A femtomolar-range suicide germination stimulant for the parasitic plant Striga hermonthica

See allHide authors and affiliations

Science  14 Dec 2018:
Vol. 362, Issue 6420, pp. 1301-1305
DOI: 10.1126/science.aau5445

A step toward control of a noxious weed

The parasitic plant Striga hermonthica causes extensive crop losses, particularly in Africa. Strigolactone hormones can be used to initiate germination of Striga seeds when no host crop is present, which causes the nascent Striga plants to die. Unfortunately, strigolactones are also used by crop plants to establish beneficial mutualisms. Uraguchi et al. developed a hybrid molecule that can initiate Striga germination without interfering with strigolactone-dependent events in the host (see the Perspective by Bouwmeester). The compound has the potential to diversify routes toward protecting fields from Striga infestation.

Science, this issue p. 1301; see also p. 1248


The parasitic plant Striga hermonthica has been causing devastating damage to the crop production in Africa. Because Striga requires host-generated strigolactones to germinate, the identification of selective and potent strigolactone agonists could help control these noxious weeds. We developed a selective agonist, sphynolactone-7, a hybrid molecule originated from chemical screening, that contains two functional modules derived from a synthetic scaffold and a core component of strigolactones. Cooperative action of these modules in the activation of a high-affinity strigolactone receptor ShHTL7 allows sphynolactone-7 to provoke Striga germination with potency in the femtomolar range. We demonstrate that sphynolactone-7 is effective for reducing Striga parasitism without impinging on host strigolactone-related processes.

Striga hermonthica (Striga) parasitizes crops widely across various parts of sub-Saharan Africa, causing loss in crop yields that result in economic pressure on millions of smallholder farmers and lead to annual losses of billions of dollars (1). Protecting crops from the numerous tiny Striga seeds buried in the soil requires integration of various approaches to suppress infestation (1). A group of host-generated small-molecule hormones, called strigolactones (SLs), provoke germination of Striga seeds. Because Striga is an obligate parasite, germination in the absence of a host is lethal, and this has prompted researchers to develop SL agonists as inducers of suicidal germination to purge the soil of viable Striga seeds (2). This approach requires the development of potent and accessible compounds that only act on Striga and do not impede normal crop development. For example, SLs are also plant chemical cues that attract root symbiotic arbuscular mycorrhizal fungi (AM fungi) that supply host plants with nutrients (3, 4). Here, we report the development of a Striga-selective SL agonist acting in the femtomolar range.

SLs are a group of plant-derived molecules whose structures consist of butenolide rings (D-rings), which are connected to cyclic moieties, usually three-ring systems (ABC-rings), through an enol-ether bridge (Fig. 1A). In vascular plants, SLs are plant hormones that optimize plant body architectures through the DWARF14 (D14) family of α/β hydrolase-fold receptors (5). D14 defines a noncanonical receptor because it initiates signal transduction by using enzymatic activity. Upon binding, SLs undergo cleavage of the enol-ether bridge through hydrolysis to leave the D-ring as a covalently linked intermediate molecule (CLIM) at the catalytic histidine residue in the receptor (68). Previous studies suggest that the ABC-portion of the SL is released from the D14 pocket, and the receptor–CLIM complex alters D14 conformation to recruit downstream negative regulators such as the SCFMAX2 protein (7). In Striga, it is thought that SLs trigger seed germination through 11 members of an independently diverged α/β hydrolase-fold receptors called Striga HYPOSENSITIVE TO LIGHT/KARRIKIN INSENSITIVE2 (ShHTL/KAI2, here called “ShHTLs”) (911). The hydrolytic activity of ShHTLs was exploited in the development of fluorogenic SL probes to uncover an ethylene-mediated amplification of a wave-like pattern of SL perception initiated during Striga germination (10). Moreover, in vitro binding suggests that the divergence of ligand preferences in ShHTLs is beneficial for Striga seeds to detect the blend of SLs exuding from preferred host species (10). Among these ShHTL isoforms, we have focused on ShHTL7 because this receptor is sensitive to picomolar levels of SLs when heterologously expressed in Arabidopsis, and its large binding pocket ensures a response to structurally diverse molecules (11, 12). These characteristics make ShHTL7 a suitable target for the development of agonists for stimulating Striga germination.

Fig. 1 Development of a femtomolar-range germination stimulant for Striga.

(A) Scheme of structure development. MEC represents the lowest concentration of the compound that produces any seed germination. (B) SAM690 induces Striga seed germination at 10 μM. Scale bar, 1 mm. (C) 10 μM aminoethoxyvinyl glycine (AVG) suppresses (+)-GR24 and SAM690. (D) Striga germination in dilution series of SPL7, H-SPL7, 5DS, and (+)-GR24. (E) Competitive bindings to ShHTLs and AtD14. IC50 value (in micromolar) in the YLG assay is presented as a heat map with SD (n = 3 technical replicates). Data for 5DS was obtained from (10). Error bars in (C) and (D) indicate SD (n = 3 biological replicates).

Chemical analysis on SLs over the past 40 years suggests that the structure of the D-ring is essential to SL activity (2, 3). By contrast, structural flexibility in the ABC-portion has led to the development of various synthetic SLs or SL mimics, including GR24 or simplified phenol–D-ring derivatives called debranones (2, 13). However, the structural element of the ABC-portion that would contribute to both potency and specificity to Striga remains elusive. To explore the chemical characteristics that define species selectivity toward Striga, we performed a small-molecule screen for compounds that germinate Striga seeds (harvests from sorghum fields in Sudan). This screening of 12,000 synthetic molecules was followed by additional synthesis of 60 analogs of hit compounds that were found from the initial screening. On the basis of median inhibitory concentration (IC50) using the fluorogenic SL-mimic Yoshimulactone Green (YLG), the binding assay resulted in the identification of N-arylsulfonylpiperazine as a molecular scaffold that selectively bound to ShHTL7 (Fig. 1, A and B, fig. S1, and table S1). A representative molecule, SAM690, which contains the arylsulfonylpiperazine moiety, exhibited potency toward Striga germination at the micromolar level. The mode of action of SAM690 was similar to that of (+)-GR24, in that germination activity was suppressed by inhibition of ethylene production (Fig. 1C). However, unlike (+)-GR24, SAM690 was not hydrolyzed by ShHTL7 (fig. S2) (10). These observations indicate that SAM690 stimulates Striga germination with selective activation of ShHTL7 through a mechanism independent of hydrolysis.

During a series of above assays, we noticed inconsistency in stimulant activities of several SAM690 derivatives depending on the purification method due to active impurity. This byproduct, although only 0.01% of the total product, appeared to be an unusually oxidized molecule that has a hybrid structure resembling SAM690 with a D-ring–like butenolide moiety (Fig. 1A and fig. S3). In order to verify the structure and potency of this derivative, we established a three-step synthetic procedure, and the resulting oxidized SAM690 exhibited potency comparable with that of (+)-GR24, as evident from its minimum effective concentration (MEC) of 10 pM (Fig. 1D). As expected from its structure, oxidized SAM690 was hydrolyzed by ShHTL7 (fig. S2). The structural similarity of this compound to SLs led us to hypothesize that attaching a methyl group to the C4′ position may enhance the potency of the molecules. Indeed, this modification improved MEC from 10 pM to 10 fM (Fig. 1, A and D). We named the D-ring/sulfonylpiperazine-hybrid molecule sphynolactone-7 (SPL7) and named its demethylated analog H-SPL7 (sulfonylpiperazine hybrid strigolactone mimic of ShHTL7) (the stability and toxicology of SPL7 are summarized in fig. S5). The name is derived from the sphinx, a mythical creature with the head of a human and the body of a lion, to represent the hybrid nature of the molecule. The IC50 values of SPL7 improved from SAM690 (0.31 versus 8.9 μM), and our liquid chromatography–mass spectrometry (LC-MS) analysis revealed that SPL7 was hydrolyzed by ShHTL7 to form CLIM at the catalytic histidine residue (Fig. 1E and figs. S2 and S4) (7, 14). The potency of SPL7 is comparable with that of (+)-5-deoxystrigol (5DS), a natural SL that is currently the most potent commercially available germination stimulant for Striga.

Despite their high potencies, the presence of the N-arylsulfonylpiperazine scaffold allows SPL7 to retain selectivity toward ShHTL7, whereas 5DS binds to all the SL receptors with different ranges of IC50 values (Fig. 1E) (10). To gain insight into this difference in selectivity, we replaced 16 active-site residues of ShHTL7 with those of ShHTL5 (11). Using the YLG binding assay, we identified seven residues that are essential for the binding with SPL7 (M139, T142, T157, L161, Y174, C194, and M219) (Fig. 2, A and B, and fig. S6). The combination of these mutations led to a distribution of IC50 values of SPL7, which was correlated with that of H-SPL7 [correlation coefficient (r = 0.81)] but not with that of 5DS (r = 0.15) (Fig. 2C). These results indicate that SPL molecules use a different subset of residues for binding compared with those of natural SLs, displaying selectivity. Our computational investigation supports the hypothesis that SPL7 could fit to the active site of the homology model of ShHTL7, whereas changes in polarity and volume through active-site mutations may impair its fit (Fig. 2A and fig. S7). These seven amino acids as a combination are specific to ShHTL7 among known HTL/KAI2 homologs, including those from a parasitic plant Orobanche minor, which also uses SLs as germination stimulants (fig. S8) (3, 9). Consistently, SPL7 exhibits nanomolar-level potency to O. minor and is effective at femtomolar range for several S. hermonthica ecotypes that parasitize to different hosts (fig. S8).

Fig. 2 Active-site residues differentiating selectivity of SPL7 and 5DS.

(A) Homology models of ShHTL7 and its septuple mutant with mutated amino acids located in the active sites. Brown circles indicate polar to nonpolar mutations. The yellow dotted circle indicates reduction of the pocket volume by T157Y. (T157Y indicates that threonine at position 157 was replaced by tyrosine). Single-letter abbreviations for the amino acid residues are as follows: C, Cys; L, Leu; M, Met; S, Ser; T, Thr, and Y, Tyr. (B) IC50 values (in micromolar) in the YLG assay with the mutant series of ShHTL7. Sixteen active-site residues were replaced with those corresponding to ShHTL5. Quadruple, hextuple, and septuple mutants are shown with SD (n = 3 technical replicates). (C) Distribution of IC50 values (in micromolar) in the series of ShHTL7 mutants.

Because SPL7 and GR24 have identical D-ring structures, the selectivity to ShHTL7 and the femtomolar-range potency must be encoded in the ABC-portion of SPL7 (Fig. 3A). In light of an activation model solely dependent on CLIM formation as proposed in D14, the ABC-portion of SPL7 possibly contributes to efficient CLIM formation on the receptor (7, 14). Alternatively, the ABC-portion may have additional functions other than accelerating CLIM formation. We assessed these possibilities through investigation of the relationship between potencies and D-ring hydrolysis using various SPL7 analogs. The potencies of two hydrolysis-resistant analogs, carba-H-SPL7 and 1′-carba-SPL7, were ≥100 nM, implying that the hydrolysis of D-ring is dispensable for activity yet essential to gain the femtomolar-level potency (Fig. 1A and fig. S9). Next, to investigate the quantitative relationship between potencies and the hydrolysis reaction rate, we performed a kinetic analysis similar to that involving surface plasmon resonance, which allows estimation of reaction rate constants k1 and k−1 independently (15). Briefly, we obtained the parameter k1CLIM and (k−1CLIM + k2) by fitting an equation formularized from a reaction scheme in Fig. 3B to experimentally obtained time-dependent CLIM-formation curves (supplementary materials, materials and methods) (8). We assumed (k−1CLIM + k2) ≈ k−1CLIM because observed stability of CLIM-ShHTL7 complex over 30 min theoretically limited k2 to <1% fraction of (k−1CLIM + k2) in our analysis. The kinetic analysis with SPL7 analogs allowed us to observe only a vague trend between potency and k1CLIM (r = −0.32), indicating that the rate of CLIM formation, although important, was not a sole factor for determining potency (Fig. 3B and figs. S10 and S11). This interpretation was supported by the observation with GR24, in which the reaction rate of the CLIM formation was higher (k1CLIM = 316 × 10−3/μM/s) than that of SPL7 (k1CLIM = 43.5 × 10−3/μM/s) despite a potency 1000 times lower than that of SPL7 (Figs. 1D and 3, B and C). These results are contradictory to the model proposed for D14, thus indicating that the ABC-portion of SPL7 has additional functions other than accelerating CLIM formation for delivering the difference in potency (7, 14). Although difference in the uptake or stability in Striga seeds could account for differences in potency, we obtained no positive results supporting this assumption (fig. S12). On the basis of these observations, we hypothesized that the function of the ABC-portion after the hydrolysis is essential to deliver femtomolar-level potency (fig. S13). Verification of this model will require detailed studies on the metabolic fate of SPL7 and crystallization of SPL7-ShHTL7 complex.

Fig. 3 Mode of action of SPL7.

(A) Annotation of structural modules identified from the structure–activity–relationship study. (B) Relationship between reaction rate constants and MEC among SPL7 analogs. (Top) Reaction scheme and (Bottom) scatter plot of k1CLIM or k−1CLIM against MEC of Striga germination are presented. (C) Time-dependent CLIM formation quantified by LC-MS. T50 indicates the half-maximal time. Error bar indicates SD (n = 3 technical replicates).

We next tested the utility of SPL7 as a Striga-selective suicide germination stimulant, using three organism-based bioassays. First, we applied 10 μM SPL7 to a SL biosynthetic mutant, more axillary growth4-1 (max4-1), to see whether SPL7 restores the increased branching phenotype (16). SPL7 failed to rescue max4-1 branching defects, although a similar concentration of GR24 did suppress axillary branch emergence (Fig. 4A). SPL7 also failed to induce root hair elongation or induce SL-inducible gene expressions in wild-type Arabidopsis (Fig. 4, B and C) (17, 18). Thus, SPL7 exhibits no hormonal SL activity in Arabidopsis assays. Second, we evaluated the effect of SPL7 on AM fungi, which are agronomically important microbes that support the growth of crops. Whereas SLs induced multiple 3° hyphal branches as in Medicago root exudate, SPL7 exhibited only a mild effect at the highest concentration, showing 800 times less activity than that of (+)-GR24 (Fig. 4D) (19). Last, we evaluated the ability of SPL7 to induce suicide germination of Striga in a pot infestation assay (Fig. 4, E and F). In the dimethyl sulfoxide (DMSO) control, Striga seeds parasitized maize and emerged from the soil at an average of one seedling per host. Soil treatment with SPL7 at a concentration of 100 pM or higher for a week before planting maize reduced the emergence of Striga and protected the host plants from senescence caused by parasitism. By contrast, GR24 requires 10 nM to obtain similar effect. Taken together, we concluded that SPL7 is effective as a Striga-selective suicide-germination stimulant, at least in laboratory experiments.

Fig. 4 Bioassays with SPL7.

(A) SPL7 does not suppress shoot-branching phenotype of Arabidopsis SL biosynthetic mutant, max4-1, at 10 μM. Arrowheads indicate axillary branches. Average numbers of axillary branches are indicated with SE; n indicates number of plants tested. Scale bar, 5 cm. (B) SPL7 fails to enhance root hair elongation in Arabidopsis wild-type at 10 μM. Average length of root hair is presented with SD (n = 7 biological replicates). Scale bar, 100 μm. (C) SPL7 fails to induce SL-inducible BRANCHED1 (BRC1) expression in Arabidopsis wild-type at 10 μM. Average expression obtained from quantitative reverse transcription polymerase chain reaction (RT-PCR) analysis is presented as relative value to DMSO control with SD (n = 3, biological replicates). (D) SPL7 shows 800 times less potency for AM fungi than that of (+)-GR24. MEC represents the lowest concentration of compound that induces multiple 3° hyphae. Data for (+)-GR24 were obtained from (19). Scale bar, 1 mm. (E) Suicide germination assay. Representative pictures taken after 2 months (left) or 3 months (right) of cocultivation of maize with Striga. The soil was pretreated with DMSO or 10 nM of SPL7. Arrowheads indicate emerged Striga. Scale bar, 5 cm. (F) Number of emerged Striga after 2 months of cocultivation. n indicates number of hosts tested. Error bar indicates SE.

The discovery of SPL7 reinforced the design principle of SL mimics as a hybrid of two functional modules, a modifiable synthetic scaffold responsible for both receptor selectivity and potency as the ABC-portion and the D-ring component of natural SLs. Implications of the strategy for basic science includes direct dissection of the roles of specific SL receptors in experimentally intractable organisms such as Striga. For practical purpose, the strategy appears applicable to other noxious parasitic weeds, including Orobanche or Phelipanche species.

Supplementary Materials

Materials and Methods

Figs. S1 to S13

Table S1

References (2030)

References and Notes

Acknowledgments: We thank A. Babiker, S. Runo, and P. Matana for providing the S. hermonthica seeds; K. Yoneyama for providing O. minor seeds; and S. Hagihara and M. Yoshimura for providing YLG. We thank N. Nakamichi for instructing quantitative RT-PCR analysis and J. X. Yap for supporting biochemistry works. We thank A. Miyazaki for proofreading. Authors contributions: The chemical aspect of the research was managed by D.U. and T.O. Conceptualization of the project and the management of biological aspect of the research was performed by Y.T. Chemical screening was performed by Y.T. under the supervision of P.M. and T.K.; Y.T. and H.I. performed Striga germination assays and YLG assays. S.AM. synthesized analog series of initial hits, and R.Y. synthesized SPL7 analogs under the supervision of D.U. and T.O. Arabidopsis assays, quantitative RT-PCR, and suicide germination assay were performed by H.I. under supervision of Y.T.; N.M. performed hyphal branching assay with AM fungi under the supervision of K.A.; K.K. performed LC- MS analyses for small molecules and proteins. Mathematical characterization of CLIM formation was performed by Y.H. Homology model and docking simulations were performed by C.R. under supervision of S.I.; Y.T. wrote the overall story of the manuscript. The manuscript was edited by D.U., P.M., T.K., and T.O. All the authors discussed the manuscript. Funding: This work was supported by a Grant in Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology (15KT0031 and 15K07102 to Y.T. and 15H059556 to T.K.) and a grant from the Advanced Low Carbon Technology Research and Development Program from the Japan Science and Technology Agency (T.K.). Support for C.R. from the Japan Society for the Promotion of Science (JSPS) and the Alexander von Humboldt Foundation (AvH) is gratefully acknowledged. P.M. was funded by the Natural Sciences and Engineering Research Council of Canada (NSERC). ITbM is supported by the World Premier International Research Center Initiative (WPI), Japan. Competing interests: Nagoya University has filed for patents regarding the following topics: “Regulators for germination in Striga species,” inventors Y.T., D.U., S.AM., S.H., M.Y., T.K., T.O., and K.I. (patent publication nos. WO 2017/002898 and JP 2017-014149); “Regulators for germination in parasitic plants,” inventors Y.T., D.U., T.O., T.K., and K.K. (patent application nos. PCT/JP2018/36785 and JP 2017-193773). We declare no financial conflicts of interest in relation to this work. Data and materials availability: All data are available in the manuscript or the supplementary materials. The complete sets of raw data underlying all figures in the main text and supplement can be found in the supplementary materials.
View Abstract

Stay Connected to Science

Navigate This Article