Nitrogen and phosphorus supply strongly reduced the control efficacy of maize against sunflower broomrape

ABSTRACT Heavy infestation by the root parasitic sunflower broomrape over wide regions constitutes a major biological constraint to sunflower production. The use of maize as a trap crop to induce sunflower broomrape germination is considered a cost-effective method for controlling sunflower broomrape. This study was designed to examine the activity of maize to induce sunflower broomrape seeds germination under varying levels of nitrogen (N) and phosphorus (P) fertilization. Orthogonal experiments were conducted during two cropping seasons with four maize cultivars under four levels of N and P application. A significant reduction in sunflower broomrape germination induced by maize root and rhizosphere soil extracts upon N and P application was observed in both 2014 and 2015, with P having the strongest effect on the broomrape germination inducing activity of maize root extracts. Structural equation modelling analysis revealed that N and P supplementation not only directly affected the germination-inducing capacity towards sunflower broomrape of maize root extracts but also had an indirect effect through the impact on abscisic acid contents. This study may contribute to a better understanding of the connection between soil fertility and the production and release of germination stimulants in maize and that can lead to improved control strategies.


Introduction
Broomrapes (Orobanche and Phelipanche spp.) are holoparasites lacking chlorophyll and functional roots. Sunflower broomrape (Orobanche cumana Wallr.) is known to infect only sunflower (Helianthus annuus L.) and is among the most devastating pests of this crop. It is widespread from central Asia to southeastern Europe and has been introduced to China (Parker 2013). Sunflower broomrape is mainly found in large parts of north and northwest China, causing devastating damage to sunflower production in the infested areas (Shi et al. 2015). As an obligate parasite, the life cycle of sunflower broomrape is strongly cued by signals from host plants (Parker and Riches 1993;Fernandez-Aparicio et al. 2009). The complex life cycle of root parasites suggests that a potential solution to the related weed problems is to modify the germination stimulant levels perceived by the parasite quantitatively, qualitatively, or temporarily to uncouple it from the host cycles (Fernandez-Aparicio et al. 2011). In recent studies, the potent and specific strigolactone (SL) biosynthesis inhibitor, TIS108, was found to be applicable for the regulation of Striga germination and infestation (Ito et al. 2011). Uraguchi et al. (2018) developed a selective SL agonist, sphynolactone-7, which is effective for reducing Striga parasitism without impinging on host SL-related processes. However, due to the high reproductive ability and longevity of broomrape seeds (Daniel 1995;Rubiales et al. 2009), a reduction in broomrape soil-seed banks could promote control approaches and manage this parasite. Although trap crops, which are false hosts, can trigger seed germination, they are incompatible in the downstream infection process and can gradually depleted the broomrape seed bank (Hooper et al. 2009). A 30% reduction in the crenate broomrape (Orobanche crenata Forsk.) seed bank was observed after one trap crop cycle (Linke et al. 1993). As cereals are important for human consumption and are cultivated at high levels, the use of a cereal with a strong inductor potential for broomrape germination as a trap crop has received considerable attention (Lins et al. 2006;Ye et al. 2016).
Factors such as chemical signal composition, fertilizer rate, and cultivar differences, influence the induction capability of the host or trap crop on broomrape seeds. Two groups of hostgenerated small-molecule signals were found to stimulate the germination of broomrape seeds. SLs, including strigol, orobanchol, and Sorgomol were the main germination stimulants of Orobanch minor, Phelipanche aegyptiaca and others . The other one is dehydrocostus lactone, which was reported to be responsible for inducing the germination of sunflower broomrape from sunflower root exudates. Nitrogen (N) and phosphorus (P) starvation has been reported to dramatically promote strigolactone production in maize (Zea mays L.) . N or P fertilization is a direct and straightforward method for controlling parasitic weeds (Ayongwa et al. 2006). Considering that the exudate of a single plant contains various compound that influence stimulatory activity, changes in the germination of broomrape seeds in response to fertilization could arise from changes in the concentration of germination stimulants or from a change in their synergistic or antagonistic actions of these stimulants (Fernandez-Aparicio et al. 2009).
Maize is the most widely planted crop in China and is important for food security in the semiarid regions . Ma et al. (2013) demonstrated that some hybrid maize lines could be used as trap crops for the control of sunflower broomrape; however, farmers in Northwest China more commonly apply high amounts of N and P fertilizer, up to 180 kg N ha −1 and 92 kg total P ha −1, respectively, during the maize season to increase crop production (Dang et al. 2006;Vitousek et al. 2009). The effect of nutrient level on the synthesis and transportation of SLs, suggests that excess fertilization may reduce the control efficiency of maize on sunflower broomrape, with varying effects among maize cultivars. Thus, we evaluated the sunflower broomrape germination inducing activity of root and rhizosphere extracts of maize grown under different fertilizer regimes. Maize growth and certain physiological traits associated with stress responses were analyzed to determine a possible correlation between broomrape inducing activities and the physiological characteristics of maize. This study could refine the current knowledge and provide further data toward an integrated, biologically based program for sunflower broomrape control using maize as a trap crop.

Plant materials
Based on preliminary results (Ye et al. 2020), the level of germination induction of the maize cultivars selected in this study was as follows: Neidan 314 (C0) > Haoyu 19 (C1) > Zhongbeiheng 6 (C2) > Jinmancang 15 (C3). All of these cultivars are suitable for cultivation in Northwest China. The seeds were purchased from Sanqin Seed Company in Yangling, Shaanxi Province, China. Seeds of sunflower broomrape for germination assays were collected in 2013 from infested sunflower fields in the Xinjiang Uygur Autonomous Region. Dry seeds were stored in cloth bags in the dark at room temperature for further use. All seeds used in the experiment were uniformly surface-sterilized in 1% (w/w) NaClO for 1 min and then soaked in 75% (v/v) ethanol for 2 min. The surface-sterilized seeds were thoroughly rinsed with autoclaved distilled water and air dried on a clean bench.

Experimental design and implementation
The experiments were conducted at the Guyuan Ecological Station of the Institute of Soil and Water Conservation (35°99′N, 106°44′E) of the Chinese Academy of Sciences, located in Guyuan, Ningxia Hui Autonomous Region, China. The experiment was conducted in 2014 and was repeated in 2015. The flow chart of experimental design was shown in Figure S1. The experimental site was a typical semiarid agricultural region in northwestern China with average annual sunshine of 2057 to 2384 h. The maize was sown on 6 May 2014 and 1 May 2015 and harvested on 27 July 2014 and 31 July 2015. The temperature during the cropping season is shown in Figure S2. All plants were grown in a rain shelter with natural light and watered every other day. The soil used in this study was common dark loessial soil with a pH of 7.41, soil organic matter of 9.41 g kg −1 , total N of 0.28 g kg −1 , available P of 2.28 mg kg −1 and available K of 206 mg kg −1 . The sand, was brought from Guyuan, sieved using mediumsized natural river sand. Sand and soil were mixed in the ratio 1:1, before the experiment. The experiment used an orthogonal design under four levels of three factors (N, P, and maize cultivars). Considering the mineral contents in the tested soil, the N and P fertilizer application rates were selected based on data available in the literature (Liang 2012). Different weights of urea were thoroughly mixed with the soil, resulting in various N addition treatments: 0 (N0), 82.5 (N1), 165 (N2) and 330 (N3) mg kg −1 soil (based on N). Super phosphate was mixed the soil resulting in varing weights, resulting in the following P addition treatments: 0 (P0), 187.5 (P1), 375 (P2), and 750 (P3) mg kg −1 soil (based on P 2 O 5 ). An L 16 (4 5 ) orthogonal array was selected to arrange the experimental runs. Each row of the orthogonal matrix represented a run with a specific set of levels to be tested (Table 1). Five seeds of each cultivar were sown per rocket pot (20 × 25 cm), and the maize population was thinned to two seedlings per pot after emergence. The pots were randomly placed to avoid subjective bias. Each treatment had three replicates.

Growth measurements
On the sampling date, heights from the base of the stem to the apex of the second fully expanded leaf were recorded. Plant and rhizosphere soil samples were collected after the harvest. The shoot samples were oven dried to a constant weight at 80°C and weighed.

Measurements of antioxidant enzyme activities and ABA contents
To estimate POD (peroxidase) and PPO (polyphenol oxidase) activity, approximately 0.5 g of fresh leaf material (the second fully expanded leaf) and 0.5 g root sample was homogenized with a mortar and pestle and 0.05 M phosphate buffer (pH 5.8). The homogenate obtained was centrifuged at 9500 × g for 20 min, and the supernatant was used to determine enzyme activity. The above steps were performed at 4°C. PPO and POD activities were assayed following the procedure described by González et al. (1999). Specific activity was reported as the absorbance change per min per mg fresh weight (Abs min −1 g FW −1 ). Abscisic acid (ABA) content were quantified by ELISA with a Phytodetekt ABA Test Kit (Agdia/Linaris) according to the manufacturer's instructions. The pre-experiment identified that the ABA content of maize whole roots was too low to quantify; therefore, the apical growing parts (3-4 mm long) were used here. Pre-weighed root samples were homogenized in an 80% methanol-water solution and then shaken at 4°C overnight in the dark. The homogenates were centrifuged at 1520 × g for 10 min at 4°C. The supernatant was dried under N flow at 4°C. The dried residues were suspended in 1 mL 0.05 M phosphate buffer (pH 7.4) and then used directly for the ABA assay. Standard ABA was added to the immune reaction system for normalization of hormone levels.

Germination assay of sunflower broomrape
To evaluate the effect of maize on inducing sunflower broomrape germination, the root and rhizosphere soil were extracted for germination assays. Rhizosphere soil that was less than 5 mm away from maize roots was collected and frozen at −20°C until use. Subsequently, 5 g of the soil was mixed with 5 mL of distilled water and ultrasonically treated for 30 min at 25°C, 50,000 Hz, and 300 W in an ultrasonic cleaner (CS-500EII, Ningbo Jiangnan Instrument Factory, Ningbo, China) and filtered. The filtrates were referred to as the undiluted rhizosphere soil extracts (1.0 g mL −1 ). Root tissues were separated, freeze-dried, milled, and passed through a 0.35 mm sieve. Root samples (1 g) and 1 mL of distilled water were added to 1.5 mL centrifuge tubes, ultrasonically treated for 30 min and centrifuged at 3900 × g for 2 min. The supernatants were referred to as the undiluted root extracts (1.0 g mL −1 ). Seed conditioning was performed using 4 mL aliquots of distilled water applied to Petri dishes (9.0 cm diameter) lined with double filter paper. Glass fiber filter disks (GFFP, 8 mm diameter, Whatman GF/A) with 40-80 seeds were placed on the filter paper (Ye et al. 2017), and the Petri dishes were sealed with parafilm, incubated in the dark at 25°C for 4 days, and used for germination assays. The solutions of rhizosphere soil and root extracts were diluted 10-and 100fold with distilled water, and applied at final concentrations of 0.01, 0.1 and 1.0 g mL −1 . Aliquots (20 µL) of maize rhizosphere soil and root extracts were added to a glass fiber disk along with sunflower broomrape seeds in Petri dishes. The Petri dishes were sealed and incubated at 25°C. Germinated and non-germinated sunflower broomrape seeds were counted under a binocular dissecting microscope at 20 × magnification after two weeks to assess the germination rate. Sunflower broomrape seeds were treated with 1.0 mg L −1 strigol (provided by Prof. Binne Zwanenburg, Radboud University Nijmegen, the Netherlands.) and were used as positive controls; distilled water was used as a negative control.

Statistical analysis
The data were processed using Excel 2016 and SPSS (SPSS software, V19, SPSS Inc., Chicago, IL, USA). An orthogonal design with three replications was used in this study. To satisfy the assumptions of the ANOVA, the germination data were arcsine-transformed before analysis. Tukey's honest significant difference (HSD) test was used to compare the means. To determine which fertilization levels of N and P optimize sunflower broomrape germination induction by maize, quadratic regression models were fitted with each nutrient input. The factors N and P were denoted by X1 and X2, respectively, and the levels are listed in (Table S1). The graphical procedures of the response surface and contour charts were performed using Origin 9.0 (OriginLab Corporation, Northampton, MA 01060). Structural equation modeling (SEM) analysis was applied to investigate the direct and indirect effects of N and P fertilization on the germination inducing activity of maize on sunflower broomrape by modifying ABA contents. All variables were tested for linearity of direct relationships, since SEM accepts only linear regressions (Bollen 1989). The analysis was performed with AMOS (AMOS, v18, IBM, Chicago, IL, USA) and the model fit was assessed using the Chi-square statistic and its associated P value, as well as the comparative fit index (CFI), goodness-of-fit index (GFI) and the root mean square error of approximation (RMSEA).

Maize growth parameters
Among the 16 treatment runs we tested, the greatest seedling height (178.2 cm) and dry mass (57.9 g) was obtained in C2 with N at 165 mg kg −1 and P at 187.5 mg kg −1 (Table 1). In 2014 and 2015, fertilizer supply exhibited a significant effect on maize growth. Although the interaction effects of N and P fertilization on shoot mass were not significant (P> 0.05) in 2014, they were significant (P< 0.01) in 2015 (Table 2). In 2015, seedling height was significantly affected only by N application (Table 2), whereas supplementation with both compounds showed significant effects on plant height in 2014. In both growing seasons, the dry mass and seedling height of maize without fertilizer was significantly lower than that of maize across all other treatments. The average shoot mass of C1 was higher than that of the other cultivars in 2014, but there was no significant difference among the various maize cultivars in 2015. The shoot mass showed an increasing trend with N addition up to the N2 treatment, followed by followed by decreases at higher concentrations in both 2014 and 2015. All percentage data were arcsin transformed prior to analysis. Significant statistical differences are indicated by asterisks: ** P < 0.01; * P < 0.05; (Tukey's HSD).
According to the orthogonal methods, the highest average level corresponded to the optimal conditions. The optimum heights of the maize seedlings were obtained with 165 mg kg −1 N and 187.5 mg kg −1 P in 2014 and 165 mg kg −1 N and 375 mg kg −1 P in 2015. The superior treatments for the dry mass of shoots were achieved with 330 mg kg −1 N and 375 mg kg −1 P in 2014 and 165 mg kg −1 N and 375 mg kg −1 P in 2015. Nitrogen presented the highest range value in seedling height and dry mass, and we inferred that N exhibited the largest effect on maize growth.

Antioxidant enzyme activities of maize
The ANOVA indicated that the effect of maize cultivars on antioxidant enzyme activities was insignificant (P> 0.05) in 2014, but significant in 2015 (P < 0.05) ( Table 2). The average POD and PPO activities of different maize organs were significant lower in C3 than other maize cultivars, and showed an increasing trend as the N levels escalated (Table 3). Our results illustrated that the antioxidant enzyme activities were significantly affected by N supplementation in both years. P supplementation did not appear to alter the activity of POD and PPO in either year (Table 2). An exception was for leaf POD activity in 2014, where the POD activity of maize leaves significantly increased with P supply. In 2015, the interaction effects of N and P on maize POD and PPO activities were significantly (P < 0.05).
The results of the range analyses showed that N had the highest range value in all enzyme activities. Higher POD activity was observed in the roots than in the leaves of maize, with the greatest average POD activity level (0.81 Abs min −1 g FW −1 ) of maize roots occurring in 2014 with 330 mg kg −1 of N, which was double that (0.37 Abs min −1 g FW −1 ) without N supply (Table 3). PPO activity was significantly higher in the leaves than in the roots. In 2015, the highest average PPO activity of maize roots (0.103 Abs min −1 g FW −1 at 330 mg kg −1 of N) increased by 37.3% over that (0.075 Abs min −1 g FW −1 ) of the treatment without N. The average PPO activities among different N rates showed little variation. For instance, in 2014, the average leaf PPO activities of various N rates ranged from 0.152 to 0.299 Abs min −1 g FW −1 , with the greatest being approximately 2-fold higher than the least, and the leaf POD activities in 2014 revealed average values between 0.387 and 0.465 Abs min −1 g FW −1 , with the highest being 1.2-fold greater than the lowest.

ABA contents in maize roots
The ABA content of maize roots ranged from 553.7 to 822.7 in 2014, 429.48 to 762.9 ng g FW −1 in 2015. In both growing seasons, the largest ABA contents among the 16 treatments was obtained from C0 with no fertilizer. ANOVA results indicated that N and P supplementation significantly affected the ABA content of maize roots in both years. Maize cultivars and the interaction effect of N and P on maize ABA content were not significant in 2014, although they were significant in 2015. During the two experimental years, ABA content showed a decreasing trend as N and P levels increase, with the highest range value of ABA content found for P. The lowest average ABA content among different P supplies was obtained at the highest P rate (541.6 ng g FW −1 ), which decreased by 36.3% compared to that with no fertilizer supplementation (738.2 ng g FW −1 ). In addition, a strong decrease in ABA content was observed for both high N and P supply.

Stimulatory activities of maize on sunflower broomrape seed germination
The mean germination rate of sunflower broomrape seeds treated with 1.0 mg L −1 strigol was 88.2%, while distilled water did not induce sunflower broomrape seeds germination (Tables S2 and S3). The aqueous extracts of maize root and rhizosphere soil at the concentration of 1.0, 0.1 and 0.01 g mL −1 all induced sunflower broomrape germination. The germination rates in the bioassays ranged between 0 and 88.2% depending on the maize cultivar and nutrient supplementation rate (Tables S2 and S3). Extracts of maize rhizosphere soil and root tissues without fertilizer induced significantly higher  germination rates than those of other treatments in both 2014 and 2015. In 2014, the average germination rate (64.0%) of sunflower broomrape induced by maize with fertilizer supply represented a 27% decrease compared to the germination rate induced by root extracts of C0 with no fertilizer (88.2%).
Although the aqueous extracts of maize rhizosphere soil and roots induced sunflower broomrape germination, the limited degree at 0.01 g mL −1 did not allow us to determine the inducing capacity of maize with various treatments and their interactions. Therefore, only the germination rates at 1 g mL −1 and 0.1 g mL −1 of extracts were analyzed in following results. In both growing seasons, sunflower broomrape germination induced by maize root and rhizosphere soil extracts were all significantly affected by N and P supplementation. The effect of maize cultivar and the interaction effects of N and P on sunflower broomrape germination were not significant except the effect of maize root extracts in 2014. The germination rate induced by various maize cultivars was significantly different, and the germination rate was in the order of C0 > C1 > C2 > C3. The range analysis demonstrated that P exhibited the largest effect on the sunflower broomrape germination inducing capacity of maize. Response surface graphs were plotted with varying N and P rates over the experimental ranges (Figures 1 and 2). For maize root extracts at 0.1 g mL −1 , the relationship between P level and germination rates of sunflower broomrape was negatively linear in 2014. A quadratic regression between the N rate and germination rates of sunflower broomrape was obtained in 2014 (Figure 1(a), R 2 = 0.47, P < 0.01).

Relationships among N and P fertilization, growth and physiological traits, and stimulatory activities on sunflower broomrape germination by maize
The germination inducing activity of maize on sunflower broomrape was not correlated with maize growth characteristics and antioxidant enzyme activities (data not shown). However, the ABA content of maize roots was significantly positively correlated with broomrape germination inducing activity in both 2014 and 2015 (Figure 3). In 2014, the coefficient of determination (R 2 = 0.41) was higher for germination induced by root extracts than that from rhizosphere soil extracts (R 2 = 0.20). Similar results were obtained in 2015.
Thus, we hypothesized that N and P fertilization would affect the broomrape germination inducing activity of maize through ABA. The SEM model explained 78% and 52% of the variance in the germination inducing activity of maize rhizosphere soil and root, respectively (Figure 4). The pathway between fertilizer application and the broomrape germination inducing activity of maize roots was through ABA. Regarding the total effect, P application (λ = −0.4) was the strongest predictor of broomrape germination inducing activity in maize roots. However, in contrast to our expectation, SEM indicated that ABA content had no significant impact on the inducing activity of maize rhizosphere soil. Among the variables used in the SEM model, N application was the key factor (λ = −0.58), contributing to the germination inducing activity of maize rhizosphere soil on sunflower broomrape.

Discussion
Nitrogen and phosphorus are primary macronutrients that are critical in various plant growth and development process. The results of two successive cropping seasons (2014 and 2015) showed that fertilizer supply significantly affected growth traits (seedling height and shoot dry mass), physiological traits (antioxidant enzyme activities and ABA content), and the broomrape germination inducing capacity of maize (Table 2).

Growth and physiological traits of maize in response to N and P fertilization
Low amounts of fertilizer positively regulate plant growth, whereas excessive fertilizer application does not necessarily result in higher plant yields (Liang et al. 2014). In our experiment, the maximum shoot mass occurred with 165 mg kg −1 N and 187.5 mg kg −1 P in 2014 and 165 mg kg −1 N and 375 mg kg −1 P in 2015. Similar results were obtained for shoot dry mass, revealing that N or P supplementation can increase growth parameters to a certain extent, but a negative effect occurs at higher levels. Studies have reported both positive and negative effects of fertilizer application on subsequent seedling growth and yield (Shah et al. 2017).
Metabolic processes in plants produce active oxygen species (Rios-Gonzalez et al. 2002). Thus, the maintenance of an active antioxidant enzyme system that scavenge the active oxygen species is critical for plant preservation (Pinhero et al. 1997). Some studies have found that high N and P levels elevated the antioxidant enzyme activity of submersed macrophytes (Zhang et al. 2010). In the present study, only N significantly increased antioxidant enzyme activity in different organs of maize, although increased POD activity under P-deficient conditions has been previously reported in maize (Tewari et al. 2004). The results of the present study showed the POD activity in maize roots and shoots was not significantly affected by P.
As important stress hormone, increased ABA in plant tissues has been reported subjected to nutrient deficiency (Radin and Ackerson 1981). In this study, a substantial accumulation of ABA in the apical parts of roots was observed in maize without fertilizer supplementation in both 2014 and 2015, and exogenously applied N and P had a decreasing effect on ABA accumulation. The interaction of N and P resulted in a significant decrease of ABA content in 2015, but no significant changes were observed in 2014. However, the influence of fertilizer on ABA accumulation in roots under stress is unclear. Schraut et al. (2005) showed that N and potassium deficiency caused ABA accumulation in roots, whereas a shortage of P and sulfur seemed to be ineffective. The results of this study indicated that maize growth and antioxidant enzyme activities are greatly dependent on N availability and that P has only a limited effect. However, in contrast to earlier findings, ABA content was more affected by P than N levels. This discrepancy might relate to differences in extraction ABA from the roots. An example of this is a study Vysotskaya et al. (2008), in which the ABA in whole durum wheat roots of nutrient-limited plants revealed no changes in comparison with well-fed plants, while a marked accumulation was observed in the apical 3-4 mm of roots.

Sunflower broomrape germination inducing capacity of maize in response to N and P fertilization
The differences in the efficiencies of various trap crops on broomrape were expected because of alternated taxonomies and environmental cues, but mainly due to the capacity to produce and exude germination stimulants (Babaei et al. 2010). In this study, root and rhizosphere soil extracts were used to assess this capacity in maize. Sunflower broomrape seeds were able to germinate in all treatments with added extracts. These results are consistent with data obtained by Ma et al. (2013) and suggest that maize can be a trap crop for sunflower broomrape. Prior studies have noted the significant difference among various plants or cultivars in their germination stimulatory activity for broomrape (Ma et al. 2012;Ye et al. 2016). The results of maize root extracts for 2014 showed a significant difference among cultivars, and germination ability was in the order of C0 > C1 > C2 > C3. However, a statistically significant difference was not found for maize rhizosphere soil extracts in the two experimental years or the root extracts in 2015. This study aimed to characterize the capacity of maize to induce sunflower broomrape germination in response to N and P supplementation. The increasing levels of nutrients showed a highly and P application on the allelopathic activity of maize root and rhizosphere soil on broomrape germination during the 2 years trial (Χ 2 = 7.669, df = 6, P = 0.263, CFI = 0.994, GFI = 0.978, RMSEA = 0.054). Solid line arrows represent significant paths (P < 0.05), dotted lines indicate non-significant. Arrow thickness represents the magnitude of the path coefficient. Double-sided gray arrow represents estimated covariance between two predictors. R 2 values represent the proportion of the variance explained for each endogenous variable. BGS, germination rate of sunflower broomrape induced by the extract of maize rhizosphere soil; BGR, germination rate of sunflower broomrape induced by maize root extracts. significant and negative effect on the germination of sunflower broomrape seeds induced by both maize root and rhizosphere soil extracts (Figure 2). These results support previous studies on Striga, which reported a reduction in germination stimulants secretion and Striga infestation with the application of mineral nutrients. The fitted curves illustrate the coupling effects of N and P on sunflower broomrape germination. The range analysis showed that the germination rate of sunflower broomrape induced by both maize roots and rhizosphere soil is highly sensitive to phosphate.

Exploring the direct and indirect influences of fertilization on sunflower broomrape germination inducing capacity of maize
The production and stimulation of germination stimulants could be directly and indirectly reduced by N and P application (Figures 1 and 2, Table S2 and S3), which is consistent with previous studies (Sato et al. 2003;Zhang et al. 2015). The main reason for this is the decreased production of strigolactone under improved soil fertility conditions (YoneyaYoneyama et al. 2013). A series of natural strigolactones have been identified and characterized from the root exudates of maize, with the typical examples as strigol, 5-deoxy-strigol, sorghumol, zealactone and zeapyranolactone (Siame et al. 1993;Awad et al. 2006;Charnikhova et al. 2017Charnikhova et al. , 2018. All of the SLs have a similar chemical structure and are derived from carotenoids by the action of carotenoid cleavage dioxygenases (Matusova et al. 2005;Alder et al. 2012). As SLs and ABA share biosynthetic origins from carotenoid precursors, the links between ABA and SLs under various stress conditions have been proposed (Tsuchiya andMcCourt 2009). Haider et al. (2018) reported that drought simultaneously induces ABA production, SLs production and the expression of SL biosynthetic genes (D27) in rice, suggesting a positive correlation between SL and ABA. We also observed a strong positive correlation between the ABA content of maize roots and the germination inducing capacity of the root and rhizosphere soil extracts in both 2014 and 2015 (Figure 3). In contrast to the above, Liu et al. (2015) found that SL biosynthesis was repressed in Lotus japonicus roots under osmotic stress and proposed that a SL decrease under osmotic stress is required to allow the increase of ABA in roots. These differences can be explained in part by stress severity, stress adaptation time, and the absence of the arbuscular mycorrhiza (Ren et al. 2018). Additionally, the SEM analysis in this study highlights that the decreased germination inducing activity of maize root extracts for sunflower broomrape is partly explained by a decrease in ABA concentration resulting from increased N and P supplementation ( Figure 4). However, the dependence of broomrape germination inducing activity of maize rhizosphere soil on ABA content was not significant. These results suggest the possible relationship between ABA contents and the potential of maize to induce suicidal germination of sunflower broomrape that significantly different responded to various environmental cues. Therefore, further studies on the relationship between SL and ABA during the induction of plant stress may be required to better understand the modes of action of germination stimulant production and secretion.

Conclusion
In summary, the results of two-year greenhouse experiment confirmed the potential of maize to induce suicidal germination of sunflower broomrape and suggested that this capacity of maize strongly depends on N and P supplementation. A low dose of N and P supplementation in maize strongly reduced the production and stimulation of germination stimulants. The germination rates of sunflower broomrape induced by maize root and rhizosphere soil extracts were correlated significantly and positively with the ABA content of maize root, with a similar response to fertilizer treatment. The pathway between fertilizer application and the broomrape germination inducing activity of maize roots through ABA was highlighted. These findings suggest that using maize without fertilizer could provide a biological control of sunflower broomrape by inducing its germination, thereby allowing for a combination of options toward an integrated, biologically controlled strategy to reduce damage by sunflower broomrape.

Disclosure statement
No potential conflict of interest was reported by the author(s).