Small Molecule Toolbox for Strigolactone Biology

Abstract

Strigolactones (SLs) are plant hormones associated with diverse developmental processes including plant architecture and stress responses. SLs are exuded to the soil as an ecological signal to attract symbiotic arbuscular-mycorrhizal fungi. This ecological mechanism is also used by parasitic plants to detect the presence of host plants and initiate germination. The functional diversity of SLs makes SL biology so extensive that a single methodology is not sufficient to comprehend it. This review describes the theoretical and practical aspects of the design of small molecule probes that have been used to elucidate the functions of SLs. The lessons from the development of small molecules to tackle the unique questions in SL biology might be instructive in the extending field of chemical biology in plants.

Introduction

Biological processes involving contact of small organic molecules with proteins are essential elements of life. Small signaling molecules, including plant hormones, bind to their specific receptors and trigger sequential signal transduction events to induce cellular responses, to regulate development of organisms or to recognize and respond to other organisms in the surrounding environment. Biosynthesis of such signaling molecules requires binding of the precursor molecules to biosynthetic enzymes, while allocating the molecules to the place where they function also requires their binding with specific transporters. By nature, approaches to these biological processes with synthetic small molecules are powerful in elucidating how such interactions lead to phenotypic consequences. The utility of small molecule probes is not limited to revealing loss-of-function phenotypes, but they can also be used to over-ride genetic redundancy by inhibiting multiple proteins with related sequences, or by selectively activating a specific isoform of redundant receptors (Park et al. 2009). In broader biology, synthetic molecules are useful for translational studies from model to non-model plants. Conversely, some small molecules can be used to distinguish phenotypes caused by a single nucleotide polymorphism within the same species (Zhao et al. 2007). Moreover, small molecules can have additional functionalities, e.g. optical control or emission of fluorescence, which enable precise control or visualization of biological processes (Chan et al. 2012). Because these biological processes largely rely on the contact of small molecules with proteins, the central issue in the field is how we find and design appropriate structures, and manipulate them to reveal novel mechanisms.

In this respect, the plant-derived signaling molecules known as strigolactones (SLs) are intriguing and important for a number of reasons. First, SLs regulate diverse developmental processes including suppression of shoot branching, root architecture, leaf senescence, secondary growth and stress responses (Gomez-Roldan et al. 2008, Umehara et al. 2008, Kapulnik et al. 2011, Agusti et al. 2012, Ha et al. 2014, Yamada et al. 2014). Additionally, SLs are released into the soil and function as communicative signals with soil microbes, including symbiotic arbuscular-mycorrhizal fungi (AM fungi) (Akiyama et al. 2005). Moreover, SLs released in the soil are sensed by seeds of parasitic plants as germination cues (Cook et al. 1966). Therefore, SLs produced by plants induce different, but mutually associated, responses by three classes of organisms. The utility of small molecule probes for cross-species functional studies enables the elucidation not only of the mechanisms of SL biosynthesis and signaling but also of evolutionary or ecological questions, which is difficult using genetics alone. Secondly, the mechanism of SL perception by DWARF14 (D14) α/β hydrolase-fold proteins is unusual as it involves enzymatic hydrolysis of SLs (Arite et al. 2009, Hamiaux et al. 2012). A crystallographic study suggested that the D-ring is covalently attached to the receptor pocket and triggers the association of the receptor with the F-box protein, MORE AXILLARY GROWTH2/DWARF3 (MAX2/D3), to ubiquitinate the downstream regulator DWARF53/SUPPRESSOR OF MAX2 1/SMAX-Like (D53/SMAX1/SMXL) proteins (de Saint Germain et al. 2016, Yao et al. 2016). However, why the hydrolysis of SLs is essential for their activity is unclear; indeed, this is the central issue in SL research today. Because there has been a wealth of knowledges in manipulating enzymatic reactions from the viewpoint of chemistry, incorporating the idea into agonists or antagonists of the SL receptors might contribute to clarifying the issue. Thirdly, the ancestral α/β hydrolase-fold protein gene known as HYPOSENSITIVE TO LIGHT/KARRIKIN INSENSITIVE2 (HTL/KAI2) defines a parallel, yet associated, pathway for responding to germination stimulants including non-natural isomers of SLs, smoke-derived karrikins and other synthetic molecules (Tsuchiya et al. 2010, Waters et al. 2012, Toh et al. 2014). Although the involvement of MAX2 as a shared component of the pathway was revealed, their endogenous ligands (known as karrikin-like or KL) and the relationship between the hydrolase activity and signal transduction are unknown (Conn and Nelson 2016). Intriguingly, the SL receptors in parasitic plants did not diverge from D14, but from HTL/KAI2 (Conn et al. 2015, Toh et al. 2015, Tsuchiya et al. 2015). Additionally, parasitic plants generally have a large number of copies, e.g. 11 HTL/KAI2 homologs have been identified in Striga hermonthica (Conn et al. 2015, Tsuchiya et al. 2015). Synthetic molecules might thus be instructive in addressing evolutionary questions related to plant parasitism. Fourthly, and most importantly, SLs are key molecules for solving the devastating problem caused by parasitic plants of the genera Striga, Orobanche and Phelipanche (Ejeta 2007). These plants parasitize economically important crops and cause massive reductions in their production. In particular, the damage caused by S. hermonthica in sub-Saharan Africa is estimated at US$10 billion dollars annually, and 1 million people are adversely affected by the parasite (Gurney et al. 2006, Ejeta 2007). The dormant seeds can survive in soil for at least 20 years, while false germination in the absence of a host results in death of the seed within 4 d (Butler 1994). Therefore, synthetic agonists for SL receptors can be used to induce suicide germination by the seeds of parasitic plants. For these reasons, use of small molecule probes in studies of SL biology will enhance food security for the growing population of Africa.

Since the discovery of strigol in 1966 as a host-derived stimulant of S. lutea germination, a number of synthetic SLs have been developed (Cook et al. 1966). In the past decade, molecular identification of the factors involved in SL biosynthesis and signaling has enabled systematic development of small molecule probes using molecular design-based approaches and/or chemical screening. Moreover, incorporation of hydrolysis-dependent functionalities into synthetic SL agonists has begun to illuminate the mechanisms of SL signaling. This review summarizes and updates the contribution of chemical biology to understanding the functionalities of SLs. Despite the rapid progress in understanding of the molecular mechanisms of SL biosynthesis and signaling, extending this to broader biology that requires assembly of data from multiple species is challenging. One of the aims of this review is to propose that small molecule probes can be used for research on the roles of SLs in diverse biological systems. The findings may be useful in the expanding field of plant chemical biology.

Design and Utility of SL Biosynthesis Inhibitors

SLs are small molecules with a butenolide ring (D-ring) substructure that is linked to the rest of the molecule by an enol–ether bridge (Xie et al. 2010) (Fig. 1). In naturally occurring SLs, the D-ring is connected to the ether oxygen in the R configuration at the 2' position of the D-ring. Chemical synthesis of SLs generates a racemic mixture with the S-configured enantiomer, although such stereoisomers have not been discovered in nature. The majority of SLs discovered to date have a tricyclic lactone (ABC-ring) as a backbone with variable modifications on the AB-ring. However, SLs with only the A-ring or with other ring systems have recently been discovered in nature (Yoneyama et al. 2018). In total >20 SLs have been identified, and each plant species produces a unique SL profile (Wang and Bouwmeester 2018). Although the physiological relevance of the structural diversity of SLs is unclear, recent research has revealed how plants produce the diverse structures. This new knowledge has provided opportunities to dissect SL biology with small molecule probes.

Fig. 1

The SL biosynthetic pathway and its inhibitors.

The SL biosynthetic pathways were determined almost 40 years after their discovery. Because SLs were known only as germination stimulants for parasitic plants, genetic screening for mutants in SL biosynthesis was at best extremely tedious and at worst impossible. The first indication of the SL biosynthetic pathway came from a study using a small molecule called fluridone, an inhibitor of the phytoene desaturase involved in carotenoid biosynthesis (Matusova 2005) (Fig. 1). Because SLs are classified as terpenoids and exhibit a degree of structural similarity to carotenoid-derived molecules including ABA, the SL biosynthesis pathway was hypothesized to involve carotenoid intermediates. Indeed, the root exudate of fluridone-treated maize had a reduced ability to induce germination of S. hermonthica. Because fluridone causes bleaching and other deleterious effects on plant growth, it was difficult to conclude from these data alone that SLs are produced via the carotenoid pathway. However, the idea was subsequently linked to independent studies with increased shoot branching mutants that had predicted the existence of carotenoid-derived unknown signaling molecules.

Since the unknown molecules appeared to be SLs, subsequent genetic studies revealed the overall SL biosynthetic pathway (Seto and Yamaguchi 2014). The biosynthetic pathway as briefly introduced below was established by the identification of biosynthetic mutants with a characteristic bushy phenotype, which are referred to as more axillary growth (max) in Arabidopsis, dwarf (d) in rice, ramosus (rms) in pea and decreased apical dominance (dad) in petunia. The initial step in SL biosynthesis after branching from the carotenoid pool is mediated by the D27 family of iron-containing isomerases that convert all-trans-β-carotene to its 9-cis isomer (Lin et al. 2009, Alder et al. 2012). Next, a pair of carotenoid dioxygenase (CCD) family proteins, CCD7 and CCD8, catalyzes sequential oxidative cleavage and cyclization reactions to generate the key intermediate, carlactone (Alder et al. 2012). The carlactone biosynthesis pathway is conserved among plant species, but the downstream steps vary among plant species. In rice, the Cyt P450 monooxygenase, carlactone oxidase (also known as Os900), catalyzes the generation of tricyclic lactone in the ABC-ring, resulting in production of 4-deoxyorobanchol (4DO), a bioactive SL (Zhang et al. 2014). 4DO is further oxidized and hydroxylated by another P450 monooxygenase, orobanchol synthase (also known as Os1400), which attaches a hydroxyl group to the B-ring to generate the SL orobanchol. In Arabidopsis, in contrast, carlactone is oxidized by MAX1 P450 monooxygenase to carlactonoic acid (Abe et al. 2014). The downstream step is unclear, but it is assumed that carlactonoic acid is methylated by an unknown methyl transferase, and the methyl carlactonic acid reportedly binds SL receptors and activates signal transduction (Abe et al. 2014). The methyl ester is further metabolized by LATERAL BRANCHING OXIDOREDUCTASE (LBO) to generate an unidentified molecule with a 16 Da greater mass (Brewer et al. 2016). Although the structure is unknown, or indeed whether it is the final product, the predicted oxidization and hydroxylation activity of LBO, along with the increased shoot branching of loss-of-function mutants, suggests that the product might be a novel active SL with an oxygen atom inserted somewhere in the methyl carlactonoic acid. In addition to these SL biosynthetic enzymes, a gene encoding a protein with a sulfotransferase domain, known as LOW GERMINATION STIMULANT1 (LGS1), was discovered in Striga-resistant sorghum cultivars (Gobena et al. 2017). Mutations in this gene result in reduced production of 5-deoxystrigol (5DS), accompanied by increased production of orobanchol, which has a different stereochemistry from 5DS at the B–C ring junctions. While its biochemical functions are unknown, LGS1 is assumed to be a regulator of, or be directly involved in the generation of, stereochemistry. Although further investigation of these biosynthetic pathways is required, their diversity is responsible for the plant species specificity of SL profiles.

In comparison with genetic studies, the background for developing the inhibitors for SL biosynthesis may be trivial as long as we have good end-products. However, it is interesting in that they provide deeper understanding of the properties of the binding pockets. This is analogous to protein engineering through directed evolution, which provides information about how proteins gain new functions through gradual accumulation of mutations.

As described above, fluridone was the first SL biosynthesis inhibitor identified. However, inhibition of carotenoid biosynthesis exerts pleiotropic effects on the metabolism of carotenoid-derived molecules, including ABA and components of the photosynthetic apparatus, and causes photo-oxidative damage. Therefore, SL biosynthesis inhibitors must target enzymes at or downstream of the 9-cis isomerization. The first such inhibitors were discovered in a study of ABA biosynthesis inhibitors (Kitahata et al. 2011). As with SLs, the biosynthesis of ABA also involves oxidative cleavage of 9-cis carotenoids, 9-cis-violaxanthin or 9-cis-neoxanthin, as a key step, which is mediated by 9-cis-epoxycarotenoid dioxygenases (NCED) belonging to the same family as CCDs (Schwartz 1997). The small molecule abamine was initially developed as an NCED inhibitor by structural optimization from an inhibitor for lignostilbene-α,β-dioxygenase that oxidatively cleaves the C–C double bond of stilbene by a mechanism similar to NCED (Han et al. 2004). Treatment of Arabidopsis with abamine reduced the level of ABA, leading to an ABA-deficient phenotype including reduced drought tolerance, increased stomatal aperture and elongation of radicles (Han et al. 2004). However, abamine-treated rice also exhibited a reduced level of 4DO, resulting in decreased seed germination in the parasitic plant Orobanche minor (Kitahata et al. 2011). These findings were confirmed in a report showing that abamine inhibited both NCED1 and CCD8 in vitro (Harrison et al. 2015). Therefore, abamine inhibits biosynthesis of both ABA and SLs. Although the promiscuity of abamine hampers dissection of SL-related processes, these studies demonstrated that its structure provides a good lead for the development of selective inhibitors of NCEDs or CCDs. In terms of ABA, abamineSG, which has three additional carbons in the linker between the methyl ester and the nitrogen, exhibited enhanced inhibition of NCED, but reduced inhibition of CCD7 (Kitahata et al. 2006). In contrast, hydroxamic acid inhibitors, in which the linker is substituted with a hydroxyl group, called D6, preferentially inhibit CCD7/CCD8 over NCED in vitro, and result in increased shoot branching in Arabidopsis (Harrison et al. 2015). Therefore, the length and composition of the linker are critical determinants of the activity for NCED and CCD. Interestingly, other hydroxamic acid analogs inhibited D27 and CCD8, and some preferentially inhibit D27 over CCD8 (B2 in Fig. 1) (Harrison et al. 2015). This is somewhat surprising, as the amino acid sequences of D27 and CCD8 do not exhibit a high level of similarity. However, it may be explained by the fact that both enzymes contact 9-cis carotenoids.

The second class of SL biosynthetic inhibitors developed were targeted P450 monooxygenases. It is well documented that some azole-containing small molecules act as effective inhibitors for P450s by co-ordination of the azole nitrogen to the heme iron at the catalytic center of the enzyme, e.g. inhibition by paclobutrazol, brassinazole and abscinazole-E2B of gibberellin biosynthesis, brassinosteroid (BR) biosynthesis and ABA catabolism, respectively (Asami et al. 2000, Rademacher 2000, Okazaki et al. 2012). In principle, optimization for the substrate-recognizing site in the pocket of MAX1 or related P450s might enable development of selective inhibitors of SL biosynthesis. Based on this idea, SL biosynthesis inhibitors were screened for in a library of azole-containing small molecules for an increased tiller number in rice (Ito et al. 2010). However, the screening again involves issues with cross-selectivity among P450s, as some putative inhibitors induced an increased number of tillers and a reduced plant height, most probably the result of retardation of both SL and gibberellin biosynthesis. As in the case of CCD inhibitors, difficulty in chemically dissecting the SL pathway from that for gibberellin suggests that these P450s have similar binding pockets. Eventually, TIS13 was selected as the most potent candidate, and structural optimization to improve its selectivity gave rise to TIS108 (Ito et al. 2011).

SL biosynthesis inhibitors have begun to reveal the role of SLs in non-model plants including Sesbania cannabina, a soil-improving legume, in which SL production is associated with the response to salt stress (Kong et al. 2017). An intriguing example was provided with the parasitic plant Phelipanche aegyptiaca, whose germination was stimulated by TIS108 (Bao et al. 2017). This is unexpected, as an inhibitor should exert an effect opposite to that of SLs. Also, SL biosynthesis inhibitors should not affect seed germination, as the seeds of parasitic plants should not produce SLs. Otherwise, the seeds would germinate autonomously, and the host-dependent mechanism for germination would not occur. Thus, the induction of seed germination by TIS108 is unlikely to be due to loss of SL production. TIS108 may influence the accumulation of other signaling molecules by inhibiting other P450s. However, given the role of gibberellin and BR in promoting seed germination, inhibition of their biosynthesis by TIS108 is unlikely (Takeuchi et al. 1995, Toh et al. 2014, Bao et al. 2017). Similarly, the catabolism of ABA is unlikely to be the target, as abscinazole-E2B inhibits the germination of P. ramosa (Lechat et al. 2014). Alternatively, parasitic plants may produce unknown SL-like molecules that inhibit their germination, and suppression of their production by TIS108 results in induction of germination. Although MAX1 has not been found in parasitic plants, the P. aegyptiaca genome harbors CCD7 and CCD8 orthologs (Aly et al. 2014). This suggests that parasitic plants are capable of producing SL-like molecules. Further investigations of hormone profiles and biochemistry are required, and the above-mentioned inhibitors should be tested in other parasitic plants such as S. hermonthica to clarify whether the trait is specific to P. aegyptiaca or general to parasitic Orobanchaceae. LGS1 is an interesting target for controlling Striga infestations. The biased production of orobanchol over 5DS in lgs1 cultivars reduces their susceptibility to Striga without changing the above-ground architecture, indicating that the diverse ecological functions of SLs can be dissected by manipulating the SL preferences (Gobena et al. 2017). Inhibition of LGS1 by small molecules might impart Striga resistance without affecting the signaling or other important functions of SLs.

Lineage of SL Agonists

Small molecules targeting SL receptors enable elucidation of the mechanism of SL signal transduction in parasitic plants and AM fungi. In non-parasitic plants, SL signaling inhibitors increase the level of SLs via feedback regulation, which may affect the interactions of these plants with microbes or parasitic plants. Therefore, manipulating SL signaling using small molecules involves different opportunities and challenges compared with biosynthetic inhibitors.

The development of agonists or antagonists of SL receptors is more oriented to application purposes, e.g. suicide germination of noxious parasitic weeds. Indeed, recent findings such as hormonal roles of SLs in plant growth and stress responses have been further accelerating research, and hundreds of molecules with SL-related structures have been registered to date. With some exceptions, these SL agonists can be organized as a lineage rooted in the discovery of strigol (Cook et al. 1966, Cook et al. 1972) (Fig. 2A). The first synthetic SL was developed during testing the functions of characteristic butenolide rings of strigol (Johnson et al. 1981). A simplified strigol analog with only the CD-ring, known as GR5, induced germination of the seeds of parasitic plants, which is consistent with our knowledge about the structural diversity of the ABC-rings. Subsequent studies revealed that the D-ring of SL agonists appeared to be the core structure for activity, and the structure of the ABC-ring can be optional (Wigchert et al. 1999). The principle has been adopted in the synthetic SL agonists today. In the same work, another agonist with a three-ringed ABC-ring system resembling strigol, known as GR24, exhibited the highest potency. Although the synthesis of GR24 is much simpler than that of strigol, even simpler SL analogs were developed after the discovery of the symbiotic and hormonal functions of SLs (MacAlpine et al. 1976, Brooks et al. 1985, Mangnus et al. 1992, Fukui et al. 2011). These molecules, debranones, are simple D-ring ether derivatives with phenol groups instead of ABC-rings that exhibit activity in rice, Arabidopsis and S. hermonthica. Hereafter, debranones and other synthetic agonists lacking an enol–ether linker are termed SL mimics, as the linker is essential for activity in the native SL scaffolds.

Fig. 2

Agonists and antagonists for SL receptors. (A) The lineage of the development of SL agonists. (B) Synthetic SL mimics or an antagonist that utilize the hydrolysis mechanism of the SL receptors. (C) SL agonists or antagonists that exhibits species selectivity.

Incorporation of Additional Functionalities in SL Mimics

Biochemical characterization of SL receptors has facilitated the design of new types of SL mimics. The D14 family of SL receptors are α/β hydrolase-fold proteins with a conserved catalytic triad (serine, histidine and aspartic acid) that hydrolyze SLs into ABC-ring and D-ring products (Arite et al. 2009, Hamiaux et al. 2012). The proposed reaction mechanism involves nucleophilic attack of the carbonyl carbon in the D-ring by the serine, which opens the ring and forms a covalent intermediate structure bridging the serine and histidine (Yao et al. 2016). Subsequently, the D-ring is regenerated on the histidine residue, which can be detected by liquid chromatography–mass spectrometry (LC-MS) (de Saint Germain et al. 2016, Yao et al. 2016). Although a crystallographic study reported the intermediate structure in the active conformation of D14 in complex with MAX2, the model is now questioned as it has insufficient resolution to capture the intermediate molecule (Carlsson et al. 2018). However, mutation of the catalytic serine residue of DAD2, a homolog of D14 in petunia, to alanine disrupted its GR24-dependent association with PhMAX2 in a yeast two-hybrid assay, and the mutant gene failed to complement the dad2 loss-of-function phenotype. A similar result was observed for Arabidopsis AtD14 (de Saint Germain et al. 2016). These results suggest that the hydrolysis of GR24 is essential for activating its receptor. The majority of the 11 SL receptors in S. hermonthica known as ShHTLs also hydrolyze GR24 (Tsuchiya et al. 2015, Yao et al. 2017). Therefore, a similar mechanism involving hydrolysis of SLs is expected in ShHTL-mediated activation of SL signaling in the parasite. Although further mechanistic studies are required, it is now clear that the hydrolysis is integrated in SL signal transduction.

Two independent studies have been reported in which fluorescent turn on functions were incorporated within SL mimics. As established in many enzyme assays, colorimetric dye or fluorogenic substrates generate pigments or fluorescence after the structure is enzymatically modulated. The idea was applied to incorporate a fluorophore into SL mimics so that the molecule generates fluorescence after hydrolysis by the SL receptors (Fig. 2A, B). Such molecular probes are useful in visualizing the hydrolase activity of SL receptors. In the first report, fluorescein was selected as a backbone fluorophore because it has phenol oxygen as the electron donor required for emitting fluorescence (Tsuchiya et al. 2015). By overlaying it with the phenol of debranone, fluorescence is quenched by the electron-accepting property of the D-ring. Hydrolytic cleavage of the D-ring by an SL receptor releases free fluorescein, which upon excitation emits green fluorescence. The fluorogenic SL mimic yoshimulactone green (YLG) exhibited SL activity in Arabidopsis and S. hermonthica, and fluorescence following hydrolysis by AtD14 and ShHTL proteins in vitro. The in vitro assay was used to reveal the unique profiles of binding between SL types and ShHTL proteins, which explains how the diversity of SL receptors enables parasitic plants to distinguish SL types. Moreover, a YLG analog with an improved on/off ratio due to the presence of an additional D-ring (YLGW) revealed the dynamic nature of SL perception in planta. The initial perception at the root tip of S. hermonthica embryo was gradually propagated to the entire embryo like a wave and, after dissipation of the initial wave, a second wave was accompanied by root elongation. Although the mechanism underlying these dynamics is unknown, it is hypothesized that the pattern is generated with the feedback regulation involving ethylene, unknown mobile factors and the ligand-dependent degradation of the receptors (Tsuchiya et al. 2018). The fluorogenic SL mimic GC242, the fluorophore of which is coumarin, was used to assess the enzymatic activity of RMS3, a D14 ortholog in pea (de Saint Germain et al. 2016) (Fig. 2B). The kinetic analysis with the fluorescence intensity as the read out of hydrolysis suggested that the receptor is a single turnover enzyme, which was confirmed by the identification of a D-ring covalently attached to the catalytic histidine residue. Interestingly, a D-ring analog with no methyl group at the 4' position was rapidly hydrolyzed by RMS3 without formation of the covalent adduct and did not suppress shoot branching in pea. This supports the hypothesis that not hydrolysis itself, but a covalent adduct of the D-ring, triggers a conformational change in the receptor. Although the single-turnover hypothesis is feasible, a free D-ring is reportedly produced by hydrolysis catalyzed by SL receptors from different species (Tsuchiya et al. 2015). Additionally, covalent adduct formation in non-parasitic HTL proteins has not been reported. Therefore, further investigations involving other species are required.

Besides these fluorogenic SL mimics, an irreversible inhibitor called TFQ0022 was designed by utilizing the hydrolysis reaction to attach itself covalently to the catalytic serine residue, by which access of SLs to the pocket is limited (Xiang et al. 2017) (Fig. 2B). Moreover, an interesting combination of gibberellic acid and D-ring was recently reported (Pereira et al. 2017) (Fig. 2B). The analog with a D-ring attached at the carboxylic acid of the GA₃ scaffold exhibited activity in stimulating parasitic seed germination. Because GA₃ alone was not effective, the activity is presumably related to SL function. However, the GA-D-ring conjugate should not exert GA activity because the attachment of the D-ring to the carboxylic acid should hinder its interaction with the gibberellic acid receptor GIBBERELLIN INSENSITIVE DWARF1 (GID1). (Murase et al. 2008, Shimada et al. 2008). Therefore, the hybrid SL mimic may be hydrolyzed by SL receptors, releasing free gibberellic acid and activating GID1-mediated signaling. The way in which the hybrid molecule set gibberellic acid accumulation downstream of the SL receptor is interesting, analogous to artificial rewiring of genetic circuits in synthetic biology. Unfortunately, endogenous gibberellic acid accumulation is indeed up-regulated by SLs in the parasitic seeds and the utility of the molecule in this context is limited (Toh et al. 2012). However, the idea of small molecule rewiring of a genetic circuit can potentially be applied to synthetic molecules or hormones with carboxylic acid or alcohol groups, e.g. ABA and auxin, which are inactivated by glycosylation at carboxylic acid (Jin et al. 2013, Tanaka et al. 2014, Liu et al. 2015).

Challenges in Controlling SL Receptors Using Synthetic Molecules

Despite the structural flexibility in the ABC-part, the structures of SLs have passed through natural selections to retain the D-ring with an enol–ether linker that is essential to exhibit activities to the three different class of organisms (Xie et al. 2010). The SL receptors in non-parasitic and parasitic plants evolved from the ancestral HTL/KAI2 gene independently to perceive SLs as endogenous or ecological signals. The SL-type preferences of parasitic plants tend to differ, as they tend to respond to the SLs produced by their preferred host. Indeed, the SL receptors of S. hermonthica diverged to perceive different types of SLs (Tsuchiya et al. 2015). Across the layers of evolution, all the SL receptors have retained the ability to recognize the core structure of SLs and activate downstream signaling. How the SL–receptor interactions evolved co-operatively within a species, among species and across kingdoms are important issues in SL biology. An in silico study of 339 related sequences reported that the transition from the ancestral HTL/KAI2 to D14 was driven by interactions with partner proteins, rather than being required for SL perception (Bythell-Douglas et al. 2017). However, the hypothesis does not explain why promiscuous intermediates obtain the ability to recognize the conserved structure of the D-ring enol–ether repeatedly. These issues can be addressed by investigating the evolution of the shape of the binding pocket in cross-species functional experiments using small molecules specific for different SL receptors. Moreover, small molecules can be used to assess the levels of similarity of the binding pockets. The challenge in this context is to manipulate the structures of ligands to give selectivity; these will also benefit the molecule for controlling the environmental impact of SLs, which must be carefully considered when they are used in the field.

A study of the structure–activity relationship of debranones was the first attempt at producing such small molecule inhibitors (Fukui et al. 2011) (Fig. 2C). The first debranone analog had bromide at the 4' position of the phenol and had activity comparable with that of GR24 in rice and Arabidopsis, but its stimulation of S. hermonthica germination was reduced 10^–4-fold more than GR24. Substitution with bromide and cyan at the 2' and 5' positions increased the ability of the analog to stimulate S. hermonthica germination without inhibiting tiller growth in rice (Fukui et al. 2017) (Fig. 2C). Although the mechanism of this selectivity is unknown, the ability of simple modifications to modulate receptor selectivity will facilitate development of selective agonists based on different scaffolds. Besides the SL mimics, the chemical space of SL ligands has been explored by screening of small molecule libraries for novel scaffolds (Fig. 2C). Yeast two-hybrid screening of Arabidopsis HTL/KAI2 and MAX2 resulted in identification of azole-containing molecules, which induced germination of S. hermonthica and Arabidopsis (Toh et al. 2014). Although their interactions with ShHTL proteins have been proven, the conserved activities of these azole-containing molecules suggest that they are agonists for HTL/KAI2 in Arabidopsis and S. hermonthica. The mechanistic conservation was applied to another small screening for the inhibitors of GR24-dependent reduction of hypocotyl growth and germination in Arabidopsis, and a novel piperizine-containing molecule, soporidine, was identified as an inhibitor of HTL/KAI2 proteins in Arabidopsis and S. hermonthica (Holbrook-Smith et al. 2016). On the other hand, inhibitors of D14 proteins have recently been screened with various screenings including in silico docking, in vitro binding with differential scanning fluorometry or YLG-based binding assay (Mashita et al. 2016, Hamiaux et al. 2018, Yoshimura et al. 2018). These screenings all yielded accelerators of shoot branching, and a particular small molecule called DL1 appeared to be ineffective in inhibiting HTL protein in Arabidopsis. Interestingly, these small molecule screenings together showed no conserved structures, nor similarity to SLs, indicating that the chemical space for the ligand of the SL receptors is much broader than our current knowledge. The inherent nature of the SL receptors in accepting diverse structures of synthetic molecules may be the key to understanding the whole picture of SL biology.

Closing Remarks

SL research was initiated with the aim to solve the problem of agricultural damage caused by parasitic plants. The efforts to address biological questions of how parasitic plants recognize the presence of their hosts led natural product chemistry to the identification of strigol as a host factor for the host-dependent germination in the parasite. Following their discovery, organic chemistry provided structural insight of SLs, which eventually encouraged biologists to identify the genetic components involved in SL biosynthesis and signaling. The biological information accelerated the development of new small molecule probes that have revealed novel mechanisms. In this way, the history of SL research has been established on mutually associated efforts between chemistry and biology. The research field thus has provided an example of chemical biology research that asks questions about how structures of small molecules lead to phenotypic consequence. The way in which it asks biological questions is similar to that in genetics which asks how genotype leads to phenotypic consequences. However, there is a difference, as one is led by contacts with small molecules while the other is led by polymorphisms in the DNA sequence. Therefore, elucidation of the mechanisms underlying the effects of small molecules relies on an understanding of their interactions with proteins, and this will require ongoing and closer collaborations between chemists and biologists.

Funding

The work was supported by KAKENHI a Grant in Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology (15KT0031 and 15K07102).

Disclosures

The authors have no conflicts of interest to declare.

Abbreviations

Abbreviations
 
• AM fungi

  arbuscular-mycorrhizal fungi

 
• BR

  brassinosteroid

 
• CCD

  carotenoid dioxygenase

 
• D14

  DWARF14

 
• D53/SMAX1/SMXL

  DWARF53/SUPPRESSOR OF MAX2 1/SMAX-Like

 
• 4DO

  4-deoxyorobanchol

 
• 5DS

  5-deoxystrigol

 
• GID1

  GIBBERELLIN INSENSITIVE DWARF1

 
• HTL/KAI2

  HYPOSENSITIVE TO LIGHT/KARRIKIN INSENSITIVE2

 
• LBO

  LATERAL BRANCHING OXIDOREDUCTASE

 
• LGS1

  LOW GERMINATION STIMULANT1

 
• MAX2/D3

  MORE AXILLARY GROWTH2/DWARF3

 
• NCED

  9-cis-epoxycarotenoid dioxygenases

 
• SL

  strigolactone

 
• YLG

  yoshimulactone green