Strigolactones, host recognition signals for root parasitic plants and arbuscular mycorrhizal fungi, from Fabaceae plants

Summary

• • 
  Both root parasitic plants and arbuscular mycorrhizal (AM) fungi take advantage of strigolactones, released from plant roots as signal molecules in the initial communication with host plants, in order to commence parasitism and mutualism, respectively.
• • 
  In this study, strigolactones in root exudates from 12 Fabaceae plants, including hydroponically grown white lupin (Lupinus albus), a nonhost of AM fungi, were characterized by comparing retention times of germination stimulants on reverse-phase high-performance liquid chromatography (HPLC) with those of standards and by using tandem mass spectrometry (LC/MS/MS).
• • 
  All the plant species examined were found to exude known strigolactones, such as orobanchol, orobanchyl acetate, and 5-deoxystrigol, suggesting that these strigolactones are widely distributed in the Fabaceae. It should be noted that even the nonmycotrophic L. albus exuded orobanchol, orobanchyl acetate, 5-deoxystrigol, and novel germination stimulants.
• • 
  By contrast to the mycotrophic Fabaceae plant Trifolium pratense, in which phosphorus deficiency promoted strigolactone exudation, neither phosphorus nor nitrogen deficiency increased exudation of these strigolactones in L. albus. Therefore, the regulation of strigolactone production and/or exudation seems to be closely related to the nutrient acquisition strategy of the plants.

Introduction

The root parasitic plants Orobanche and Striga spp. are devastating pests in agricultural production throughout the world (Joel et al., 2007). These root parasites depend on host plants for nutrients and water and cannot survive without parasitizing hosts. Their tiny seeds contain limited resources so that the parasites must connect to hosts within a week of germination. The seeds of these parasites germinate only when they perceive host-derived chemicals, termed ‘germination stimulants’, released from plant roots. The first described germination stimulant for Striga, named strigol, was isolated from the root exudates of a false host, cotton (Gossypium hirsutum) (Cook et al., 1966, 1972), and later identified from genuine hosts, sorghum (Sorghum bicolor), maize (Zea mays), and proso millet (Pennisetum glaucum) (Siame et al., 1993). Subsequently, sorgolactone was isolated from S. bicolor root exudates (Hauck et al., 1992). Alectrol was purified from cowpea (Vigna unguiculata) root exudates (Müller et al., 1992) and recently identified as orobanchyl acetate (Xie et al., 2008a). The first described Orobanche germination stimulant, orobanchol, was isolated by Yokota et al. (1998) from red clover (Trifolium pratense) root exudates. Recently, 2′-epiorobanchol and solanacol were characterized from root exudates of tobacco (Nicotiana tabacum), a host of Phelipanche ramosa (formally called Orobanche ramosa) (Xie et al., 2007). In addition, two novel stimulants, sorgomol (Awad et al., 2006; Xie et al., 2008b) and a putative didehydro-orobanchol (strigol) isomer (Sato et al., 2003; Xie et al., 2007), were identified in the root exudates of several Poaceae species and the Solanaceae species N. tabacum and tomato (Solanum lycopersicum). The structure of this didehydro-orobanchol isomer has not yet been clarified. These strigol-related germination stimulants are collectively called strigolactones (Fig. 1).

Figure 1. Chemical structures of natural strigolactones and the synthetic analog GR24.

Among the strigolactones, 5-deoxystrigol was originally isolated as a branching factor of arbuscular mycorrhizal (AM) fungi from root exudates of Lotus japonicus (Akiyama et al., 2005). We also identified 5-deoxystrigol as one of major germination stimulants of S. bicolor, Z. mays, and pearl millet (Pennisetum typhoideum) (Awad et al., 2006).

The AM fungi, which are obligate symbionts, are incapable of completing their life cycle without residing in their host roots. After spore germination and hyphal growth, the hyphal branching of AM fungi occurs in the vicinity of host roots. Because the phenomenon does not occur in the vicinity of roots of nonhosts, including rapeseed (Brassica napus) and white lupin (Lupinus albus), hyphal branching is considered to be the host recognition process (Giovannetti et al., 1993). In addition, strigolactones were found to induce a rapid increase in mitochondrial density and changes in the shape and movement of the organelles in AM fungi (Besserer et al., 2006). Such activation of the mitochondria might lead to the oxidation of lipids, which are the main form of carbon storage in AM fungal spores. Therefore, strigolactones may be crucial components of root exudates that switch on lipid catabolism at the presymbiotic stage of the fungus (Besserer et al., 2006; Akiyama, 2007). Furthermore, strigolactones induce gene expression of Gigaspora margarita CuZn superoxide dismutase (GmarCuZnSOD) (Lanfranco et al., 2005) and chemotrophic growth of Glomus mosseae hyphae (Sbrana & Giovannetti, 2005). The AM association is by far the most widespread association between microorganisms and higher plants. Within the angiosperms, at least 80% of the species are able to form AM symbioses (Harrison, 2005). Therefore, if strigolactones are indispensable for host recognition of AM fungi, they may be widely distributed in the plant kingdom (Akiyama, 2007). However, characterizations of strigolactones have been conducted for only a few plant species as plants exude trace amounts of unstable strigolactones. Accordingly, it is necessary to clarify the distribution of strigolactones in the plant kingdom to understand the chemical communications in the rhizosphere between plants and AM symbionts, and plants and root parasites.

We have developed a specific and rapid analytical method for known strigolactones using high-performance liquid chromatography (HPLC) connected to tandem mass spectrometry (LC/MS/MS) (Sato et al., 2003). LC/MS/MS analyses revealed the presence of 5-deoxystrigol in the root exudates of S. bicolor, Z. mays, and P. typhoideum (Awad et al., 2006). Besserer et al. (2006) identified sorgolactone as a branching factor in S. bicolor root exudates using LC/MS/MS. Furthermore, using LC/MS/MS, we demonstrated that nutrient deficiencies affect strigolactone exudation: in T. pratense, phosphorus (P) deficiency significantly promoted orobanchol exudation (Yoneyama et al., 2007a), while in S. bicolor, nitrogen (N) deficiency as well as P deficiency enhanced 5-deoxystrigol exudation (Yoneyama et al., 2007b).

To date, we have examined a wide range of plant species including crops, weeds, and even trees for the production of strigolactones and have found that plants produce diverse mixtures of known and unknown strigolactones. Although some of these unknown strigolactones have been purified and subjected to structural elucidation (Yokota et al., 1998; Xie et al., 2007, 2008a,b), at least several novel strigolactones remain to be characterized (K. Yoneyama, unpublished). Therefore, to compare strigolactone production among different plant species, the major strigolactones in each plant species should first be identified. As in the case of the Poaceae species (Awad et al., 2006), the major strigolactones of plant species within the same family are expected to be similar. We thus focused on the Fabaceae for the comparison of strigolactone production within the family as we had already identified major strigolactones produced by T. pratense (Yokota et al., 1998).

In this paper we extended the characterization of strigolactones in root exudates of 12 Fabaceae plant species, including L. albus, a nonhost of AM fungi, by comparing retention times of germination stimulants on reverse-phase (RP)-HPLC with those of standards and by using LC/MS/MS. In addition, the effects of N deficiency and P deficiency on strigolactone exudation by L. albus were examined to clarify whether the regulation of strigolactone exudation is related to the nutrient acquisition strategy of plants.

Materials and Methods

Chemicals

(+)-Orobanchol was a generous gift of Emeritus Prof. Kenji Mori (The University of Tokyo, Tokyo, Japan). GR24 was kindly provided by Prof. Binne Zwanenburg (Radboud University, Nijmegen, the Netherlands). (+)-Orobanchyl acetate was purified from root exudates of T. pratense (Xie et al., 2008a). (+)-Solanacol and the didehydro-orobanchol isomer were purified from root exudates of N. tabacum (Xie et al., 2007). (±)-5-Deoxystrigol was prepared as reported previously (Akiyama et al., 2005). The other chemicals of analytical grade and HPLC solvents were obtained from Kanto Chemical Co. Ltd (Tokyo, Japan) and Wako Pure Chemical Industries Ltd (Osaka, Japan).

Plant material

Orobanche minor Sm. seeds were collected from mature plants that were parasites of Trifolium pratense L. grown in the Watarase basin of Tochigi Prefecture, Japan. Seeds of Glycine max (L.) Merrill, Phaseolus vulgaris L., Vicia faba L., Arachis hypogaea L., Astragalus sinicus L., Medicago sativa L., Pisum sativum L., Trifolium incarnatum L., Vigna angularis (Willd.) Ohwi & Ohashi, and Psophocarpus tetragonolobus L. were obtained from a local supplier. Seeds of Cicer arietinum L., and Lupinus albus L. were generously supplied by Dr Yaakov Goldwasser (Hebrew University of Jerusalem, Jerusalem, Israel) and Dr Jun Wasaki (Hiroshima University, Hiroshima, Japan), respectively.

Growth conditions and collection of root exudates

Plant seeds were surface-sterilized in 70% ethanol for 2 min and then 1% NaClO for 2 min. After thoroughly rinsing with sterile distilled water, the seeds were germinated on moistened filter paper in a container at 23°C (L. albus and C. arietinum) or 25°C (the remainder) in the dark for 3 d. Seedlings (n = 20–100) were transferred to a strainer (28 × 23 × 9 cm, width × length × height (W × L × H)) lined with a sheet of gauze moistened by placing it in a slightly larger container (28.5 × 23.5 × 11 cm, W × L × H) containing 1 l of tap water as the culture medium in a growth chamber with a 14 : 10 h photoperiod at 120 µmol photons m⁻² s⁻¹ at 23 : 20°C (L. albus and C. arietinum) or 25 : 23°C (the remainder). The plants were grown in tap water for 7 d and then transferred to 1/2 Tadano and Tanaka medium (Tadano & Tanaka, 1980) without P and grown for another 3 d. Two strainers were used for each plant species. Then, the two strainers were transferred to a larger container (53.5 × 33.5 × 14 cm, W × L × H) containing 10 l of tap water and 1 mM CaCl₂. Root exudates released into culture medium were adsorbed by activated charcoal with using circulation pumps (Akiyama et al., 2005). The plants were grown for 8 d, during which the culture medium and activated charcoal were exchanged every other day.

Extraction of root exudates

Root exudates adsorbed on the charcoal were eluted with acetone. After the acetone was evaporated in vacuo, the residue was dissolved in 50 ml of water and extracted 3 times with 50 ml of ethyl acetate. The ethyl acetate extracts were combined, washed with 0.2 M K₂HPO₄ (pH 8.3), dried over anhydrous MgSO₄, and concentrated in vacuo. These crude extracts were stored in sealed glass vials at 4°C until use.

Characterization of strigolactones

Characterization of strigolactones in the root exudates was conducted by comparing retention times of germination stimulants on RP-HPLC with those of synthetic (or natural) standards and by using LC/MS/MS (Sato et al., 2003, 2005; Awad et al., 2006).

HPLC separation

HPLC separation was conducted with a U980 HPLC instrument (Jasco, Tokyo, Japan) fitted with an ODS (C₁₈) column (Mightysil RP-18, 2 × 250 nm, 5 µm; Kanto Chemicals Co., Ltd., Tokyo, Japan). The crude extracts were dissolved in 60% methanol and filtered through spin columns (Ultra-Free MC, 0.45 µm pore size; Milipore, Tokyo, Japan), and 10 µl was injected. The mobile phase was 60% methanol in water and was changed to 100% methanol 30 min after injection. The column was then washed with 100% methanol for 20 min. The flow rate was 0.2 ml min⁻¹ and the column temperature was set to 40°C.

Mass spectrometry

Mass spectrometry was performed with a Quattro LC mass spectrometer (Micromass, Manchester, UK) equipped with an electrospray source. The drying and nebulizing gas was nitrogen generated from pressurized air in an N2G nitrogen generator (Parker-Hanifin Japan, Tokyo, Japan). The nebulizer gas flow was set to approx. 100 l h⁻¹, and the desolvation gas flow to 500 l h⁻¹. The interface temperature was set to 400°C, and the source temperature to 150°C. The capillary and cone voltages were adjusted to orobanchol and to the positive ionization mode. MS/MS experiments were conducted using argon as the collision gas and the collision energy was set to 16 eV. The collision gas pressure was 0.15 Pa. For the detection of known strigolactones, we used six-channel multiple reaction monitoring (MRM). The six transitions of m/z 339 > 242, 353 > 256, 365 > 268, 367 > 270, 369 > 272, and 411 > 254 were monitored for sorgolactone, 5-deoxystrigol, solanacol, didehydro-orobanchol (strigol), strigol (orobanchol and sorgomol), and orobanchyl acetate (strigyl acetate), respectively. Data acquisition and analysis were performed with the MassLynx software (ver. 4.1). Quantification of strigolactones was carried out using synthetic or natural standards in a manner similar to that for orobanchol, strigol, and 5-deoxystrigol (Sato et al., 2003, 2005; Yoneyama et al., 2007a,b).

Germination assays

A portion of the ethyl acetate extracts dissolved in 60% methanol was fractionated by RP-HPLC operated under the same conditions as for LC/MS/MS analyses and the fractions collected every minute were examined for O. minor seed germination stimulation (Goldwasser et al., 2008). For this, preliminary germination assays were conducted to determine an optimal dilution of each crude extract, because, when an excessive amount of crude extract was loaded for HPLC, germination stimulation activity was distributed broadly in many fractions. The amounts of crude extracts used for germination assays were 1/1000 that used for LC/MS/MS analyses for P. tetragonolobus, 1/500 for A. hypogaea and V. faba, and 1/100 for the other species.

Germination assays on O. minor seeds were conducted as reported previously (Yoneyama et al., 2007a). The surface-sterilized O. minor seeds, c. 20 each, were placed on 6-mm glass fiber discs (Whatman GF/A) and c. 90 discs were incubated in a 9-cm sterile Petri dish lined with a sheet of filter paper (No. 2, Advantec, Tokyo, Japan) and wetted with 6 ml of sterile Milli-Q water in the dark at 23°C for 7 d as a ‘conditioning period’ during which the seeds become responsive to germination stimulants. Then, four discs carrying the conditioned seeds were transferred to a 5-cm sterile Petri dish prepared as follows. An aliquot of root exudate samples in 60% methanol or methanol solutions of authentic strigolactones was added to a 5-cm Petri dish lined with a filter paper. The solvent was allowed to evaporate before the discs carrying the conditioned seeds were placed on the filter paper and treated with sterile Milli-Q water (650 µl). The Petri dishes were sealed, enclosed in polyethylene bags, and placed in the dark at 23°C for 4–5 d. Seeds treated with or without GR24 (10⁻⁶ M) were always included as positive and negative controls. Seeds were considered germinated when the radicle protruded through the seed coat.

Hyphal branching assay

Hyphal branching assays for Gigaspora margarita Becker & Hall were conducted as reported previously (Akiyama et al., 2005). Spores of G. margarita (CGC1411; Central Glass Co., Tokyo, Japan), surface-sterilized with 0.2% NaClO and 0.05% Triton X-100, were inserted into a 0.2% Phytagel gel (Sigma-Aldrich, Tokyo, Japan) containing 3 mM MgSO₄ in 60-mm plastic Petri dishes. The dishes were incubated vertically for 5–7 d in a 2% CO₂ incubator at 32°C. Secondary hyphae emerging from a primary hypha, which grew upward in a negative geotropic manner in the gel, were used for assay. Test samples were first dissolved in acetone and then diluted with 70% ethanol in water. Paper discs (6 mm) loaded with 15 µl of test sample solution were placed in front of the tips of the secondary hyphae. The control was 70% ethanol in water. The hyphae branch patterns were observed 24 h after treatment. The sample was scored as positive for hyphal branching if new hyphal branches formed from the treated secondary hyphae or primary hyphae located proximal to the paper discs. The assay was repeated at least twice, using between three and five dishes for each concentration.

The clusters of hyphal branches consisting of higher orders of hyphae (third, fourth, and fifth hyphae) are induced from the secondary hyphae by treatment with positive samples such as root exudates and pure strigolactones. In the control treatment (70% ethanol in water), no hyphae or an occasional single branching hypha is induced from the treated secondary hyphae located proximal to the paper discs. This distinct difference in hyphal morphology enables us to distinguish positive samples from negative controls qualitatively.

Effects of nutrient deficiencies on strigolactone exudation by L. albus

After germination, 20 seedlings of L. albus were grown in tap water for 10 d and then subjected to each nutrient condition. Half-strength Tadano and Tanaka media (Tadano & Tanaka, 1980) was used as the control medium containing 2.43 mM N and 160 µM P. Low-N and low-P nutrient media had 120 µM N and 8 µM P, respectively. After 10 d of acclimatization in the test media, the growth media containing root exudates (plus washings) collected daily (1.5 l) were extracted with ethyl acetate as reported previously (Yoneyama et al., 2007a,b).

Results

Characterization of strigolactones in root exudates of Fabaceae plants

For the characterization of strigolactones produced by the Fabaceae species, we collected relatively large amounts of root exudates as we did not know how much strigolactones they would produce. The plants were grown in tap water and root exudates were collected using activated charcoal, because the production of strigolactones seemed to be enhanced under these conditions (Yoneyama et al., 2001; Akiyama et al., 2005). We first tried to examine qualitative but not quantitative differences in strigolactone production among plant species. In the germination assays, we used O. minor because this parasite could be handled without any strict quarantine restrictions and in addition O. minor has been reported to have a relatively broad host range (Parker, 1986; Parker & Riches, 1993) and thus its seeds are expected to be sensitive to various strigolactones. Not all of the 12 Fabaceae species examined in this study have been reported as hosts for Orobanche spp. Some of them are hosts of Orobanche crenata, Phelipanche ramosa and Phelipanche aegyptiaca (Parker, 1986; Parker & Riches, 1993).

In the preliminary germination and hyphal branching assays with crude extracts, all of the 12 Fabaceae species were found to exude germination stimulants of O. minor and branching factors of G. margarita. Then, each crude extract was examined for the presence of known strigolactones by comparing retention times of germination stimulants on RP-HPLC with those of synthetic or natural standards and by using LC/MS/MS. The results clearly demonstrated that all of the plants examined, including L. albus, a nonhost of AM fungi, produced known strigolactones.

A four-channel MRM chromatogram and the distribution of germination stimulation activity on O. minor after RP-HPLC separation of L. albus root exudates are shown in Fig. 2. Among known strigolactones, orobanchol, orobanchyl acetate, and 5-deoxystrigol were identified in the transitions of m/z 369 > 272, 411 > 254, and 353 > 256 at retention times of 7.9, 16.4 and 27.3 min, respectively, by LC/MS/MS (Fig. 2a). These assignments were confirmed by co-chromatography with synthetic and natural standards (data not shown). Germination stimulation activities on O. minor were associated with the fractions whose retention times corresponded to those of orobanchol (fraction 8), orobanchyl acetate (fractions 15 and 16), and 5-deoxystrigol (fraction 28) (Fig. 2b). In addition to these known strigolactones, a distinct peak was detected in the MRM channel monitoring the transition of m/z 367 > 270 at a retention time of 6.6 min, and the activity was observed in fraction 6 corresponding to this peak in the MRM chromatogram. This was identified as a didehydro-orobanchol isomer which had been detected in the root exudates of the Solanaceae plants N. tabacum (Xie et al., 2007) and S. lycopersicum (Sato et al., 2003). This strigolactone was the third, minor stimulant produced by T. pratense (Yokota et al., 1998) but its structure has not yet been elucidated. Furthermore, there was a distinct germination stimulation activity eluted in fraction 11 (11–12 min), whereas there were no peaks at the corresponding retention time in the six-channel MRM chromatogram, indicating the presence of an unknown germination stimulant. Accordingly, L. albus was found to exude the didehydro-orobanchol isomer, orobanchol, orobanchyl acetate, 5-deoxystrigol and an unknown germination stimulant (Fig. 2a,b).

Figure 2. (a) Four-channel multiple reaction monitoring (MRM) chromatogram of the root exudates from Lupinus albus, a nonhost of arbuscular mycorrhizal (AM) fungi, where the transitions of m/z 367 > 270, 369 > 272, 411 > 254, and 353 > 256 were monitored for didehydro-orobanchol, orobanchol, orobanchyl acetate, and 5-deoxystrigol (and their isomers), respectively. (b) Distribution of germination stimulation activity on Orobanche minor after reverse-phase high-performance liquid chromatography (HPLC) separation of the root exudates. All the fractions were tested for germination stimulation activity but only the numbers of active fractions are presented.

Figure 3 shows a four-channel MRM chromatogram and the distribution of germination stimulation activity on O. minor after RP-HPLC separation of P. sativum root exudates. Orobanchol, orobanchyl acetate, 5-deoxystrigol, and the didehydro-orobanchol isomer were detected (Fig. 3a) by LC/MS/MS and their identities were confirmed by co-chromatography with strigolactone standards (data not shown). However, germination stimulation activity was not observed in the fractions corresponding to the retention times of the didehydro-orobanchol isomer, orobanchol, and 5-deoxystrigol (Fig. 3b). Fraction 16 (16–17 min), corresponding to the retention time of orobanchyl acetate, showed weak germination stimulation. Instead, potent germination stimulation activity was observed in fraction 12 (12–13 min), indicating that the major germination stimulant of P. sativum was an unknown compound.

Figure 3. (a) Four-channel multiple reaction monitoring (MRM) chromatogram of root exudates from Pisum sativum. The channels were set for the same transitions as in Fig. 2a. (b) Distribution of germination stimulation activity on Orobanche minor after reverse-phase high-performance liquid chromatography (HPLC) separation of the root exudates. All the fractions were tested for germination stimulation activity but only the numbers of active fractions are presented.

Distinct germination stimulant activity in fraction 12 after RP-HPLC separation was detected not only in the root exudates of P. sativum but also in those of A. sinicus, A. hypogaea, C. arietinum, M. sativa, T. incarnatum, and V. faba. Therefore, this novel stimulant seems to be distributed widely in the Fabaceae. However, it remains unclear whether these plants exude the same compound.

Table 1 summarizes the distribution of strigolactones among the Fabaceae plants examined. In this table, strigolactones identified by LC/MS/MS are designated as ‘MS’ and ‘G’ means that germination stimulation activity was detected in the fraction corresponding to the retention time of the strigolactone standard. It should be noted that germination assays were conducted at a particular dilution of the root exudates at which major stimulants were clearly separated by RP-HPLC. Therefore, the lack of G in Table 1 means that the contribution of the corresponding strigolactone to the overall germination stimulation activity of the root exudates was rather small.

Table 1.  Distribution of strigolactones in Fabaceae plants

Scientific name	Solanacol	Didehydro-orobanchol	Orobanchol	Orobanchyl acetate	5-Deoxystrigol
Strigolactones were characterized by using tandem mass spectrometry (LC/MS/MS) and by comparing retention times of germination stimulation activity on reverse-phase high-performance liquid chromatography (HPLC) with those of natural and synthetic standards. MS indicates that the strigolactone was detected by LC/MS/MS. G indicates that germination stimulation activity on Orobanche minor seeds was observed at the retention time corresponding to that of the strigolactone. It should be noted that only major germination stimulants could be detected in the germination assays.
Arachis hypogaea	–	–	MS/G	MS/G	MS/G
Astragalus sinicus	–	MS/G	MS/G	MS/G	MS/G
Cicer arietinum	–	MS/G	–	MS/G	MS/G
Glycine max	–	–	MS/G	MS/G	MS
Lupinus albus	–	MS/G	MS/G	MS/G	MS/G
Medicago sativa	–	–	MS/G	MS/G	MS/G
Phaseolus vulgaris	–	–	MS/G	MS/G	MS
Pisum sativum	–	MS	MS	MS/G	MS
Psophocarpus tetragonolobus	–	–	MS/G	MS/G	MS/G
Trifolium incarnatum	MS/G	–	MS/G	MS/G	–
Vicia faba	–	–	MS/G	MS/G	MS
Vigna angularis	–	–	MS/G	MS/G	MS/G

Orobanchyl acetate was detected in the root exudates of all the Fabaceae plants examined. Orobanchol and 5-deoxystrigol were present in the root exudates of 11 of the 12 Fabaceae plants. Therefore, these three strigolactones, orobanchol, orobanchyl acetate, and 5-deoxystrigol, were found to be major strigolactones in the Fabaceae plants. In addition to these major strigolactones, the didehydro-orobanchol isomer was detected by LC/MS/MS and its activity was confirmed by germination assays in the root exudates of A. sinicus, C. arietinum, and L. albus. Solanacol was identified only in the root exudates of T. incarnatum. Although 5-deoxystrigol was detected by LC/MS/MS in the root exudates of G. max, P. vulgaris, P. sativum, and V. faba, the RP-HPLC fractions corresponding to the retention time of 5-deoxystrigol were inactive in the germination assays (Table 1), indicating that 5-deoxystrigol was a minor germination stimulant in these species. None of these Fabaceae plants produced detectable levels of strigol, sorgolactone, or sorgomol.

Effects of N and P deficiencies on strigolactone exudation by L. albus

The nonmycotrophic L. albus species was found to exude the didehydro-orobanchol isomer, orobanchol, orobanchyl acetate, 5-deoxystrigol, and an unknown stimulant (Fig. 2a,b). Among the strigolactones present in the root exudates of L. albus, orobanchol, orobanchyl acetate and 5-deoxystrigol could be quantified by LC/MS/MS. Therefore, the effects of P and N deficiencies on the exudation of orobanchol, orobanchyl acetate, and 5-deoxystrigol by L. albus were examined.

The P contents of shoots and roots of L. albus plants grown under low-P conditions were 2.6 and 3.2 mg (g DW)⁻¹, which were 40% and 20% those of the controls, respectively, and thus the plants were in a state of P deficiency. Similarly, the N contents of the plant material subjected to N deficiency were < 50% those of the controls.

In the case of the mycotrophic Fabaceae plant T. pratense, a reduced supply of P but not of other minerals significantly increased the exudation of orobanchol by the roots (Yoneyama et al., 2007a). In contrast, deficiency of neither P nor N promoted the exudation of orobanchol, orobanchyl acetate, or 5-deoxystrigol by L. albus roots (Fig. 4a). In fact, both N and P deficiencies tended to reduce exudation of these strigolactones. The nonmycotrophic L. albus grown hydroponically under P deficiency released 1.5 ± 0.5 pg g⁻¹ root FW of orobanchol and 37.5 ± 7.5 pg g⁻¹ root FW of orobanchyl acetate over 3 d (mean ± SE, n = 3) as shown in Fig. 4a. By contrast, under the same growth conditions, the mycotrophic T. pratense exuded 36 000 ± 4200 pg g⁻¹ root FW of orobanchol and 153 000 ± 19 700 pg g⁻¹ root FW of orobanchyl acetate over 3 d (mean ± SE, n = 3) (Yoneyama et al., 2007a).

Figure 4. (a) Exudation of orobanchol (white bars), orobanchyl acetate (gray bars), and 5-deoxystrigol (dark gray bars) by Lupinus albus roots and (b) germination stimulation activity of L. albus root exudates from plants grown under nutrient deficiencies in hydroponic culture medium over 3 d. The control medium (white circles) contained 160 µM phosphorus (P) and 2.43 mM nitrogen (N) which were reduced to 1/20 in the low-P (8 µM P; triangles) and low-N (120 µM N; black circles) media. (a) Root exudate samples collected daily were analyzed separately by tandem mass spectrometry (LC/MS/MS). The data are sums of exudations over 3 d. The experiments were repeated three times. Values represent the means ± SE. (b) Root exudate samples collected daily for 3 d were combined and used in germination assays with Orobanche minor seeds. Aliquots of methanol solutions of the root exudate samples corresponding to 4.5, 45, and 450 ml of medium were transferred to 5-cm Petri dishes lined with a filter paper, and the solvent was allowed to evaporate before the discs carrying the conditioned O. minor seeds were placed on the filter paper and treated with distilled water (650 µl). The experiments were repeated three times. Values represent the means ± SE.

Lupinus albus exudes the didehydro-orobanchol and an unknown germination stimulant in addition to these strigolactones, as shown in Fig. 2(a,b), and thus deficiency of N and/or P may affect exudation of those stimulants which cannot be quantified as their pure standards are not yet available. The amounts of the pure didehydro-orobanchol isomer isolated from N. tabacum root exudates were too small to weigh accurately. Therefore, crude extracts of root exudates from the plants grown under P or N deficiency were examined for their germination stimulation on O. minor seeds (Fig. 4b).

A 4.5-ml equivalent of the control culture medium induced c. 50% germination, while those of N and P deficiencies at the same dosage elicited < 10% germination (Fig. 4b). At 45- and 450-ml equivalents of culture medium, there were no significant differences in the germination stimulation activities among the root exudates from the plants grown under the different nutrient conditions. The reduction in germination rate at the highest dosage of root exudate from the control was attributable to germination inhibitors produced simultaneously. The germination stimulation activities in the culture media therefore seemed to be nearly proportional to the amounts of orobanchol, orobanchyl acetate, and 5-deoxystrigol, suggesting that the exudation of the didehydro-orobanchol isomer and the unknown germination stimulant appeared to be affected similarly to the other strigolactones.

Discussion

In the present study, strigolactones exuded from 12 Fabaceae plants were characterized by an RP-HPLC separation–germination assay and by LC/MS/MS. Orobanchyl acetate (alectrol) was detected in root exudates from all of these plants, and most of them exuded orobanchol and 5-deoxystrigol. Therefore, orobanchol, orobanchyl acetate, and 5-deoxystrigol appear to be major germination stimulants in the Fabaceae. In addition to these strigolactones, a didehydro-orobanchol isomer and solanacol were identified in the root exudates from four and one species, respectively. Furthermore, two novel stimulants, one from P. sativum and the other from L. albus root exudates, were detected. The stimulant found in the root exudate of P. sativum seems to be present in the root exudates of A. sinicus, A. hypogaea, C. arietinum, M. sativa, and V. faba. After several steps of purification, this stimulant appeared to be a novel strigolactone with a molecular weight of 404 Da, while its structure has not yet been fully elucidated as the amounts purified from P. sativum root exudates were not sufficient for comprehensive spectroscopic analyses.

In the case of P. sativum, although distinct peaks of the didehydro-orobanchol, orobanchol, orobanchyl acetate, and 5-deoxystrigol were observed in the MRM chromatogram (Fig. 3a), germination stimulation activities were not detected in the RP-HPLC fractions corresponding to the retention times of orobanchol and 5-deoxystrigol. Only a weak germination stimulation activity was associated with fraction 16, corresponding to the retention time of orobanchyl acetate (Fig. 3b). This implies that these strigolactones contributed to the overall germination stimulation activity in the root exudate only to a small extent as compared with the novel strigolactone eluted in fractions 12 and 13 and orobanchyl acetate in fraction 16.

In the LC/MS/MS analyses, each strigolactone was monitored with the MRM channel specific to it at an independent sensitivity. For example, in Fig. 3a, the ion intensity in the channel for orobanchyl acetate (1.95 × 10⁷) was 100-fold larger than those for orobanchol (1.15 × 10⁵) and 5-deoxystrigol (7.14 × 10⁴). Given the fact that, in electrospray ionization (ESI) mass spectrometry, ion intensities are not always proportional to the concentrations of analytes, and the ionization efficiency of orobanchol is c. 10-fold lower than those of orobanchyl acetate and 5-deoxystrigol, the amounts of orobanchyl acetate would be at least 10-fold larger than those of orobanchol and 5-deoxystrigol. We could not increase amounts of root exudate samples for RP-HPLC separation, because the activity was distributed broadly in many fractions when an excessive amount of sample was loaded for HPLC.

Similar results were obtained with the root exudates of G. max, P. vulgaris, and V. faba, where 5-deoxystrigol was detected by LC/MS/MS but germination stimulation activity was not observed in the fraction(s) corresponding to the retention time of 5-deoxystrigol. In addition to the differences in the contents of strigolactones as discussed above, the differences in sensitivity of root parasite seeds to various strigolactones and the presence of inhibitors also affected the results of germination assays. In fact, O. minor used in the germination assays was c. 100-fold more sensitive to orobanchol than to 5-deoxystrigol; 0.1 nM orobanchol elicited c. 60% germination and a similar level of germination was observed at 10 nM 5-deoxystrigol.

Although many authors have suggested that strigolactones are distributed widely in the plant kingdom, the isolation and identification of strigolactones have been hampered by the extremely low concentrations produced and exuded by host roots as well as their relative instability. The recent development of an analytical method using LC/MS/MS enables identification and quantification of known strigolactones (Sato et al., 2003, 2005; Awad et al., 2006; Besserer et al., 2006; Yoneyama et al., 2007a,b; Xie et al., 2007) and, in addition, the search for novel strigolactones (Awad et al., 2006; Xie et al., 2007). Here we demonstrated that the 12 Fabaceae plants, including the nonmycotrophic L. albus, exude at least three strigolactones, and that, among the known strigolactones, orobanchol, orobanchyl acetate, and 5-deoxystrigol are the major ones produced by these plants grown hydroponically.

It is intriguing that all the plants examined so far exude mixtures of at least two strigolactones. Although the question of why plants produce and exude mixtures of strigolactones has not yet been answered, quantitative and/or qualitative differences in the strigolactone compositions may be one of the key factors determining the host specificity of AM fungi and root parasites. Therefore, the effects of various combinations of strigolactones on both hyphal branching of AM fungi and seed germination of root parasites need to be examined.

Orobanchol and orobanchyl acetate (alectrol) were first isolated from root exudates of T. pratense and V. unguiculata as germination stimulants, respectively. Orobanchol was also identified in root exudates from Solanaceae plants including N. tabacum (Xie et al., 2007) and S. lycopersicum (Sato et al., 2003), and from the Compositae marigold (Targetes patula) (K. Yoneyama, unpublished). Orobanchyl acetate was detected in several Compositae plants (K. Yoneyama, unpublished). However, these strigolactones have not been found in the root exudates from any members of the Poaceae examined to date (Awad et al., 2006; K. Yoneyama, unpublished). These results suggest that orobanchol and orobanchyl acetate seem to be distributed in dicotyledons but not in monocotyledons.

5-Deoxystrigol, originally isolated as a hyphal branching inducer for AM fungi from the root exudate of Lotus japonicus (Akiyama et al., 2005), has been shown to be one of major strigolactones in monocotyledonous plants including S. bicolor, Z. mays and P. typhoideum (Awad et al., 2006). As all the Fabaceae plants examined in this study, with the exception of T. incarnatum, were found to exude 5-deoxystrigol, 5-deoxystrigol appears to be widely distributed in both monocotyledons and dicotyledons. This is in good agreement with the proposed biosynthetic pathway for strigolactones, originating from carotenoids, where 5-deoxystrigol serves as the precursor of all the other known strigolactones (Matusova et al., 2005; Bouwmeester et al., 2007). The oxidation of 5-deoxystrigol produces mono-hydroxy-strigolactones such as orobanchol, strigol, and the recently identified sorgomol (Xie et al., 2008b). It is likely that sorgomol is then converted to sorgolactone by subsequent oxidation and decarboxylation (Xie et al., 2008b; Rani et al., in press). Among the plants examined to date, T. incarnatum, N. tabacum (Xie et al., 2007) and S. lycopersicum (unpublished data) did not produce detectable amounts of 5-deoxystrigol. It is interesting that these three plant species exuded solanacol, a strigolactone containing a benzene ring.

AM fungi supply mineral nutrients, especially P, to host plants by extending beyond the depletion zone for P around the root, the external mycelia improving P absorption. However, AM fungi are absent under all environmental conditions in the Brassicaceae and Chenopodiaceae, and are also quite rare or absent in many members of the Proteaceae and other typical root cluster-forming plant species including L. albus (Marschner, 1993). The proximity of roots of nonhosts including L. albus did not elicit hyphal branching of AM fungi (Giovannetti et al., 1993). An early hypothesis that nonhost plants do not induce the morphogenetical event suggested that roots of nonhosts secrete compounds into the rhizosphere that inhibit AM colonization (Fontenla et al., 1999). Schreiner & Koide (1993) showed that members of Brassicaceae have the potential to produce significant quantities of antifungal compounds in roots, probably isothiocyanates. An alternative explanation for the lack of AM colonization of nonhosts is that nonhosts fail to produce branching factors required by AM fungi for host recognition. In this study, it was demonstrated that the nonmycotrophic L. albus grown hydroponically also produces and exudes strigolactones, and therefore the hypothesis that nonhosts of AM fungi fail to exude strigolactones may be excluded.

The amounts of strigolactones released from L. albus, however, were quite low compared with mycotrophic plants. Indeed, the amounts of orobanchol and orobanchyl acetate, per unit root fresh weight, released from L. albus grown under P starvation were only 1/24 000 and 1/4080 that from T. pratense, respectively. Although sizes and growth stages of plants and the composition of strigolactones produced were different and thus direct comparisons were not possible, the total amount (or activity) of all strigolactones exuded by L. albus would be c. 1/1000 that exuded by T. pratense. For example, in the case of T. pratense, approx. 80% germination of O. minor was induced at 500-µl equivalent of the control culture (1/2 Tadano & Tanaka) medium (Yoneyama et al., 2007a) and a similar level of germination was achieved at 450-ml equivalent of the control medium of L. albus (Fig. 4b).

Another clear difference between mycotrophic and nonmycotrophic plants was observed in the response of strigolactone exudation to nutrient deficiency. In the cases of T. pratense and S. bicolor, which are host plants of AM fungi, P deficiency (and also N deficiency in S. bicolor) significantly promoted strigolactone exudation (Yoneyama et al., 2007a,b). In the case of L. albus, however, P and N deficiency slightly reduced strigolactone exudation (Fig. 4). Such a decrease in strigolactone production may be attributable to a reduction in metabolic functions in L. albus under N and P deficiencies. In another nonmycotrophic plant, Spinacia oleracia, neither P nor N deficiency enhanced strigolactone production (K. Yoneyama, unpublished).

Recently, the strigolactone orobanchol was identified from the root exudates of Arabidopsis thaliana, a host of Orobanche spp. but not of AM fungi (Goldwasser et al., 2008). These results suggest that strigolactones have other unknown functions indispensable for the normal growth and development of plants themselves.

Legumes are thought to be very versatile in their symbioses (Sprent & James, 2007). Therefore, further work on the characterization of strigolactones from nodulating and nonnodulating legumes and legumes that produce ectomycorrhiza, and the effects of mineral nutrients on their strigolactone production, can be expected to unveil the functions of strigolactones in the rhizosphere community and in plants.

Acknowledgements

Part of this study was supported by a Sasakawa Scientific Research Grant from The Japan Science Society and a Grant-in-Aid for Scientific Research (1820810) from the Japan Society for the Promotion of Science (JSPS).