Extracts from cotton over the whole growing season induce Orobanche cumana (sunflower broomrape) germination with significant cultivar interactions

Five cotton cultivars and their parents were tested for induction of germination of Orobanche cumana Wallr. (sunflower broomrape) seeds in pot and field experiments. Germination rates induced by cotton root extracts were the highest followed by stem extracts then leaf extracts. Cotton seedlings at the six-leaf stage induced higher germination than seedling at the twoand four-leaf stage, in all five cotton cultivars and their parents. In the field, the highest concentration of cotton root extracts gave the highest germination rate of O. cumana, and the lowest concentration of cotton root extracts gave the lowest germination rate. Methanol extracts of cotton rhizosphere soil gave the highest germination ofO. cumana. In general, the root, stem and leaf extracts were more active at the sixleaf stage than other seedling stages. In conclusion, extracts of cotton rhizosphere soil and tissues have high activity in the seedling stage. Extracts of cotton across the whole growing season were able to induce O. cumana germination but displayed significant cultivar interactions.


Introduction
Broomrapes (Orobanche spp.) are holoparasitic weeds that completely depend on their hosts for water and nutrients [1] . They heavily infest many important crops and have negative impacts on crop yield and quality, thus causing economic losses worldwide [2,3] . Of the 20 broomrape species in China, Phelipanche aegyptiaca Pers. (Egyptian broomrape), O. cumana Wallr. (Sunflower broomrape) and O. ramosa L. are the most common and have the widest host range [4] .
Suitable temperature and moisture conditions are required for host plants (such as sunflower) to grow, but these conditions alone are not sufficient to cause O. cumana seeds germination. This parasite has evolved to require a chemical signal from a host or non-host plant to induce germination. In the absence of these germination stimulants, the seeds will revert back to secondary dormancy [2] .
Cotton is a non-host plant that can induce germination of clover broomrape [5] . Botanga et al. showed that in planta production of the germination stimulant of Striga hermonthica (Del.) Benth in cotton was a qualitatively inherited trait, and the genes encoding this stimulant are monogenic and simply inherited [6] . It might therefore be possible to select and breed certain cotton genotypes that produce high concentrations of highly active germination stimulants, while maintaining or improving other agronomic attributes.
Efficient and economical control of Orobanche is extremely difficult because of the infested soil usually contains high seed reservoir densities. At present, mechanical and manual removal, pesticide spraying and other measures are used to control Orobanche, but these methods are costly and labor intensive. They also pose unwanted side effects of crop phytotoxicity and environmental pollution caused by chemical pesticide residues. Therefore, there is currently no consistent and sustainable method for controlling Orobanche anywhere in the world [2] .
Trap crops are non-host crops that induce germination of Orobanche, and consequently lead to seed reservoir attrition. Research has shown that many crops including carrot, cucumber, maize, onion, soybean and wheat can be used as trap crops for clover broomrape [7,8] . Wheat can be used as a trap crop with germination rates as high as 25% to 40%, and even up to 70% [9] . Control of parasitic weeds with trap crops is by far the most economical and practical method for small-scale commercial farming [10] . Our research group has conducted investigations on trap crops for Orobanche spp. including wheat [9,11] , maize [12] , cotton [13] , soybean [14] , rice [15] and switchgrass (Panicum virgatum) [16] . In this study, pot and field experiments were conducted to study the effects of rhizosphere soil, root, stem and leaf extracts of different cotton cultivars on the germination of O. cumana. The objective was to find a new method for using cotton as a trap crop for the biocontrol of O. cumana.

Source of seeds and chemicals
Seeds of the cotton cultivars were provided by the Cotton Institute Henan Academy of Agricultural Science. Strigol was provided by Prof. K. Yoneyama, Utsunomiya University, Japan and synthetic strigolactones GR24 by Prof. B. Zwanenburg, Radboud University, The Netherlands.

Seed surface sterilization
Cotton and O. cumana seeds were surface sterilized in 1% sodium hypochlorite for 3 min and then soaked in 75% (V/V) ethanol for 3 min. After thoroughly rinsing with sterile distilled water, seeds were air-dried.

Preconditioning of O. cumana seed
Five milliliter gibberellin (10 -4 mol$L -1 ) was added into a Petri dish (9 cm in diameter) containing two filter papers. Glass fiber filter disks (8 mm Whatman GF/A) were laid uniformly on the double filter paper and 20-50 O. cumana seeds placed on each disk. The Petri dishes were then sealed with Parafilm and incubated at 25°C for 6 d as described in Parker et al. [17] .

O. cumana germination assay
Aqueous solutions were assayed directly, by applying 20 μL aliquots of the respective test solution to conditioned O. cumana seeds on glass fiber filter disks in the Petri dishes. For solutions and extracts containing organic solvents, aliquots (20 μL each) of the test solution were applied to an 8 mm disk of glass fiber filter paper without seeds and allowing it to dry. Then a disk of conditioned O. cumana seed was placed on top of each and moistened with 40 μL of distilled water. After 10 d incubation, the germination rates were examined under a binocular dissecting microscope. An O. cumana seed was considered as germinated when the germ-tube protruded from the seed coat. GR24 was also used at 10 mg$L -1 on seed stock in separate assays to establish a standard proportion of O. cumana seeds that were responsive to germination stimulants [18] . A distilled water control was also included. Individual treatments were replicated three times unless otherwise mentioned.

Field and pot experiments
Five cotton cultivars and their parents (Table 1) . The loosely held soil was gently shaken off the roots, which was referred to as rhizosphere soil [19,20] and the soil as well as plant samples were sampled at each stage.
Pot experiments were conducted at the of Institute of Soil and Water Conservation, Northwest A&F University, Yangling, Shaanxi Province, China in May 2009. Eight kilogram of soil was put in each pot. Lou soil (Eum-orthic Anthrosols) was collected from the field at Institute of Soil Five grams of the rhizosphere soil [19,20] and 1.5 mL distilled water were added to Petri dishes (3.5 cm in diameter). Five disks of glass fiber filter paper (8 mm Whatman GF/A) with conditioned O. cumana seeds were put on the surface of the soil and the Petri dishes were sealed and incubated at 25°C for 10 d and subsequently examined for germination using a binocular dissecting microscope.

Methanol and distilled water extracts of cotton roots, stems and leaves on O. cumana germination
Cotton root, stem, and leaf were freeze-dried, milled and sieved (0.35 mm). Samples (100 mg) were weighed into 1.5 mL centrifuge tubes and 1 mL of methanol or distilled water added. The tubes were sonicated for 30 min then centrifuged at 6400 rpm for 2 min. The supernatant was collected, and used undiluted and at 10-and 100-fold dilutions in O. cumana germination tests.

Data analysis
The germination rates among different cultivars at each growth stage were subjected to an analysis of variance. Data processing was done with Excel 2007 and DPS 9.5. Tukey's honest significant difference (HSD) test was used to compare the means.

Results and discussion
GR24 (1 mg$L -1 ) gave the high germination rates (over 60%) in all experiments and distilled water did not induce germination.
In pot experiments, undiluted extracts did not induce any germination, but the 100-fold dilution gave the highest germination rate, higher than that induced by the 10-fold dilution ( Table 2). Root extracts from the three cotton seedling stages (two-, four-and six-leaf stages) all induced O. cumana germination. There were significant interactions with cultivar and their parents. Generally, extracts from cotton plants at the six-leaf stage induced the highest germination, followed by the four-leaf stage and the twoleaf stage. The highest germination rates reached at each stage were 76.5% for Male 5 at the six-leaf stage, 68.8% for Female 4 at the four-leaf stage, and 37.7% for Female 4 at the two-leaf stage. At least five cotton cultivars and/or parents induced over 60% germination at the four-and sixleaf stages ( Table 2).
There were similar trends for the methanol extracts of cotton stem and leaf. Also there were significant Table 2 Germination rates of sunflower broomrape induced by 10-and 100-fold methanol extracts of cotton roots at two-, four-and six-leaf stages in pot experiment unit: % interaction with cultivar and their parents. Basically, extracts of the six-leaf stage induced the highest germination, followed by the four-leaf stage and the two-leaf stage ( Table 3; Table 4).
The field experiment showed that extracts from the whole growth period of cotton, rhizosphere and tissues could induce O. cumana germination, but there were significant interactions with cotton cultivar (Tables 5-8).
The undiluted methanol extracts of rhizosphere soil induced the highest O. cumana germination compared to the 10-and 100-fold dilutions. The average germination rates were over 30% at squaring stage, with Male 1, Male 2 and Female 3 inducing over 35% germination at this stage. In contrast to the pot experiments, the germination induced by extracts collected at squaring stage were higher than germination for the four-leaf stage, followed by the twoleaf stage and the six-leaf stage ( Table 6).
The 100-fold diluted methanol extracts of rhizosphere soil gave the highest O. cumana germination rates compared to the undiluted and 10-fold dilutions. As for the pot experiments, the germination induced by the soils collected at the six-leaf stage were highest for the whole growth period, while the extract from the flowering-boll stage showed the lowest germination rates. The Male 2 cotton parent induced 81.0% germination at the six-leaf stage ( Table 6).
The 100-fold diluted methanol extracts of stem tissues gave the highest O. cumana germination rates compared to the undiluted and 10-fold dilutions. Among the five stages, the average germination rates induced by the two-leaf stage samples were the highest, while those from the floweringboll stage were the lowest. The Hybrid 5 cotton induced 67.7% germination at the two-leaf stage (Table 7).
Again, the 100-fold diluted methanol leaf extract gave the highest O. cumana germination rates compared to the undiluted and 10-fold dilutions. Among the five stages, the germination rates induced by the extracts from the two-leaf stage were highest, while the squaring stage gave the lowest germination rates. The Female 5 cotton parent induced 53.1% germination at the two-leaf stage ( Table 7).
The plants collected at flowering-boll cotton stage were divided into eight parts: root, upper stem xylem, lower part of stem xylem, upper stem phloem, and lower part of stem phloem, upper leaves, lower leaves, and flowering-boll. The methanol extracts of the tissues induced O. cumana germination. For nine cotton cultivars, extracts of the different organs gave the lowest germination rates. This demonstrated that the germination activity was weak at the flowering-boll stage. (Table 8) Our observations are consistent with those of Botanga et al. [6] that showed cotton genotypes significantly affect the induction of O. cumana germination. Integrated management of parasitic weeds like Orobanche using trap crops is needed to reduce the use of herbicides in agriculture. Trap cropping is done to induce the germination of parasitic weed seeds, while preventing the parasite from producing its own seeds. Orobanche produces a large number of seeds that have prolonged dormancy in the soil. A chemical stimulant is required to break the seed dormancy and induce germination. This chemical is Table 3 Germination rates of sunflower broomrape induced by 10-and 100-fold methanol extracts of cotton stem at two-, four-and six-leaf stages in pot experiments unit: % synthesized in planta and released as root exudates by hosts of the parasitic weed and other plants, and the latter can serve as trap crops. The primary consideration in parasitic weed management is the reduction of the parasitic weed seed reservoir in the soil [21] . The high heritability and simple inheritance of suicidal germination in S. hermonthica under cotton suggests that this cotton trait should be easily incorporated into cultivars with good agronomic attributes for use in Striga-infested areas [6] . It has been reported that the production of strigol by proso millet gradually increased, reaching a maximum of 25 pg per plant per day on days 5-7, and then decreased   to a constant level of 14 pg per plant per day [22] . Strigol is produced by cotton, and is a germination stimulant for the root parasitic plants, Striga and Orobanche, which was first isolated by Cook et al. [23,24] . Strigol was later identified in the root exudates of sorghum, maize and proso millet [Pennisetum glaucum (L.) R. Br.] [25] .
Recently, the quantification of strigolactones produced by cotton was reported, which confirmed that strigol was the main stimulant for the germination of Striga and Orobanche seeds [22] . Since the Striga germination stimulant strigol was isolated from cotton, and in our experiment we observed   [26] found the fast-neutron-mutagenized tomato mutant, SL-ORT1 was highly resistant to various Phelipanche and Orobanche spp., but no toxic activity or inhibition of Phelipanche germination was detected in the SL-ORT1 root extracts. They concluded that the SL-ORT1 resistance was due to its inability to produce and secrete natural germination stimulants into the rhizosphere [26] . This conclusion suggested that resistance to the parasites was due to the inability of the host plant to produce germination stimulants. Recently, it was reported that the resistance of sunflower to O. cumana might be associated with a hypersensitive reaction which was activated by exogenous salicylic acid treatment [27] . In contrast, in our experiments all the tested cotton cultivars were able to produce O. cumana germination stimulants, albeit with significant cultivar interactions ( Table 2; Table 5; Table 6). Importantly, however, it was observed that not all cotton cultivars were able to induce O. cumana germinate under the pot and field experiment conditions. Strigolactones, including strigol, can stimulate branching of arbuscular mycorrhizal fungi [28] and inhibit shoot branching in plants [29,30] , and so the consensus is that most of the strigolactone-producing plants (80% of all land vegetation) could be used as trap crops for broomrapes. In reality, the situation is far more complex. Broomrape germination induced by strigolactones is an allelopathic phenomenon. Therefore the strigolactones produced in planta have to be released into the soil in suitable quantities to reach concentrations able to induce germination of parasitic weeds under various environmental stresses. As we gain greater knowledge of allelopathy and trap cropping for broomrape control, new questions arise. Further research to understand the required concentration and 3-dimensional structure of strigolactones; how they are released into the environment; their stability in any given environment; how they are translocated; how they are degraded in soil or fixed in organic soil matter; the effect of the stage of host and/or non-host plant production; local soil temperature, moisture, and mineral conditions; and how strigolactones are adsorbed by broomrape seeds, is required. Answering these questions is of paramount importance. In addition, the criteria used to recommend crop cultivars to farmers as trap crops for broomrape is complicated by the fact that each crop have numerous cultivars, and many of them may not be able to induce broomrape germination. Transferring knowledge about trap crops and allelopathy for broomrape weed control to individual farmers remains a major challenge.

Conclusions
The pot experiment indicated that there were significant interactions with cotton cultivar. Specifically, germination induced by the samples collected at the six-leaf stage was the highest, followed by the four-leaf stage and the twoleaf stage. The field experiment showed that extracts over the whole growth period of the cotton, rhizosphere and tissues were capable of inducing O. cumana germination,