ARTICLES. Inhibition of shoot branching by new terpenoid plant hormones

Transcription

1 Vol 455 Septemer 28 doi:.8/nature7272 ARTICLES Inhiition of shoot ranching y new terpenoid plant hormones Mikihisa Umehara, Atsushi Hanada, Satoko Yoshida, Kohki Akiyama 2, Tomotsugu Arite, Noriko Takeda-Kamiya, Hiroshi Magome, Yuji Kamiya, Ken Shirasu, Koichi Yoneyama 4, Junko Kyozuka & Shinjiro Yamaguchi Shoot ranching is a major determinant of plant architecture and is highly regulated y endogenous and environmental cues. Two classes of hormones, auxin and cytokinin, have long een known to have an important involvement in controlling shoot ranching. Previous studies using a series of mutants with enhanced shoot ranching suggested the existence of a third class of hormone(s) that is derived from carotenoids, ut its chemical identity has een unknown. Here we show that levels of strigolactones, a group of terpenoid lactones, are significantly reduced in some of the ranching mutants. Furthermore, application of strigolactones inhiits shoot ranching in these mutants. Strigolactones were previously found in root exudates acting as communication chemicals with parasitic weeds and symiotic aruscular mycorrhizal fungi. Thus, we propose that strigolactones act as a new hormone class or their iosynthetic precursors in regulating aove-ground plant architecture, and also have a function in underground communication with other neighouring organisms. Shoot ranching involves the formation of axillary uds in the axil of leaves and susequent outgrowth of the uds. Previous studies have suggested the involvement of a novel, as yet unidentified, hormone in inhiiting the outgrowth of axillary uds, using a series of recessive mutants that show enhanced shoot ranching. These mutants include ramosus (rms) of pea (Pisum sativum) 4, more axillary growth (max) of Araidopsis 5 9, decreased apical dominance (dad) of petunia (Petunia hyrida), and dwarf (d)orhigh-tillering dwarf (htd) of rice (ryza sativa) 2 4. Reciprocal grafting experiments, doule mutant analysis and cloning of these genetic loci suggested that the novel hormone is iosynthesized from carotenoids and moves acropetally to inhiit axillary ud outgrowth 5. In the proposed iosynthesis pathway, MAX, RMS5 and HTD/D7 encode CARTENID CLEAVAGE DIXYGENASE 7 (CCD7) 4,7,, whereas MAX4, RMS, D and DAD encode another suclass of CCDs designated as CCD8 (refs,, 4) (Fig. a). CCD7 and CCD8 might catalyse sequential carotenoid cleavage reactions, although their endogenous sustrates and exact enzymatic function in plants have not een conclusive 7,,7. MAX is a cytochrome P45 monooxygenase presumaly involved in a later iosynthetic step 8 (Fig. a). Unlike the iosynthetic mutants, the ranching phenotype of the max2, rms4 and dad2 mutants is not rescued y grafting onto a wild-type rootstock, suggesting that they are insensitive to the ranch-inhiiting hormone 2,8,. MAX2, RMS4 and D are orthologous memers of the F-ox leucine-rich repeat (LRR) protein family 4,5,2 (Fig. a), which proaly act as the sustrate recognition suunit of SCF uiquitin E ligase for proteasomemediated proteolysis 8. The predicted iochemical function of MAX2, RMS4 and D is consistent with their role in signal transduction of the novel hormone. Strigolactones are a group of terpenoid lactones (Fig. ) which have een found in root exudates of diverse plant species and were initially characterized as seed germination stimulants of root parasitic plants such as Striga and roanche species 9 2. More recently, strigolactones were shown to act as root-derived signals for symiotic interaction with aruscular mycorrhizal fungi 22, which facilitate the uptake of soil nutrients y plants. This symiosis is oserved in more than 8% of terrestrial plants, coinciding with the wide distriution of this class of terpenes. Strigolactones may have additional unidentified function(s) in plants, ecause they induce seed germination of non-parasitic plants as well 2,24 and are also produced y non-hosts of aruscular mycorrhizal fungi, including Araidopsis 25,2. Little is known aout the iosynthesis of strigolactones. Recent work has indicated that the ABC part (Fig. ) is derived from carotenoids, presumaly y means of the formation of oxidatively cleaved product(s) 2,27,28. Taken together, current lines of evidence suggest that strigolactone iosynthesis involves a (epoxy)carotenoid cleavage enzyme conserved across diverse plant species. Although CCD7 and CCD8, encoded y the MAX/RMS/DAD/D loci, fulfil these criteria 29, their role in strigolactone iosynthesis had not een examined. Therefore, we set out to examine whether the carotenoidderived ranching inhiitor shares its iosynthetic pathway with strigolactones using rice d mutants. Strigolactone levels in rice d mutants To explore the potential role of D (CCD8) and D7 (CCD7) in strigolactone iosynthesis in rice, we analysed strigolactones in root exudates of wild-type and d mutants (Supplementary Fig. a) y liquid chromatography-quadruple/time-of-flight tandem mass spectrometry (LC/MS MS). Because our survey of known strigolactones in hydroponic culture media of rice seedlings (cv. Shiokari) identified 29-epi-5- deoxystrigol (epi-5ds), we synthesized deuterium-laelled epi-5ds (Supplementary Fig. 2) and used it as an internal standard for quantification on LC/MS MS. We selected [M H] (m/z 2 and for d -epi-5ds and cold epi-5ds, respectively) as parent ions on quadruple mass spectrometry and detected [M H 5] (m/z 27. and 2. for d -epi-5ds and cold epi-5ds, respectively) as fragment ions on time-of-flight mass spectrometry after collision-induced dissociation for quantification (Fig. 2a, ). Full-scan spectra of fragment ions RIKEN Plant Science Center, Tsurumi, Yokohama 2-45, Japan. 2 Graduate School of Life and Environmental Sciences, saka Prefecture University, - Gakuencho, Naka-ku, Sakai, saka , Japan. Graduate School of Agricultural and Life Sciences, The University of Tokyo, Yayoi, Bunkyo, Tokyo -852, Japan. 4 Weed Science Center, Utsunomiya University, Utsunomiya 2-855, Japan. 28 Macmillan Pulishers Limited. All rights reserved 95

2 ARTICLES NATURE Vol 455 Septemer 28 a H Carotenoid MAX, RMS5, HTD/D7 (CCD7) MAX4, RMS, D, DAD (CCD8) (+)-5-Deoxystrigol Carotenoid cleavage product MAX (P45) MAX2, RMS4, D (F-ox) C A B a D 4 (+)-Strigol Novel hormone Branchinginhiition 7 (+)-roanchol GR24 (Synthetic analogue) Figure The novel ranching inhiitor pathway (a) and chemical structures of representative strigolactones (). confirmed the identity of these compounds (Fig. 2c). As oserved for strigolactones in other species 28,,, the levels of epi-5ds in root exudates of wild-type seedlings were elevated when inorganic phosphate was depleted in the media (Fig. 2d). However, epi-5ds was nearly undetectale in exudates of d- and d7- mutants, regardless of the nutrient conditions (Fig. 2d). Reduced levels of another strigolactone species (29-epi-oroanchol or its isomer) in root exudates were also evident for the d-2 allele on the Nipponare ackground (Supplementary Fig. ). To determine whether the production of epi- 5DS was decreased or only the secretion from roots was defective in these mutants, we quantified endogenous epi-5ds in roots. We found that endogenous levels of epi-5ds were also decreased in d- and d7- seedlings relative to the wild-type control (Fig. 2e). These results demonstrate that oth D (CCD8) and D7 (CCD7) are required for the production of normal levels of strigolactones in rice seedlings. In contrast to the d- and d7- mutants, d- seedlings accumulated higher levels of epi-5ds oth in culture media and in roots than did wild-type plants under inorganic phosphate deficiency (Fig. 2d, e). These results are correlated with the upregulation of D (CCD8) transcript levels in d- and other tillering d mutants 4, and further support the idea that D (CCD8) participates in strigolactone iosynthesis. Similar transcriptional regulation of RMS (CCD8) was also found in the rms4 mutant of pea, proaly through a feedack inhiition mechanism in the ranching inhiitor pathway 4. The elevated strigolactone production in the d mutant suggests that the decreased strigolactone levels in the d and d7 mutants are attriutale to a direct lockage of the iosynthesis pathway, rather than a secondary consequence of the decreased ranching inhiitor activity, ecause in the latter case, strigolactone levels would e reduced also in the d mutant. Pre-conditioned seeds of the parasitic plant Striga hermonthica require germination stimulants, including strigolactones, released 9 H 28 Macmillan Pulishers Limited. All rights reserved from the host roots to complete germination. We used a highly sensitive germination assay using S. hermonthica seeds to estimate strigolactone concentrations in root exudates of d mutants 27,2. In agreement with the LC/MS MS data, the culture media of d- and d7- seedlings contained weaker germination-stimulating activity than did those of wild-type plants (Fig. 2f). By contrast, d- root exudates exhiited stronger germination-stimulating activity than the wild-type control. The reduced germination-stimulating activity in d- root exudates is not due to increased germination inhiitors, ut to decreased germination stimulants, ecause the addition of d- exudates did not inhiit germination induced y ()-strigol (Fig. 2f). These results indicate that overall strigolactone levels released from roots are decreased in the d- and d7- mutants. Strigolactones inhiit tillering in rice To investigate further the relationships etween the D/D7- derived ranching inhiitor and strigolactones, we examined the effect of strigolactone treatment on rice d mutants. We developed a hydroponic culture system using rice seedlings, where we oserved outgrowth of first and second tiller (axillary) uds in the d mutants, ut not in the wild type. An application of GR24 (a strigolactone analogue; Fig. ) to the media inhiited tiller ud outgrowth of 2-week-old d- and d7- seedlings in a dose-dependent manner (Fig. a, ). The inhiitory effect was detectale in response to as low as nm GR24, and tiller ud outgrowth was nearly fully inhiited at mm GR24. In contrast to d- and d7-, the d- mutant, defective in a proale signalling component (Fig. a), was insensitive to this chemical. No morphological anormalities were evident in wildtype seedlings after GR24 treatment. Similar effects were oserved when we used naturally occurring strigolactones, ()-strigol and ()-5DS, as well (Figs and c, d). The insensitivity of the d- mutant to strigolactones indicates that their inhiitory effects on tiller ud outgrowth are specific to the proposed ranching inhiitor pathway. These results illustrate that strigolactones or downstream metaolites act as the novel ranching inhiitor. The tillering dwarf phenotype of the d mutants is more drastic in appearance at a later stage 2,. We found that the ranching phenotype as well as the plant height of -week-old d- mutants were complemented y including 2 mm GR24 in the culture media, whereas no visile effect of this chemical was recognizale in d- mutant plants (Fig. e g). These results confirm the role of strigolactones in inhiiting tiller ud outgrowth in the ranching inhiitor pathway in rice. In many cases, hormonal responses are dose-dependent within a certain range and oth hormone-deficiency and hormone-excess phenotypes are oserved. We next examined the effect of a high dose of GR24 on tillering of wild-type seedlings. We found that tiller outgrowth was severely inhiited when mm GR24 was supplemented to the culture media, without affecting the growth of main leaves (Supplementary Fig. 4). These oservations further support the role of strigolactones in inhiiting axillary ud outgrowth and suggest the potential usefulness of strigolactones as plant growth regulators that specifically inhiit ranching. As mentioned aove, D transcript levels were previously shown to e elevated in the d- and d- mutants, suggesting a negative feedack control in the ranch inhiitor pathway 4. ur quantitative polymerase chain reaction with reverse transcription (qrt PCR) analysis revealed that GR24 treatment decreased D transcript levels in d- and wild-type seedlings, ut not in the d- mutant (Fig. h). These results, together with the elevated strigolactone production in the d- mutant (Fig. 2), indicate that endogenous strigolactone levels are under homeostatic control y means of the D-dependent signalling pathway and further support the idea that strigolactones (or downstream metaolites) act as the ranching inhiitors in rice. Effect of strigolactones in Araidopsis To determine whether strigolactones participate in the ranching inhiitor pathway in Araidopsis, we examined the effect of GR24

3 NATURE Vol 455 Septemer 28 ARTICLES a m/z 47. m/z 27. D c IS purified from root exudates (d -laelled) m/z 25. Relative intensity m/z 97. [M + H] + : m/z 2. m/z 2 m/z 27. m/z m/z 2. m/z 2 m/z 27. m/z IS purified from root exudates (d -laelled) Endogenous (from root exudates) Authentic IS (d -laelled) Authentic (cold) m/z Retention time (min) Relative intensity Endogenous (from root exudates) Authentic IS (d -laelled) Authentic (cold) m/z d epi-5ds (ng l ) Root exudates d- d- d7- d- d- d7- e epi-5ds (pg per g FW) Roots d- d- d7- DW S d- d- d7- d-+s d- d- +P i P i +P i P i +P i P i d- d- d7- f Striga germination (%) d7- d-+s Figure 2 Strigolactone analysis in rice seedlings. a, Predicted major fragmentation patterns of d -epi-5ds on LC/MS MS., Selected reaction monitoring for d -epi-5ds (internal standard, IS) or epi-5ds. c, Full-scan spectra of fragment ions. Asterisks indicate deuterium-laelled ions. d, e, LC/ MS MS analysis of epi-5ds levels in wild type () and d mutants in the presence (P i ) or asence (2P i ) of inorganic phosphate (mean s.d., n 5 ). FW, fresh weight. f, Estimation of germination stimulant levels in culture media using Striga seeds (means and s.d., n 5 ). DW, distilled water; S, ()-strigol ( 27 M); d-s, co-incuation with ()-strigol ( 27 M) and d- culture media. on the ranching phenotype of max mutants (Supplementary Fig. ). The MAX genes are required for selective repression of axillary shoots and max mutants exhiit ushier shoots than do wild-type plants 5,9. ur data show that the enhanced ranching phenotype of max and max4 mutants (defective in CCD7 and CCD8, respectively; Fig. a) is rescued y supplementing the hydroponic culture media with 5 mm GR24, whereas max2 mutants are insensitive to GR24 treatment (Fig. 4a, ). Next, we estimated the levels of strigolactones in root exudates of max mutants y determining germination-stimulating activity using S. hermonthica seeds. In root exudates from max and max4 seedlings, the levels of germination stimulants are significantly lower than those from wild-type seedlings. By contrast, the max2 mutant exuded germination stimulants at slightly higher levels than the wild type (Fig. 4c). Collectively, these results suggest that strigolactones are iosynthesized from carotenoid cleavage products y CCD7 and CCD8 and inhiit shoot ranching through the MAXdependent pathway in Araidopsis. d roots are infected y fewer Striga plants Here we have identified strigolactone-deficient and -insensitive mutants. To explore the impact of altered strigolactone levels on 28 Macmillan Pulishers Limited. All rights reserved the interaction with parasitic weeds, we used rice d mutants to oserve germination, infection and developmental processes of S. hermonthica plants. S. hermonthica is an oligate root parasite that infects cereals 4,5, including rice (Fig. 5a). In the vicinity of d- roots, fewer seeds germinated than did those co-incuated with wildtype or d- roots (Fig. 5), consistent with the finding that d- roots exude lower levels of strigolactones (Fig. 2d, f). As a consequence of the reduced germination frequency, fewer S. hermonthica plants estalished parasitism with d- mutants in 2 weeks than with wild type or d- (Fig. 5). When S. hermonthica seeds were co-incuated with d- seedlings after the induction of germination y ()-strigol, there was no significant difference in the frequency of successful parasitism among the three genotypes (Fig. 5c). Aleit at a very low frequency, some S. hermonthica seeds germinated in the vicinity of d- roots in the asence of ()-strigol and then achieved successful infection. Together, these results indicate that compared to wild-type roots, fewer S. hermonthica plants can infect d- roots principally due to lower levels of germination stimulants released from this host. ur results also suggest that, once the S. hermonthica seeds germinate, strigolactone deficiency does not significantly affect the susequent infection processes. We cannot rule out the possiility 97

4 ARTICLES NATURE Vol 455 Septemer 28 that a small amount of strigolactone owing to residual CCD8 activity might exist in the d- mutant and affect the germination and infection of S. hermonthica, ecause the d- mutation results in a single amino acid sustitution (Supplementary Fig. ) and may not e a null allele. Discussion utgrowth of axillary uds is, in part, regulated y the interaction of multiple hormonal signals 5 ; auxin is actively transported downwards in the shoots and inhiits ud outgrowth, whereas cytokinins move upwards in plants and activate ud outgrowth. In this study, we have shown that the d and max ranching mutants of rice and a Control e Control Araidopsis are deficient in or insensitive to strigolactones, and that exogenously applied strigolactones inhiit shoot ranching. Thus, we propose that strigolactones or downstream metaolites act as the long-sought-after hormones in the D/MAX pathway. However, it should e noted that the ioactive form(s) of this new hormone class has not een clarified in the current study. Extensive survey of natural strigolactones as seed germination stimulants of root parasites and hyphal ranching inducers of aruscular mycorrhizal fungi revealed highly diverse structures, attriutale to modifications on ring ABC and the C29 configuration 29, (Fig. ). Moreover, it is not known how these diverse strigolactones are further metaolized in plants. Elucidation of the ioactive form(s) of the ranch-inhiiting hormones is a critical next question to explore the distriution, movement and perception of this chemical signal in plants. Considering the chemical structures, more enzymes are likely to e required for a Control +GR24 ( µm) +GR24 (2 µm) d- d- d7- c d GR d- d- d7- (+)-Strigol.... Concentration (µm).. d- d- d7- (+)-5DS (.2 µm) Concentration (µm) d- d-.... f Height (cm) 4 2 g 8 4 h Relative D mrna level 2 d- d- week 2 weeks weeks 4 weeks 5 weeks weeks d- d d- d d- d- Figure Effect of strigolactones on rice tillering. a d, Two-week-old wild type () and d mutants. Scale ar in a indicates cm. Blue and red arrowheads (a) or ars ( d) indicate first and second tillers, respectively. Total numer of tillers (over 2 mm) in six seedlings is shown (mean s.d., n 5 ). e g, Six-week-old plants with () or without ( ) 2 mm GR24. Scale ar in e indicates cm. Weekly changes in the numer of tillers (f) and the plant height at the sixth week (g) is shown (mean s.d., n 5 4). h, D transcript levels in roots of 8-day-old seedlings with () or without ( ) mm ()-GR24 treatment for 24 h (mean and s.d., n 5 ). c Striga germination (%) No. of axillary shoots GR24 (5 µm) +GR24 (5 µm) max2- max- max DW S max2- max2- max- max2-4 max- max4-7 max-2 Figure 4 Effect of GR24 on axillary ud outgrowth of Araidopsis. a, Thirty-day-old wild type () and max mutants. Red arrowheads indicate outgrowth of axillary uds. Scale ar, cm., Numer of axillary shoots (over 5 mm) is shown (mean s.d., n 5 2 ). c, Estimation of germination stimulant levels in culture media of 2-week-old seedlings using Striga seeds (mean and s.d., n 5 ). DW, distilled water; S, ()-strigol ( 27 M). max4-7 max Macmillan Pulishers Limited. All rights reserved

5 NATURE Vol 455 Septemer 28 ARTICLES a Strigol ( ) c Strigol (+) Relative numer of S. hermonthica (%) * d- d- * * * NG SC LD D d- d- NG SC LD D Figure 5 Infection of rice d mutants y Striga hermonthica. a, Striga parasitizing to rice roots 4 weeks after inoculation. Scale ar, 5 mm., c, Ratio of Striga plants at each developmental stage 2 weeks after the inoculation of ()-strigol-treated (c) or non-treated () seeds (mean and s.d.;, n 5 7; c, n ). D, died after penetration; LD, leaf developed after the estalishment of parasitism; NG, no germination; SC, penetration succeeded and seed coat remained attached;, wild type. Asterisks indicate significantly different from wild type (Student s t-test, P,.5). the iosynthesis of strigolactones, in addition to CCD7, CCD8 and MAX (Fig. a). Identification of genetic loci defined y additional ranching mutants 4,5 may reveal new enzymes in the strigolactone iosynthesis pathway. Shoot ranching is influenced y a wide range of environmental signals 7. ur findings suggest that strigolactones may have a principal role in mediating the detection of nutrient availaility y roots and the resulting alterations in shoot architecture, provided that strigolactone levels are increased in response to inorganic phosphate deficiency, particularly in hosts of aruscular mycorrhizal fungi 2,28,, (Fig. 2d f); on inorganic phosphate (and possily other nutrients) deficiency, a proale adoptive strategy of plants would e to synthesize strigolactones for minimizing shoot ranching and maximizing the symiotic interaction with aruscular mycorrhizal fungi that facilitate the uptake of mineral nutrients. Seeds of root parasitic plants ause these chemical signals secreted for the successful symiosis with aruscular mycorrhizal fungi to find their potential hosts in soil. In many parts of the world, the parasitic weeds Striga and roanche are serious agricultural pests 4,5. Strigolactones have een an important target for parasitic weed control in generating lowgermination-stimulant varieties 2. Although strigolactones have een chemically recognized for decades, the iosynthetic pathway had not een genetically defined. The identification of several D/MAX loci as strigolactone iosynthesis genes now allows us to take a first step towards designing new varieties with reduced risk of parasite infections in molecular reeding. In fact, our results show that, at least in an experimental condition, the rice d- mutant is infected y significantly fewer S. hermonthica plants in comparison with wild-type plants, as a consequence of decreased germination frequency of the parasite seeds near the host root (Fig. 5). The use of strigolactonedeficient mutants will also facilitate our understanding of the exact roles of this class of terpenes in communication with aruscular mycorrhizal fungi in the rhizosphere. METHDS SUMMARY Plant materials. Rice and Araidopsis mutants used in this study are shown in Supplementary Fig.. Mutations in new mutant alleles used for this study were determined y DNA sequencing. Genotyping was carried out y a PCR-ased method using the primers listed in Supplementary Tale. Growth conditions and strigolactone treatment. Rice and Araidopsis seeds were surface-sterilized and the seedlings were first grown aseptically on agar media. Plants were then grown hydroponically in growth chamers. For oth rice and Araidopsis, strigolactones were added to the hydroponic culture medium. Strigolactone analysis. The levels of strigolactones released to hydroponic culture media were estimated y germination-stimulating activity using S. hermonthica seeds as descried previously 2. Strigolactones were identified and quantified on LC/MS MS y comparing the retention time and full-scan spectrum with those of authentic standards. We synthesized deuterium-laelled epi-5ds (d -epi-5ds) and used it as an internal standard for quantitative analysis using LC/MS MS. Gene expression analysis. We performed qrt PCR analysis to determine D transcript levels, according to the method descried previously 8. Full Methods and any associated references are availale in the online version of the paper at Received 4 June; accepted 2 July 28. Pulished online August Macmillan Pulishers Limited. All rights reserved. Beveridge, C. A., Ross, J. J. & Murfet, I. C. Branching mutant rms-2 in Pisum sativum. Grafting studies and endogenous indole--acetic acid levels. Plant Physiol. 4, (994). 2. Beveridge, C. A., Ross, J. J. & Murfet, I. C. Branching in pea (Action of Genes Rms and Rms4). Plant Physiol., (99).. Beveridge, C. A., Symons, G. M. & Turnull, C. G. Auxin inhiition of decapitationinduced ranching is dependent on graft-transmissile signals regulated y genes Rms and Rms2. Plant Physiol. 2, (2). 4. Johnson, X. et al. Branching genes are conserved across species. Genes controlling a novel signal in pea are coregulated y other long-distance signals. Plant Physiol. 42, 4 2 (2). 5. Stirnerg, P., van De Sande, K. & Leyser,. MAX and MAX2 control shoot lateral ranching in Araidopsis. Development 29, 4 (22).. Sorefan, K. et al. MAX4 and RMS are orthologous dioxygenase-like genes that regulate shoot ranching in Araidopsis and pea. Genes Dev. 7, (2). 7. Booker, J. et al. MAX/CCD7 is a carotenoid cleavage dioxygenase required for the synthesis of a novel plant signaling molecule. Curr. Biol. 4, (24). 8. Booker, J. et al. MAX encodes a cytochrome P45 family memer that acts downstream of MAX/4 to produce a carotenoid-derived ranch-inhiiting hormone. Dev. Cell 8, (25). 9. Turnull, C. G., Booker, J. P. & Leyser,. Micrografting techniques for testing long-distance signalling in Araidopsis. Plant J. 2, (22).. Snowden, K. C. et al. The Decreased apical dominance/petunia hyrida CARTENID CLEAVAGE DIXYGENASE8 gene affects ranch production and plays a role in leaf senescence, root growth, and flower development. Plant Cell 7, (25).. Simons, J. L., Napoli, C. A., Janssen, B. J., Plummer, K. M. & Snowden, K. C. Analysis of the DECREASED APICAL DMINANCE genes of petunia in the control of axillary ranching. Plant Physiol. 4, 97 7 (27). 2. Ishikawa, S. et al. Suppression of tiller ud activity in tillering dwarf mutants of rice. Plant Cell Physiol. 4, 79 8 (25).. Zou, J. et al. The rice HIGH-TILLERING DWARF encoding an ortholog of Araidopsis MAX is required for negative regulation of the outgrowth of axillary uds. Plant J. 48, (2). 4. Arite, T. et al. DWARF, an RMS/MAX4/DAD ortholog, controls lateral ud outgrowth in rice. Plant J. 5, 9 29 (27). 5. ngaro, V. & Leyser,. Hormonal control of shoot ranching. J. Exp. Bot. 59, 7 74 (28).. Schwartz, S. H., Qin, X. & Loewen, M. C. The iochemical characterization of two carotenoid cleavage enzymes from Araidopsis indicates that a carotenoidderived compound inhiits lateral ranching. J. Biol. Chem. 279, (24). 7. Auldridge, M. E. et al. Characterization of three memers of the Araidopsis carotenoid cleavage dioxygenase family demonstrates the divergent roles of this multifunctional enzyme family. Plant J. 45, (2). 8. Lechner, E., Achard, P., Vansiri, A., Potuschak, T. & Genschik, P. F-ox proteins everywhere. Curr. pin. Plant Biol. 9, 8 (2). 9. Cook, C. E. et al. Germination stimulants II. The structure of strigol a potent seed germination stimulant for witchweed (Striga lutea Lour.). J. Am. Chem. Soc. 94, (972). 2. Humphrey, A. J. & Beale, M. H. Strigol: Biogenesis and physiological activity. Phytochemistry 7, 4 (2). 2. Bouwmeester, H. J., Matusova, R., Zhongkui, S. & Beale, M. H. Secondary metaolite signalling in host-parasitic plant interactions. Curr. pin. Plant Biol., 58 4 (2). 22. Akiyama, K., Matsuzaki, K. & Hayashi, H. Plant sesquiterpenes induce hyphal ranching in aruscular mycorrhizal fungi. Nature 45, (25). 2. Bradow, J. M., Connick, W. J. Jr, Pepperman, A. B. & Wartelle, L. H. Germination stimulation in wild oats (Avena fatua L.) y synthetic strigol analogues and gierellic acid. J. Plant Growth Regul. 9, 5 4 (99). 24. Bradow, J. M., Connick, W. J. & Pepperman, A. B. Comparison of the seed germination effects of synthetic analogs of strigol, gierellic acid, cytokinins and other plant growth regulators. J. Plant Growth Regul. 7, (988). 25. Goldwasser, Y., Yoneyama, K., Xie, X. & Yoneyama, K. Production of strigolactones y Araidopsis thaliana responsile for roanche aegyptiaca seed germination. Plant Growth Regul. 55, 2 28 (28). 99

6 ARTICLES NATURE Vol 455 Septemer Yoneyama, K. et al. Strigolactones, host recognition signals for root parasitic plants and aruscular mycorrhizal fungi, from Faaceae plants. New Phytol. 79, (28). 27. Matusova, R. et al. The strigolactone germination stimulants of the plant-parasitic Striga and roanche spp. are derived from the carotenoid pathway. Plant Physiol. 9, (25). 28. López-Ráez, J. A. et al. Tomato strigolactones are derived from carotenoids and their iosynthesis is promoted y phosphate starvation. New Phytol. 78, (28). 29. Bouwmeester, H. J., Roux, C., Lopez-Raez, J. A. & Bécard, G. Rhizosphere communication of plants, parasitic plants and AM fungi. Trends Plant Sci. 2, (27).. Yoneyama, K. et al. Nitrogen deficiency as well as phosphorus deficiency in sorghum promotes the production and exudation of 5-deoxystrigol, the host recognition signal for aruscular mycorrhizal fungi and root parasites. Planta 227, 25 2 (27).. Yoneyama, K., Yoneyama, K., Takeuchi, Y. & Sekimoto, H. Phosphorus deficiency in red clover promotes exudation of oroanchol, the signal for mycorrhizal symionts and germination stimulant for root parasites. Planta 225, 8 (27). 2. Sugimoto, Y. & Ueyama, T. Production of ()-5-deoxystrigol y Lotus japonicus root culture. Phytochemistry 9, (28).. Zou, J. et al. Characterizations and fine mapping of a mutant gene for high tillering and dwarf in rice (ryza sativa L.). Planta 222, 4 2 (25). 4. Gressel, J. et al. Major heretofore intractale iotic constraints to Africa food security that may e amenale to novel iotechnological solutions. Crop Prot. 2, 89 (24). 5. Joel, D. M. The long-term approach to parasitic weeds control: manipulation of specific developmental mechanisms of the parasite. Crop Prot. 9, (2).. Xie, X. et al. 2 -Epi-oroanchol and solanacol, two unique strigolactones, germination stimulants for root parasitic weeds, produced y toacco. J. Agric. Food Chem. 55, (27). 7. Cline, M. G. Apical dominance. Bot. Rev. 57, 8 58 (99). 8. Magome, H., Yamaguchi, S., Hanada, A., Kamiya, Y. & da, K. dwarf and delayedflowering, a novel Araidopsis mutant deficient in gierellin iosynthesis ecause of overexpression of a putative AP2 transcription factor. Plant J. 7, (24). Supplementary Information is linked to the online version of the paper at Acknowledgements We are grateful to S. Ishikawa for sequencing the d7- allele; K. Fujiwara for assistance in preparing plant materials; N. Makita and H. Sakakiara for their advice on rice hydroponic culture; and Y. Tsuchiya for advice on germination assays. We thank the Salk Institute and the Araidopsis Biological Resource Center for providing Araidopsis T-DNA insertion lines; T. Yokota, K. Yoneyama and X. Xie for sharing information on strigolactone analysis; M. Maekawa for propagating rice seeds; and K. Mori, P. McCourt and A. Gaar Baiker for providing ()-strigol and 2 -epi-oroanchol, ()-GR24, and S. hermonthica seeds, respectively. This work was supported in part y grants from the MEXT of Japan (828 to K.Y., 978 to K.S. and 9784 to Sa.Y.) and the MAFF of Japan (Genomics for Agricultural Innovation, IPG to J.K.). M.U. is supported y the RIKEN Special Postdoctoral Researchers Program. Author Contriutions M.U. and T.A. developed and performed rice ranching assays. T.A. and J.K. prepared rice genetic materials. M.U. performed Araidopsis ranching assays. A.H. carried out LC/MS MS analysis. Sa.Y. and K.S. designed and performed S. hermonthica infection assays. K.A. synthesized laelled epi-5ds. N.T.-K. and H.M. carried out qrt PCR analysis. K.Y. and K.A. provided strigolactones and assisted strigolactone analysis y A.H. Y.K., K.S., K.Y. and J.K. contriuted to the experimental design. Sh.Y. directed the project and designed the experiments. Sh.Y. and M.U. wrote the manuscript. Author Information Reprints and permissions information is availale at Correspondence and requests for materials should e addressed to Sh.Y. (shinjiro@postman.riken.jp) Macmillan Pulishers Limited. All rights reserved

7 doi:.8/nature7272 METHDS Rice hydroponic culture. We used rice normal cultivars (ryza sativa L. cv. Shiokari and cv. Nipponare) and tillering dwarf mutants (Supplementary Fig. ) in this study. Rice seeds were washed in 7% ethanol for s, sterilized in 2.5% sodium hypochlorite solution for 5 min, rinsed with sterile water, and then incuated in sterile water at 28 uc in the dark for 2 days. Germinated seeds were transferred into hydroponic culture media 9 solidified with.% agar (ph 5.7) and cultured at 25 uc under fluorescence white light (5 2 mmol m 22 s 2 ) with a h light/8 h dark photoperiod for 5 days. Each seedling was then transferred to a glass vial containing a sterilized hydroponic culture solution ( ml), fixed with a piece of sponge at the root shoot junction to the top of the vial, and grown under the same condition for an additional 7 days (total 2 weeks). The hydroponic solution was supplemented every days. For large-scale cultures, the 2-week-old seedlings were transferred into a 4-l porcelain pot containing the same hydroponic solution and grown under the same condition. After the transfer to pots, the solution was renewed weekly. Araidopsis hydroponic culture. We used Araidopsis thaliana ecotype Col- as the wild type and max mutants (Supplementary Fig. ) in this study. Seeds were sterilized in % sodium hypochlorite solution for 5 min, rinsed with sterile water, and stratified for one day at 4 uc. The seeds were placed on the halfstrength Murashige and Skoog (MS) medium 4 containing % sucrose and.8% agar (ph 5.7) at 22 uc under fluorescence white light ( 7 mmol m 22 s 2 ) with a h light/8 h dark photoperiod for 5 days. Plants were then transferred to a glass pot containing 4 ml hydroponic solution 4 and grown under the same environmental condition for an additional 5 days. The solution was renewed every days. To measure germination stimulants, sterilized and stratified seeds were placed on glass eads ( ml) wetted with / strength MS liquid media ( ml) in a Petri dish (9 cm diameter) and grown for 4 days under the same conditions aove. The culture media were collected and sujected to S. hermonthica germination assay. LC/MS MS analysis. The hydroponic culture media were collected and extracted with ethyl acetate twice after adding d -epi-5ds as an internal standard. The ethyl acetate phase was concentrated in vacuo after drying over sodium sulphate. The roots were homogenized in acetone containing d -epi-5ds. The filtrates were dried up under nitrogen gas and dissolved in % acetone. The extracts were loaded onto asis HLB ml cartridges (Waters) and eluted with acetone after washing with de-ionized water. The eluates were loaded onto Seppak Silica ml cartridges (Waters), washed with ethyl acetate:n-hexane (5:85) and then eluted with ethyl acetate:n-hexane (5:5). The epi-5ds-containing fractions from culture media and roots were dissolved in 5% acetonitrile and sujected to LC/MS MS analysis using a system consisting of a quadruple/timeof-flight tandem mass spectrometer (Q-Tof Premier, Waters) and an Acquity Ultra Performance liquid chromatograph (Waters) equipped with a reversephase column (Acquity UPLC BEH-C 8, 2. 5 mm,.7 mm; Waters). The moile phase was changed from % acetonitrile containing.5% acetic acid to 4% and 7% in 5 and min after the injection, respectively, at a flow rate of.2 ml min 2. Data analysis was performed as we descried previously for gierellin analysis using MassLynx software (v. 4.) 42. Chemicals. GR24, 5DS and 5DS isomers were synthesized as descried previously 22,4.()-Strigol and 29-epi-oroanchol were provided y K. Mori. For experiments in Fig. h, we used ()-GR24 (courtesy of P. McCourt). The synthesis of d -(epi)-5ds was carried out as descried previously for non-laelled 5DS 22. The ABC ring was formylated with deuterium-laelled methyl formate and the following alkylation with racemic 4-romo-2-methyl-2-uten-4-olide provided [9-d]-5DS and its 29-epimer (Supplementary Fig. 2). ()-[9-d ]-epi- 5DS was purified y a silica gel column (Wakogel C-2, Wako Pure Industries; n-hexane-ethyl acetate stepwise) and semi-preparative high-performance liquid chromatography on reverse-phase (Inertsil DS-, GL Sciences; 7% acetonitrile in water) and normal-phase (Inertsil SIL-A, GL Sciences; 5% ethanol in n-hexane) columns. Germination assay. Germination assays using S. hermonthica were performed as descried previously 2. For each ioassay, de-ionized water and ()-strigol solution were used as negative and positive controls, respectively. Gene expression analysis. Total RNA was extracted from roots using the RNeasy Maxi kit (Qiagen). qrt PCR was carried out to determine D transcript levels using gene-specific primers and a Taq-Man proe (Supplementary Tale ). Uiquitin expression was used as an internal standard. S. hermonthica infection assay. S. hermonthica infections were analysed using a rhizotron system as descried previously 44, with slight modifications. Briefly, -week-old rice seedlings were transferred to root-oserving rhizotron chamers (225 mm 225 mm Petri dish filled with rockwool and nylon mesh) supplied with 5 ml half-strength MS media, and grown for 2 weeks in a green house with a 2 h photoperiod (7 45 mmol m 22 s 2 ) at day/night temperature cycles of 28 uc/2 uc. S. hermonthica seeds were preconditioned on moist glass fire filter papers (GF/A, Wattman) at 2 uc in the dark for 2 weeks, and treated with or without 29 M()-strigol for 5 h in the dark. After rinsing with excess water, approximately 5 parasite seeds were carefully placed along rice roots and the rhizotrons were incuated under the same growth conditions descried aove. The status of germination, infection and development of S. hermonthica was evaluated after 2 and 4 weeks of co-cultivation. 9. Kamachi, K., Yamaya, T., Mae, T. & jima, K. A role for glutamine synthetase in the recomination of leaf nitrogen during natural senescence in rice leaves. Plant Physiol. 9, 4 47 (99). 4. Murashige, T. & Skoog, F. A revised medium for rapid growth and ioassays with toacco tissue cultures. Physiol. Plant. 5, (92). 4. Norén, H., Svensson, P. & Andersson, B. A convenient and versatile hydroponic cultivation system for Araidopsis thaliana. Physiol. Plant. 2, 4 48 (24). 42. Varanova, M. et al. Methylation of gierellins y Araidopsis GAMT and GAMT2. Plant Cell 9, 2 45 (27). 4. Mangnus, E. M., Jan Dommerholt, F., de Jong, R. L. P. & Zwaneurg, B. Improved synthesis of strigol analogue GR24 and evaluation of the iological activity of its diastereomers. J. Agric. Food Chem. 4, 2 25 (992). 44. Gurney, A. L., Slate, J., Press, C. & Scholes, J. D. A novel form of resistance in rice to the angiosperm parasite Striga hermonthica. New Phytol. 9, (2). 28 Macmillan Pulishers Limited. All rights reserved