losses in crop yield (Lewis and Lumsden, 2001;Zheng et al., 2013;Mayo et al., 2015;. It can also produce sclerotia, a special resting structure surviving in soil for many years, which is the main source of infection in the disease cycle (Townsend and Willetts, 1954;Feng et al., 2016;Moni et al., 2016). Currently, the most common strategy for disease management is the application of chemical fungicides. Their overuse, however, has posed serious threats to human health, environmental safety and ecological balance (Nicolopoulou-Stamati et al., 2016;Rohr et al., 2017). Bio-pesticides are effective alternatives to the chemical fungicides to overcome their adverse impacts (Kumar, 2018;Lengai and Muthomi, 2018). Moreover, searching for bioactive compounds from microbes for control of the difficult-tocontrol plant pathogens is becoming a promising lead in the development of novel antimicrobial agents for agriculture production (Kim and Hwang, 2007;Mnif and Ghribi, 2015;Kanagaraj Muthu-Pandian et al., 2018).
Previous studies revealed that the cell-free culture of X. nematophila YL001 exhibited potent antimicrobial activity against some plant pathogens in vitro and in vivo, including P. infestans and B. cinerea (Guo et al., 2017;Zhang et al., 2018). However, no purified bioactive compounds were obtained from this bacterial strain, although many metabolites of novel structures and strong antimicrobial activity were purified and identified from strains of genus Xenorhabdus (Dreyer et al., 2018) including odilorhabdins (Pantel et al., 2018), cabanillasin (Houard et al., 2013), and bicornutin-A (Böszörményi et al., 2009). These studies mainly focused on their medical applications as antimicrobial agents. Currently, there is still limited information available about the antimicrobial potency of the metabolites of X. nematophila against phytopathogens (Li et al., 1997a;Ji et al., 2004;Lang et al., 2008;Gualtieri et al., 2009;Dreyer et al., 2018), especially the in vivo antimicrobial data (Yang et al., 2011;. In this paper, we tried to isolate and identify metabolites of antimicrobial potency against agricultural pathogenic fungi and oomycetes from X. nematophila YL001 and determine their antimicrobial activity in vitro. Subsequently, we evaluated the in vivo control efficacy of (±)-nematophin against R. solani. The effects of nematophin on the sclerotial development of R. solani, as well as on the hyphal morphology and ultrastructure, were also observed in this paper. The results revealed that N,N′-dimethylcyclo (L-Phe-L-Leu) possessed efficient antifungal activity against Exserohilum turcicum, and that nematophin could be used as a potential candidate for bio-pesticides.
The tested strains of eight phytopathogenic fungi (R. solani, E. turcicum, Fusarium graminearum, Verticillium dahlia, Botrytis cinerea, Sclerotinia sclerotiorum, Alternaria alternata, and Gaeumannomyces graminis) and two fungal-like oomycetes (Phytophthora infestans and Phytophthora capsici) were collected from fields in different regions of China which were identified and preserved in the laboratory of the Research and Development Center of Biorational Pesticides, Northwest A&F University, Yangling, Shaanxi, China. The information of their collection places and host plants can be seen in Supplementary Table S3. Each fungal or oomycete strain was activated by growing on the petri dishes containing potato-dextrose agar (PDA) culture medium for 2 or 3 days at 26°C before use.

Microbial Fermentation
A single colony of cells of X. nematophila YL001 was cultured in a 500-ml flask containing 200 ml of fresh Luria-Bertani medium (LB: 1.0% Bacto tryptone, 0.5% yeast extract, and 1% NaCl in water; pH 7.2) for 12 h at 28°C with shaking at 180 rev/min. Then, 3.5 L of the culture broth (OD 600 = 0.8) was transferred as a seed into a 70-L fermenter (Eastbio, China) Frontiers in Microbiology | www.frontiersin.org 3 August 2019 | Volume 10 | Article 1765 containing 35 L of TSB medium (per liter of distilled water: peptone from soymeal 3 g, peptone from casein 17 g, NaCl 5 g, glucose 2.5 g, and K 2 HPO 4 2.5 g; pH 7.2) and the system was incubated at 28°C with continuous agitation of 160 rpm and aeration of 0.25 v/v/min. After 48-h culture, the broth of X. nematophila YL001 was centrifuged (10,000 g, 20 min, 4°C) to remove the bacterial cells. Then, the cell-free broth was concentrated to 7 L on a rotary evaporator at 35°C and stored at 4°C for use.
Structural identification of the three metabolites was performed on the basis of spectroscopic analysis. The mass spectra (MS) of the three compounds were obtained by a Mariner Mass 5,304 instrument (California, USA) or an AB Sciex TripleTOF 5,600 + System (Framingham, MA). 1 H and 13 C nuclear magnetic resonance (NMR) data were acquired on a Bruker AVANCE III 500 spectrometer (Rheinstetten, Germany) with tetramethylsilane as the internal standard. The optical rotation data were recorded on an Anton Paar MCP 300 polarimeter (Graz, Austria). The circular dichroism (CD) spectra were collected on an Applied Photophysics Chirascan spectropolarimeter (Leatherhead, UK).

Effects of Three Isolated Compounds on Mycelial Growth of the Phytopathogenic Fungi and Oomycetes in vitro
The antimicrobial activity of three isolated compounds against the mycelial growth of 10 plant pathogens was performed as described in previous research with slight modifications (Zhang et al., 2018). Briefly, the three compounds were dissolved in dimethyl sulfoxide (DMSO) to prepare their respective stock solutions with a concentration of 10,000 μg/ml. A stock solution (150 μl) was mixed with 15 ml of molten PDA medium at a low temperature (about 40°C) to make a final drug concentration of 100 μg/ml (1% DMSO), and then the mixture was immediately poured into a petri dish (60-mm diameter) to form a plate of 2-3 mm thickness. PDA supplemented with the same volume of DMSO (1%) served as the blank control. Carbendazim (MBC) and mancozeb (MZ) at a concentration of 0.5 μg/ml were used as positive controls. After cooling, a 5-mm diameter mycelial disc was placed in the center of each plate with the inoculum side down. The plates were incubated at 28°C in dark until the mycelia reached the edges of the control dishes.
Each experiment was repeated three times. The mycelia growth inhibition rate was calculated using the following formula: where D c and D t represent the mycelial growth diameter of the control and treatment group, respectively. The diameter of the plug is 0.5 cm.
The antimicrobial activity of NEP-2 and -3 against five plant pathogens (R. solani, S. sclerotiorum, P. infestans, F. graminearum, and P. capsici) was also evaluated using the method as described above. The final concentration of NEP-2 or -3 in each test medium was 100 μg/ml and PDA plates supplemented with 1% DMSO were used as the controls.
The half-maximal effective concentration (effective dose for 50% inhibition, EC 50 ) was derived from analysis of the concentration-dependent inhibition rates. PDA plates with serial concentrations of NEP-1 (2.5, 5, 10, 20, 40, 60, 80, and 100 μg/ml) were used for test by the method described above. All experiments were independently performed three times under the same conditions.

Antimicrobial Activity Assay of NEP-1 Against R. solani and P. infestans by a Detached Leaf Method
The efficacies of NEP-1 on broad bean leaves infected with R. solani and potato leaves infected with P. infestans were evaluated according to the studies (Arfaoui et al., 2018) with some modifications. Broad bean (Vicia faba L.) was grown in a growth chamber (25 ± 1°C, 75 ± 10% RH, and a 12:12 LD photoperiod) for 21 days. Potato tuber was grown for 30 days under the same conditions. NEP-1 solutions of 500 and 1,000 μg/ml were prepared by dissolving it in an aqueous solution (sterile water composed of 0.5% tween 20, v/v). MBC solution of 500 μg/ml in the same mixture was used as the positive control against R. solani and MZ solution of 500 μg/ml against P. infestans. Sterile water containing 0.5% tween 20 was used as the blank control. All experiments were arranged in a plant growth chamber (25 ± 1°C, 75 ± 10% RH, and a 12:12 LD photoperiod) with nine leaves per treatment and were repeated three times.
For the protective activity assay, nine leaves of broad bean or potato were as a group and sprayed with a pre-prepared sample solution (1 ml for each leaf). After incubation for 12 h, a 5-mm diameter mycelial disc was inoculated onto the center of each leaf that was wounded at the inoculation site using a sterilized needle (avoiding the main vein of the leaves) in advance. After incubation for another 2 days, the lesion area was quantified from the captured images with Image J 1.38x software 1 , and the control efficacy was also calculated according to the following formula: where A c and A t represent the disease area of the blank control and treatment group, respectively. The area of the mycelial disc is 0.25 cm 2 . For the curative activity assay, nine leaves of broad bean or potato were as a group and each leaf was inoculated with a diameter mycelial disc as described above. After incubation for 12 h (~3 mm lesion around each disc), nine leaves of a group were sprayed with a pre-prepared sample solution of NEP-1 (500 or 1,000 μg/ml), MBC (500 μg/ml) or MZ (500 μg/ml) (1 ml for each leaf). The subsequent operations were as described above.

R. solani on Rice and Broad Bean Plants Grown in a Greenhouse
A pot experiment was performed to assess the in vivo biocontrol efficiency of NEP-1 against R. solani. Potted rice plants (35 days after seeding) and broad bean plants (30 days after seeding) grown in a greenhouse were used as the host plants. The temperature of the greenhouse ranged from 28°C (±2°C) in the daytime to 22°C (±2°C) during nighttime, and the relative humidity was controlled at 85% (±5%). The procedures of fungal inoculum, pathogen inoculation, and disease evaluation were according to the reported literature (Park et al., 2008;Wu et al., 2019) with some modifications. Briefly, 10 agar blocks (0.5-cm diameter) from the outer edge of a 3-day-old inoculated PDA plate were inoculated in a 250-ml Erlenmeyer flask with 100 ml of PDB (potato dextrose broth) medium. After incubation on a shaker (160 rpm) under darkness for 7 days, liquid cultured mycelia were harvested and cut into small mycelial balls (0.1 g in wet weight). A mycelial ball was placed beneath the leaf sheath and covered with aluminum foil immediately. When typical lesions appeared after 3 days, the aluminum foil was removed. Then the plants of a group were sprayed with a sample solution of NEP-1 (500 or 1,000 μg/ml), MBC (500 μg/ml), or sterile water. Three pots with more than five rice plants of each and nine pots with 1 https://imagej.nih.gov/ij/ one broad bean plant of each were used as a group, respectively. Each experiment was performed with three replications. After 4-day incubation, the lesion length in each stem of the inoculated plants was recorded.

Effects of NEP-1 on the Sclerotial Development of R. solani
The effects of NEP-1 on sclerotial formation were performed according to a previously published procedure (Soltani et al., 2017). Briefly, serial concentrations of NEP-1 in DMSO were added to PDA medium at final concentrations of 15, 20, 30, and 40 μg/ml, respectively. PDA supplemented with DMSO (1%) served as the control. A 5-mm diameter plug of R. solani was inoculated onto the center of each plate maintained at 28°C for 6 days until sclerotia manifested in the control plates. Additionally, the dynamic developmental process of sclerotia treated with 40 μg/ml of NEP-1 was also observed at fixed time intervals of 3, 4, 6, and 12 days. All plates were photographed with a Nikon D500 camera (Tokyo, Japan), and the sclerotia phenotypes were observed by a LECIA M165 FC stereoscopic microscope (Leica-ULTRACUT, Wetzlar, Germany).
The effects of NEP-1 on the sclerotial germination were measured according to a previously described method (Kazempour, 2004;Soltani et al., 2017). Uniform and healthy sclerotia were collected from PDA plates and soaked in solutions of various concentrations of NEP-1 (15, 20, and 30 μg/ml) with 1% DMSO for 5 min. Then, the sclerotia treated with NEP-1 at a certain concentration were placed onto PDA plates (one sclerotia per plate) and maintained at 28°C to observe their germination. The sclerotia treated with 1% DMSO were used as the controls. Each concentration was tested with 12 sclerotia and each test was repeated three times.

Effects of NEP-1 on Hyphal Morphology and Ultrastructure of R. solani
Determination of the effects of NEP-1 on the hyphal morphology and ultrastructure was conducted according to the method of Soner Soylu (Soylu et al., 2010). A prepared mycelial agar disc from a 2-day-old culture was inoculated in the center of the PDA plate with 40.00 μg/ml (EC 50 ) of NEP-1, and the plate was incubated at 28°C for 2 days in the dark. PDA plates without NEP-1 treatment were used as the controls.
For the scanning electron microscopy (SEM) observations, mycelial discs (2 mm × 4 mm × 4 mm) were fixed with 2.5% glutaraldehyde in a phosphate buffer (0.1 M, pH = 7.2) overnight at 4°C. Each sample was washed three times (20 min each) with 0.1 M phosphate buffer (pH = 7.2) to remove the excess glutaraldehyde. After fixation, the samples were dehydrated in a graded ethanol series (twice at 30, 50, 70, 80, and 90% and three times at 100%, v/v) for 20 min in each solution. After dehydration, the samples were dipped into isoamylacetate three times (20 min each) for replacement of ethanol. Finally, each sample was dried at the critical point using supercritical carbon dioxide. All samples were fixed on a holder using double-sided tape and coated with gold using an E1010 sputter coating machine (Hitachi, Tokyo, Japan) for 90 s at 9 mA. Samples were then imaged using a JSM-6360LV SEM instrument (JEOL, Tokyo, Japan).
For the transmission electron microscopy (TEM) observations, the pre-fixed mycelia were the same as described in the SEM method. After washing with a phosphate buffer (0.1 M, pH = 7.2) three times, the samples were post-fixed with 1% osmic acid for 2 h. Then, the samples were washed again with the phosphate buffer (0.1 M, pH 7.2) immediately followed by dehydration as described in the SEM method. Samples were dipped into ethanol twice (30 min each), after which the specimens were passed through a solution of epoxy resin/epoxy propane (1:1, v/v) for 1 h and embedded in epoxy media at 55°C for 48 h. Blocks were sectioned using a diamond knife with an Ultramicrotome (Leica-ULTRACUT, Wetzlar, Germany) into ultrathin sections of approximately 70 nm. The ultrathin sections were contrasted with 2% uranyl acetate and 2% lead citrate for 30 min prior to examination on a JEM-1230 TEM (Hitachi, Tokyo, Japan). At least three samples from each of the treated and control groups were examined by SEM and TEM.

Data Processing and Analysis
All datasets were analyzed by ANOVA using Statistica Software (Statsoft, Tulsa, OK, USA). The EC 50 was calculated by linear regression of the log of the colony diameter versus the various NEP-1 concentrations. When the ANOVA was significant, means were separated with the least significant difference test (LSD, p < 0.05).
Their circular dichroism (CD) spectra were recorded with the results shown in Supplementary Figure S2. An obvious positive absorption band was observed from 300 to 400 nm for the NEP-3 solution. Interestingly, no absorption signals were detected in this wavelength range for both NEP-1 and NEP-2 solutions. Moreover, different from NEP-1 and NEP-2 of racemic property, NEP-3 also exhibited a specific rotation a [ ] D 25 value of 31.66 (c 0.58) in CHCl 3 . These findings indicate that NEP-1 is an enantiomeric mixture.

Examination of Inhibitory Effects of Three Isolated Compounds on Mycelia Growth of Agricultural Pathogenic Fungi and Oomycetes in vitro
A previous study revealed that NEP-1 had strong in vitro bioactivity against a series of fungal and bacterial species (Li et al., 1997a). Piperazine-2,5-diones including PDKP and MDKP are an important class of cyclodipeptide compounds of versatile bioactivities (Borthwick, 2012). However, scarce work described their antimicrobial activity against plant pathogens. Thus, to further explore their potential as bio-pesticides, we evaluated their inhibitory effects on mycelial growth of 10 common phytopathogens in vitro with results listed in Table 1. NEP-1 showed a broad antimicrobial spectrum and exhibited high inhibitory effects against R. solani, P. infestans, and F. graminearum, with inhibition rates of 82.74, 80.41, and 80.93%, respectively. R. solani was most sensitive to NEP-1. Moreover, the EC 50 values of NEP-1 against R. solani and P. infestans were determined to be 40 and 51.25 μg/ml, respectively ( Table 2). Noteworthily, NEP-1 also exhibited higher inhibition rates at 100 μg/ml than MBC at 0.5 μg/ml against the selected pathogens except F. graminearum, P. capsica, and S. sclerotiorum. Similarly, compared to the antimicrobial activity of MZ at 0.5 μg/ml, higher inhibition rates were observed for NEP-1 at 100 μg/ml against the tested pathogens except E. turcicum and G. gramini. PDKP displayed a moderate inhibitory effect against G. graminis (inhibition rate of 35%) and weak inhibitory effects against R. solani, P. infestans, S. sclerotiorum, and A. alternate (inhibition rate < 40%) at 100 μg/ml. MDKP, however, remarkably inhibited the mycelial growth of E. turcicum with an inhibition rate of 67.60% at 100 μg/ml, higher than that of NEP-1. These results indicate that NEP-1 is an excellent antifungal agent with a broad spectrum and that MDKP has the potential to control the maize leaf spot disease caused by E. turcicum.
To explore whether the configuration of nematophin affects its antimicrobial activity, we tested the inhibition of three nematophin compounds (NEP-1, -2, and -3) on mycelia growth of five plant pathogens. As revealed in Table 3, compared to NEP-1 and -2, NEP-3 exhibited stronger inhibition against the five selected pathogens and significantly higher inhibition rates against S. sclerotiorum and P. capsica. Besides, no obvious differences were observed between NEP-1 and -2 in the inhibition rates. These results reveal that the (S)-configuration structure may possess higher antimicrobial activity than the (R)-enantiomer of nematophin and that NEP-1 is an enantiomeric mixture. Given that NEP-1 can be easily obtained through the bacterial fermentation of low cost, it was selected as the subject for subsequent experiments.
Examination the Control Efficacies of NEP-1 Against R. solani and P. infestans on Their Respective Host Plant Leaves As described above, NEP-1 exhibited strong in vitro antimicrobial activity against R. solani and P. infestans. However, the antimicrobial activity in vivo was scarcely reported in previous studies. Thus, we explored the control efficacy of NEP-1 against R. solani on broad bean leaves and against P. infestans on potato leaves. As shown in Figures 3A,C, NEP-1 could protect the bean leaves from the infection of R. solani effectively, with preventive efficacies of 53.05% at 500 μg/ml and 93.01% at 1,000 μg/ml, which is comparable to that of the MBC treatment at 500 μg/ml (98.01%). Moreover, NEP-1 also exhibited high curative activity for the diseased leaves infected by R. solani, with control efficacies of 83.59 and 94.93% at 500 and 1,000 μg/ml, respectively (Figures 3B,D).
The control efficacy of NEP-1 against P. infestans on potato leaves is shown in Supplementary Figure S3. The results showed that NEP-1 possessed weak preventive efficacy, with control efficacies of 15.58% at 500 μg/ml and 36.56% at 1,000 μg/ml (Supplementary Figures S3A,C), lower than that of MZ treatment at 500 μg/ml. Additionally, NEP-1 exhibited moderate curative activity with control efficacies of 39.63% at 500 μg/ml and 56.41% and 1,000 μg/ml, lower than that of referenced MZ at 500 μg/ml (Supplementary Figures S3B,D). Overall, the disease development control assay indicated that NEP-1 exhibited high control efficacy against R. solani in vivo. Therefore, R. solani was selected as the target pathogen for further experiments.

R. solani on Rice and Broad Bean Plants
To further explore the potency of NEP-1 as a new biopesticide, we evaluated the biocontrol potential of NEP-1 against the infestation of R. solani on rice and broad bean plants using a pot experiment. The morphological variations as well as the average lesion length of the inoculated plants in different treatment groups are illustrated in Figure 4. In the blank control groups, both of the rice and broad bean plants were severely infected by R. solani, with an apparent symptom of sheath blight disease (Figures 4C,D). However, the disease severity was effectively alleviated in the NEP-1 treatment groups (Figures 4C,D). Correspondingly, the lesion length on the inoculated plants in each NEP-1 treatment group was much smaller than that in the corresponding blank control group (Figures 4A,B).   Moreover, no obvious difference was observed in the lesion length on the rice plants between the NEP-1 treatment group (1,000 μg/ml) and the MBC treatment group (500 μg/ml) ( Figure 4B). These results indicated that NEP-1 had a favorable biocontrol efficacy against the disease extension on rice and broad bean plants caused by R. solani.

Effects of NEP-1 on the Sclerotial Development of R. solani
Suppressing sclerotial formation and germination is a crucial strategy for control of R. solani (Soltani et al., 2017). Thus, we tested the capacity of NEP-1 to suppress the sclerotial development of R. solani. As illustrated in Figure 5, on the plates with 40 μg/ml of NEP-1, few sclerotia were observed at the 6th day after inoculation and only a few abnormal sclerotium associated with large droplet secretion appeared after 12-day incubation. We further tested the effects of NEP-1 on the sclerotial development of R. solani at low concentrations (15, 20, and 30 μg/ml) with results shown in Supplementary Figure S4. Compared with the control plates, sclerotial formation was considerably suppressed with an increase of the NEP-1 concentration after 6 days of incubation. There were a very limited number of sclerotia occurring in the plates with 15.00 μg/ml of NEP-1 (Supplementary Figure S4B). Meanwhile, the sclerotial maturation was also prevented by NEP-1 and no mature sclerotia were observed when the concentration of NEP-1 was above 20.00 μg/ml (Supplementary Figure S4). Moreover, NEP-1 could also suppress the sclerotial germination significantly even at a very low concentration (15.00 μg/ml) NEP500, 500 μg/ml of NEP-1; NEP1000, 1,000 μg/ml of NEP-1; MBC, 500 μg/ml of carbendazim; and CK was the blank control. Scale bar: 2.5 cm.
( Figure 6). The above results revealed that NEP-1 could inhibit the sclerotial development of R. solani including the formation, maturation, and germination, and that NEP-1 might be an excellent antifungal agent for control of such sclerotia-forming fungal pathogens as R. solani.

Effects of NEP-1 on Hyphal Morphology and Ultrastructure of R. solani
To better understand the mode of anti-fungal action of NEP-1, the effects of NEP-1 on the hyphal morphology of R. solani were observed by SEM. As shown in Figure 7A, in the absence of NEP-1, the mycelia of R. solani exhibited a normal morphology of smooth, uniform, and robust hyphae with plump growing points. However, the NEP-1 treated sample displayed an altered hyphal morphology characterized by irregular hyphae of the twisted, shriveled and distorted morphology, or with deformities at the growing points (Figures 7B-D).
TEM was employed to evaluate the ultrastructural alterations of R. solani with the results illustrated in Figure 7. The control sample represented a typical fungal ultrastructure of intact cell walls of normal thicknesses, evenly distributed cellular cytoplasm, and regularly shaped organelles in the mycelial cells (Figures 7E,F). For the sample treated with NEP-1 (EC 50 of 40.00 μg/ml), the mitochondrial abnormalities were clearly observed of hazy outlines and vacuolar degeneration, as well as the reduction in number (Figures 7G,H). Furthermore, the vacuolization and disorganization of the cytoplasm were also found in the mycelial cells exposed to NEP-1 (Figures 7G,H).

DISCUSSION
X. nematophila YL001 is a valuable producer of natural compounds with pesticidal properties. Its fermentation broth was indicated to possess strong antimicrobial activity against some plant pathogens in vitro and in vivo, including P. infestans and B. cinerea (Guo et al., 2017;Zhang et al., 2018). Additionally, several pesticidal compounds have been identified from other X. nematophila stains, such as xenocoumacins (McInerney et al., 1991;Yang et al., 2011;, nematophin (Li et al., 1997a,b), PAX (peptide-antimicrobial-Xenorhabdus) peptides (Gualtieri et al., 2009;Fuchs et al., 2011), benzylideneacetone (Ji et al., 2004), xenortides, and xenematides (Lang et al., 2008). Xenocoumacins and nematophin were also present in the culture supernatant of X. nematophila YL001, which was confirmed by the HPLC-MS analysis in our previous study (Guo et al., 2017). Based on this, we try to isolate and identify new metabolites from the strain of X. nematophila YL001. In the present study, three dipeptide compounds were isolated and identified from the cell-free culture. NEP-1 is an antibiotic which was first isolated from strain X. nematophila BC1 (Li et al., 1997a). PDKP, a cyclodipeptide of proline and glycine, has been isolated previously from a sea marine sponge of undescribed species of Callyspongia (Chen et al., 2014), as well as many microorganisms such as Quambalaria cyanescens (Bishay et al., 2018) and Bacillus amyloliquefaciens (Li et al., 2018). MDKP, a dimethylated cyclodipeptide of phenylalanine and leucine, was first synthesized by Nakao et al. (2016) and its core structure piperazine-2,5-dione was found in a variety of natural products from fungi, bacteria, plants, and mammals (Borthwick, 2012). Despite this, to the best of our knowledge, it was first identified as a natural compound in this study. Moreover, this is also the first report of the presence of PDKP and MDKP in the X. nematophila fermentation broth. To gain more information about their pesticidal properties, we also evaluated their antimicrobial activity against 10 common phytopathogens. NEP-1 was demonstrated to possess a broad-spectrum antimicrobial activity in vitro ( Table 1) and high control efficacy against R. solani in vivo (Figures 3, 4). Despite the weak antimicrobial activity against the selected plant pathogens, PDKP was determined to be a potent acaricide against Tetranychus urticae with a LC 50 value of 95.96 μg/ml in a recent study (Li et al., 2018). MDKP was proved to have moderate antifungal activity against E. turcicum ( Table 1). These results reveal the potential value of the three compounds in the crop protection.
Previous studies revealed that the racemization of S-KMVA did not occur in vivo, and that L-alloisoleucine (L-allo-Ile) was derived from L-Ile rather than R-KMVA (Mamer and Reimer, 1992;Mamer and Lépine, 1996). Thus, the other pathway may be that NEP-1 is derived from the authigenic L-Ile and L-allo-Ile, which can be converted to S-KMVA and R-KMVA, respectively. Conversion of L-Ile to L-allo-Ile was verified in bacteria, fungi, plants and mammalian systems and L-allo-Ile is also present in many natural cyclic peptide antibiotics (Mamer, 2001;Li et al., 2016), providing evidence for the second pathway. However, the actual mechanism of the racemization needs further clarification. NEP-1 has great potential to be developed as a new biopesticide for agricultural production. NEP-1 not only represented broadspectrum antimicrobial activity against 10 agricultural pathogenic fungi and oomycetes in vitro ( Table 1) but also exerted a comparable control efficacy against R. solani in vivo at 1,000 μg/ml to MBC at 500 μg/ml on the rice plants (Figure 4). Moreover, NEP-1 could also inhibit the sclerotial formation, maturation and germination of R. solani substantially even at a low concentration (15 μg/ml) (Supplementary Figures S4, S5), beneficial to control the disease extension during its epidemic period and its outbreak in the next year effectively. Besides high bioactivity (in vitro and in vivo), easy production at an industrial scale is also another key factor affecting the development of a microbial pesticide (Montesinos, 2003). NEP-1 and its analogs can be synthesized through chemical methods due to its simple structure (Li et al., 1997b). In addition, researchers have also illuminated its biosynthesis pathway and achieved its heterologous production in Escherichia coli (Cai et al., 2017). In this paper, we also synthesized an enantiomeric mixture of nematophin (NEP-2) and its (S)-configuration structure (NEP-3), with high yields and easy operating conditions. The ultrastructure data suggest that NEP-1 may diffuse into the exposed cells and then disrupt the cells by altering the subcellular structures, especially the mitochondrial structure, which are similar to those induced by camptothecin (CPT) against R. solani, indicating that they may possess a similar antifungal mechanism FIGURE 8 | Two deductive pathways involved in the biosynthetic origin of (±)-nematophin (NEP-1). (Sirikantaramas et al., 2008;Dai et al., 2017). However, more work is needed to explore the effects of NEP-1 on the mitochondria-related pathways of R. solani to further clarify its mechanism of action as a potent antifungal agent.

CONCLUSIONS
Three dipeptide compounds were purified from the cultural broth and identified as (±)-nematophin (NEP-1), cyclo (L-Pro-Gly) and N,N′-dimethyl-cyclo (L-Phe-L-Leu). NEP-1 demonstrated a wider antimicrobial spectrum than the other two compounds. It exhibited strong inhibitory effects on mycelial growth of R. solani and P. infestans. Its (S)-configuration structure (NEP-3, (+)-nematophin) was also synthesized and exhibited higher antifungal activity than the enantiomeric mixture. Besides, NEP-1 not only possessed significant preventive and curative efficacy against R. solani on broad bean leaves, but also could effectively inhibit the disease extension on rice and broad bean plants. The sclerotial formation, maturation and germination of R. solani was also significantly inhibited by NEP-1 at even a low concentration (15.00 μg/ml). The electron microscopic observations showed that the mycelium morphology of R. solani was adversely affected by nematophin at 40.00 μg/ml, as well as the subcellular structures. These results indicate that NEP-1 has great potential to be used as a bio-pesticide in agricultural production.

DATA AVAILABILITY
The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.