Strigolactones in Rhizosphere Communication: Multiple Molecules With Diverse Functions

Abstract

Strigolactones (SLs) are root-secreted small molecules that influence organisms living in the rhizosphere. While SLs are known as germination stimulants for root parasitic plants and as hyphal branching factors for arbuscular mycorrhizal fungi, recent studies have also identified them as chemoattractants for parasitic plants, sensors of neighboring plants and key players in shaping the microbiome community. Furthermore, the discovery of structurally diverged SLs, including so-called canonical and non-canonical SLs in various plant species, raises the question of whether the same SLs are responsible for their diverse functions ‘in planta’ and the rhizosphere or whether different molecules play different roles. Emerging evidence supports the latter, with each SL exhibiting different activities as rhizosphere signals and plant hormones. The evolution of D14/KAI2 receptors has enabled the perception of various SLs or SL-like compounds to control downstream signaling, highlighting the complex interplay between plants and their rhizosphere environment. This review summarizes the recent advances in our understanding of the diverse functions of SLs in the rhizosphere.

Introduction

Plant roots release a diverse range of molecules that affect other organisms living in the rhizosphere, the soil surrounding the roots (Bais et al. 2006). Among these molecules, strigolactones (SLs) have gained significant attention due to their crucial roles in plant growth, development and rhizosphere communication (Al-Babili and Bouwmeester 2015, Waters et al. 2017) The first identified SL, strigol, was isolated from cotton root exudates as a germination stimulant for the Orobanchaceae parasitic plant Striga lutea, hence the name ‘strigol’ (Cook et al. 1966). Subsequent studies have shown that SLs generally stimulate the germination of obligate parasites in the Orobanchaceae, including Striga spp. and broomrapes (Orobanche and Phelipanche spp.) (Wang and Bouwmeester 2018). These parasitic plants are noxious weeds that infest important crops and vegetables, such as maize, sorghum, rice, cowpea, carrot and tomato, causing billions of dollars in annual yield losses (Runo and Kuria 2018, Mutuku et al. 2020). The nature of obligate parasitism of Striga and broomrapes makes SLs ideal molecular cues to recognize nearby host roots, as SLs are rapidly degraded in soil due to chemical instability (Yoneyama et al. 2009, Akiyama et al. 2010) and are only secreted by living plants. Analyses of various plant species showed that most plants secrete SLs, but such general conservation of these molecules made people wonder why plants secrete SLs as their function does not seem to be beneficial to the plants themselves but instead awakens dangerous parasites. More than 40 years since their discovery, SLs were rediscovered as activators of hyphal branching in the mutualistic symbionts, arbuscular mycorrhizal fungi (AMF), and the long-standing question was solved (Akiyama et al. 2005). AMF colonize plant roots and deliver phosphates and nitrogen to the host plants in exchange for carbohydrates provided by the plant. Thus, the secretion of SLs into the rhizosphere is not for germinating parasitic weeds, but for their beneficial symbionts. In 2008, SLs were found to suppress shoot branching in plants by acting as endogenous plant hormones (Gomez-Roldan et al. 2008, Umehara et al. 2008) and have since become the subject of intense research due to their diverse and complex functions in plants. SLs are now known to regulate various developmental or physiological processes, such as shoot and root architecture, leaf senescence and stress responses as well as interactions with microbes and plants (Aliche et al. 2020) (Fig. 1A).

Understanding the function of endogenous SLs has led to the identification of key genetic components involved in the biosynthesis and perception of these molecules. Research into shoot-branching mutants, including Arabidopsis more axillary growth (max), rice dwarf/high-tillering dwarf (d/htd), pea ramosus and petunia decreases apical dominance (dad), has enabled the identification of primary enzymes involved in SL biosynthesis (reviewed in Mashiguchi et al. 2021). Further biochemical studies confirmed the core biosynthetic pathway in which the SL precursor carlactone (CL) is synthesized by DWARF27 (D27) and CAROTENOID CLEAVAGE DIOXYGENASE (CCD) 7 and 8 (Alder et al. 2012), Subsequently, CL is converted to carlactonoic acid (CLA) by the cytochrome P450s (CYP) in the family of CYP711A, also called as MAX1 homologs (Abe et al. 2014, Seto et al. 2014, Yoneyama et al. 2018a) (Fig. 2).

The core biosynthetic pathway is followed by further enzymatic reactions that are diversified among plant species, leading to the diversification of SLs. Over 30 SLs have been identified to date from various plant species, including canonical SLs and non-canonical SLs (Chesterfield et al. 2020) (Fig. 2). Canonical SLs have a tricyclic lactone ring (ABC-ring) connected to a methyl butenolide ring (D-ring) by an enol–ether bridge and are classified into two types: strigol-type and orobanchol-type, depending on the stereochemistry of the C-ring (Wang and Bouwmeester 2018, Mashiguchi et al. 2021). Non-canonical SLs, including CL and CLA, lack A-, B- and/or C-rings and have different structures based on a β-ionone ring linked to a D-ring. In all natural SLs identified to date, the connection to the D-ring is conserved in the 2ʹR configuration, which is essential for the biological activities of SLs (Thuring et al. 1997, Waters et al. 2017). It should be noted that the synthetic SL analog GR24 has been used extensively over the years as a substitute for natural SLs. However, GR24 can be configured as four stereoisomers that differ in stereochemistry at the 8b/3a and 2ʹ positions (Fig. 1B). The commonly used GR24 has 8b/3a stereochemistry identical to 5-deoxystrigol (5DS), but a racemate of GR24^5DS and its 2ʹS enantiomer GR24^ent−5DS, hereafter referred to as rac-GR24. Because unnatural 2ʹS configured SLs sometimes stimulate non-SL signaling pathways, biological activities assessed with synthetic SLs should be interpreted with caution (Scaffidi et al. 2014). Moreover, different plant species secrete a mixture of various SLs, with compositions varying between and sometimes even within species (Wang and Bouwmeester 2018, Aliche et al. 2020). Recent studies suggest that the non-canonical SLs primarily inhibit shoot branching, while both non-canonical and canonical SLs are involved in rhizosphere communication (Yoneyama et al. 2018b, Wakabayashi et al. 2019, Ito et al. 2022). Although the modification and stereochemistry of SLs contribute to the diversification of their biological activities, the relationship between their activities and structures is just beginning to be revealed.

SLs Stimulate the Germination of Orobanchaceae Root Parasitic Plants

Root parasitic plants have evolved an SL-based system to recognize their nearby hosts through the hijacking of the AMF hyphal branching signals (Akiyama et al. 2005), which is crucial for their survival, due to their tiny seeds with limited nutrients. However, the huge structural diversity of SLs raises the question of whether there is any structural specificity in their ability to stimulate the germination of parasitic plants. While both canonical and non-canonical SLs are active as germination stimulants for different parasite species (Yoneyama et al. 2018b), accumulating evidence indicates that SL precursors, CL and CLA, are not major germination stimulants. For example, knockout of rice MAX1-900 (referred to as OsMAX1-900), the cytochrome P450 CYP711A catalyzing carboxylation of CL, significantly reduced the production of canonical SLs, 4-deoxyorobanchol (4DO) and orobanchol, resulting in reduced germination of Striga hermonthica and Phelipanche ramosa (Ito et al. 2022). Similarly, the application of a chemical inhibitor of OsMAX1-900, 6-phenoxy-1-phenyl-2-(1H-1,2,4-triazol-1-yl) hexan-1-one (TIS108), reduced the production of canonical SLs and mitigated Striga infestation without obvious growth phenotype, suggesting that canonical SLs are effective for parasite germination in rice (Ito et al. 2022). Furthermore, knockout of CYP722C in tomatoes resulted in reduced orobanchol and increased CLA accumulation, resulting in lower germination induction to S. hermonthica, Orobanche crenata and Phelipanche aegyptiaca (Wakabayashi et al. 2019). Importantly, all of these mutants do not show increased tiller outgrowth/shoot-branching phenotype, suggesting that the SLs acting in rhizosphere communication and regulation of shoot architectures are different. The phytohormonal SLs are possibly derivatives of CLA or of methyl carlactonoate (MeCLA), which is produced by CLA methyltransferase (CLAMT) (Mashiguchi et al. 2022, Fig. 2A). For example, hydroxymethyl carlactonoate is produced by the oxidation of MeCLA by the enzyme LATERAL BRACHING OXIDOREDUCTASE (LBO), and the Arabidopsis lbo mutants showed increased shoot-branching phenotypes (Brewer et al. 2016, Mashiguchi et al. 2021, 2022). Whether MeCLA contributes to parasite germination activity remains unclear, as it showed higher germination induction activity to Orobanche minor than CL and CLA, but less than a synthetic SL, rac-GR24 (Abe et al. 2014). The maize mutant of ZmCYP706c37, lacking oxidation enzyme functional at several steps from MeCLA to zealactone and CL to zealactol, accumulates CLA and MeCLA but shows lower S. hermonthica germination induction (Li et al. 2023).

The structural diversity of SLs plays a significant role in host specificities of parasitic weeds through their germination-inducing activities (Xie et al. 2010). Root exudate surveys of various plants have revealed that no single plant species can universally induce germination of all broomrape species, suggesting the existence of host–parasite specificities between root exudate composition and broomrape germination (Fernandez-Aparicio et al. 2009). Consistently, none of the known natural SLs show universal germination stimulation activity for various parasite species (Fig. 2). For example, S. hermonthica germinates in response to strigol-type SLs, such as 5DS stronger than orobanchol (Wakabayashi et al. 2022), which has opposite stereochemistry at the C-ring. Mutation in the LOW GERMINATION STIMULANT 1 (LGS1) locus of sorghum resulted in a shift of its SL production, from 5DS to orobanchol, leading to lower germination induction of S. hermonthica without affecting the plant growth phenotype (Gobena et al. 2017). In contrast, Striga gesnerioides responds to orobanchol-type SLs, such as orobanchol and alectol (also known as orobanchyl acetate), but not strigol-type, such as strigol and 5DS (Ueno et al. 2011, Wakabayashi et al. 2022). Strigol-type SL even can inhibit the germination of S. gesnerioides (Nomura et al. 2013). The major host of S. gesnerioides, cowpea, produces orobanchol and alectol, and thus it is possible that parasitic plants evolved to recognize SL structures secreted by their preferred host plants. Structural variation in non-canonical SLs also affects the germination induction specificities. Maize cultivars with genomic deletion in the region including ZmMAX1b and ZmCLAMT1, the enzymes responsible for conversion from CL to CLA and subsequently to MeCLA, reduced the production of zealactone, a major non-canonical SL in maize with a strong activity for S. hermonthica germination induction (Li et al. 2023). The same cultivars, however, increased zealactol and zealactonoic acids, which have different C-19 modifications compared to zealactone and a lower activity for parasite germination (Li et al. 2023), suggesting that diversification of pathways after CL determines the biological activity of SLs. Heliolactone, a non-canonical SL isolated from sunflower, can highly induce germination of S. hermonthica, Orobanche crenata and O. minor but to a lesser extent Orobanche cumana and Phe. aegyptiaca (Ueno et al. 2014). It should be noted that O. cumana is a serious constraint to sunflowers. Thus, sunflower-producing SL seems to have a structure not to awaken its enemy. However, O. cumana primarily responds to non-SL sesquiterpene lactones, such as dehydrocostus lactone and costunolide, tomentosin and 8-epixanthatin, isolated from sunflower root exudates (Joel et al. 2011, Raupp and Spring 2013, Brun et al. 2018; Fig. 2C). It is tempting to speculate that such an irregular recognition of germination stimulants may result from an evolutionary arms race between sunflower and O. cumana.

Recognition of SLs by Parasitic Plants in the Orobanchaceae

Various SL structures may indicate the presence of structurally diversified receptors in parasitic plants. Indeed, parasitic plants in the Orobanchaceae show a unique evolution of SL receptors. In autotrophic plants, such as Arabidopsis and rice, SLs are recognized by a dual functional receptor/hydrolase, DWARF 14 (D14)/AtD14, which interacts with an F-box protein MAX2/D3 ubiquitin E3 ligase (Waters et al. 2017). Upon SL perception, the ubiquitin–proteasome complex degrades the D53/SMAX repressors to switch on the downstream signaling (for details on SL perception mechanisms, please see recent excellent reviews: Bürger and Chory 2020, Nelson 2021, Guercio et al. 2023). KAI2 is a paralog of D14 and was originally identified from an Arabidopsis mutant that shows insensitivity to karrikins (KARs), smoke-derived germination inducers (Waters et al. 2012). Although SLs and KARs share a butenolide structure, Arabidopsis KAI2 recognizes KARs or yet-unidentified endogenous KAI2-ligand (KL), but not natural SLs, to stimulate seed germination and repress hypocotyl elongation by forming a complex with MAX2/D3 F-box protein (Nelson 2021). Genomes of parasitic species in the Orobanchaceae possess a unique clade of KAI2, containing rapidly duplicated KAI2 genes, named KAI2d (divergent) (Conn et al. 2015, Nelson 2021). In contrast to KAI2 in other plants, Orobanchaceae KAI2d recognize exogenous SLs and their analogs, but not KARs, and contributes to the germination of Orobanchaceae parasites (Conn et al. 2015, Tsuchiya et al. 2015). While the Orobanchaceae have retained a few conserved KAI2 (KAI2c) with KAR recognition activity (Xu et al. 2018), KAI2d genes were highly duplicated in Orobanchaceae genomes; at least 8 KAI2/HYPOSENSITIVE TO LIGHT (HTL) genes are found in the S. hermonthica transcriptome and 17 were annotated in the Striga asiatica genome for the KAI2d clades (Tsuchiya et al. 2015, Yoshida et al. 2019). Each KAI2d/HTL showed different recognition activities against different SLs (Toh et al. 2015, Tsuchiya et al. 2015), suggesting that the duplication of KAI2 genes endowed a unique SL recognition system, which presumably affects parasite–host specificities to Orobanchaceae parasites (Conn et al. 2015).

With the identification of the SL receptors responsible for S. hermonthica germination, chemical approaches have become a powerful tool for elucidating its perception mechanisms and for developing highly active Striga germination inducers, which can be used for Striga control by promoting suicidal germination of the parasites without the presence of host crops. Among the S. hermonthica HTLs (ShHTLs), ShHTL7 has emerged as the most potent SL receptor for the induction of seed germination when introduced into Arabidopsis kai2 mutants (Toh et al. 2015, Tsuchiya et al. 2015, Uraguchi et al. 2018). Chemical screening and subsequent careful structure–activity association analysis successfully identified a hybrid molecule called sphynolactone-7 (Fig. 2B), which induced the germination of S. hermonthica at the femtomolar level without disturbing the shoot branching in Arabidopsis thaliana or hyphal branching in AMF (Uraguchi et al. 2018). Although the hydrolysis activity of ShHTL7 is important for high potency for seed germination induction, it is not required for the germination induction itself (Uraguchi et al. 2018). Recently, another agonist called aryloxyacetyl piperazines has been discovered with seed germination potency higher than rac-GR24 for S. hermonthica and Phe. aegyptiaca (Wang et al. 2022). The stereochemical structure of the compound affects ShHTL7-mediated hydrolysis, which correlates with its potency (Wang et al. 2022), implying that the efficiency of hydrolysis, which is influenced by structural differences in the SLs, plays a role in their germination potency.

The KAI2d clade emerges at the node of the Orobanchaceae root parasites, including facultative parasites that usually do not require SLs for their germination, implying that KAI2d may have functions in facultative parasites. Under nitrate-depleted and dark conditions, where germination without stimulants is suppressed, the seed germination of the facultative hemiparasite Phtheirospermum japonicum is promoted by rac-strigol, but not by KARs (Ogawa and Shirasu 2022). Thus, under certain environmental and/or seed conditions, SLs can act as germination stimulants on facultative parasitic plants. Of note, SL-induced germination in Arabidopsis has been observed in thermo-stressed seeds or dormant ecotypes (Toh et al. 2012). However, because racemate SLs can activate both D14 and KAI2 pathways, interpreting the germination response to SLs and KARs requires extra caution (Scaffidi et al. 2014, Waters et al. 2017). Given that the parasitic Orobanchaceae possess KAI2-derived SL receptors (KAI2d), KAR-recognizing KAI2c and function-unknown intermediate KAI2 (KAI2i) in their genome (Nelson 2021), investigating the response of the facultative parasites could shed light on their evolution as intermediates between autotrophs and obligate parasites.

SLs Induce Chemotropism of Roots in Orobanchaceae Parasitic Plants

After germination, Orobanchaceae parasites extend their roots toward the host to reach the host tissues, a process known as host tropism (Williams 1961, Yoshida and Shirasu 2009). SLs act as chemoattractants for Orobanchaceae parasitic plants, as evidenced by Pht. japonicum (Ogawa et al. 2022). SL-soaked filter papers placed on agar media induced chemotropic activity in Pht. japonicum seedlings, as their roots grew toward the filter paper. This chemotropism to SLs is also present in the obligate hemiparasite S. hermonthica, but not in A. thaliana and a non-parasitic Orobanchaceae Lindenbergia philippensis (Ogawa et al. 2022). Among the four stereoisomers of 5DS, 5DS, ent-5DS, 4DO and ent-4DO, naturally occurring isomers, 5DS and 4DO with a 2ʹR stereochemistry, had stronger chemotropism-inducing activity in Pht. japonicum than unnatural isomers, ent-5DS and ent-4DO, with a 2ʹS stereochemistry. Neither Pht. japonicum nor S. hermonthica showed chemotropism to KARs, suggesting that hemiparasites have evolved to specifically respond to SLs for tropism. Pht. japonicum has seven KAI2d homologs in its genome, which can be divided into two groups based on their similarity to KAI2d in obligate hemiparasites such as S. hermonthica (hemiparasite type) and KAI2d in holoparasites such as Orobanche spp. and Phelipanche spp. (holoparasite type) (Ogawa et al. 2022). At least two KAI2d in the hemiparasite type function as receptors for exogenous SLs, presumably leading to chemotropism. This implies that Orobanchaceae hemiparasitic plants have specifically modified KAI2d to recognize exogenous SLs for host tropism (Ogawa et al. 2022).

The host tropism mechanisms of other parasite species may be similar. Intriguingly, the holoparasite O. cumana recognizes costunolide (Fig. 2C) as a chemoattractant. Costunolide is a sunflower-derived sesquiterpene lactone that effectively induces the germination of O. cumana (Krupp et al. 2021). This may suggest that parasitic plants use host-derived signal compounds for both germination and host tropism, allowing them to detect the presence of proximate hosts. However, O. cumana did not show chemotropism to rac-GR24, which can also induce the germination of O. cumana at similar concentrations (Joel et al. 2011, Krupp et al. 2021). While the receptor for costunolide in O. cumana remains unidentified, it is possible that KAI2 homologs recognize sesquiterpene lactones (Rahimi and Bouwmeester 2021). Isothiocyanates, non-SL germination stimulants specific to the holoparasite Phe. ramosa, are also proposed to be perceived by KAI2 (de Saint Germain et al. 2021). The evolution of KAI2 genes may confer various ligand recognition abilities, other than KARs/KL or SLs. A recent report showed that only three amino acid mutations in Arabidopsis KAI2 change its ligand specificity from KARs only to both SLs and KARs (Arellano-saab et al. 2021). The structural analysis of KAI2 families from various parasitic species could reveal the evolution of the receptors for germination stimulants and chemoattractants.

Interestingly, chemotropism to SLs in Pht. japonicum was suppressed in ammonium-replete conditions, irrespective of other nutrient components (Ogawa et al. 2022). However, the recognition of exogenous SLs, measured by the fluorescent probe Yoshimulactone Green for monitoring SL hydrolysis upon recognition (Tsuchiya et al. 2015), was still observed in such conditions, indicating that ammonium ions repressed the downstream signaling of KAI2d. Consistently, the expression of PjSMAX1.2, a homolog of a repressor of KAI2 signaling whose overexpression in Pht. japonicum roots resulted in attenuated chemotropism to SLs, was enhanced in ammonium ion–rich conditions. In addition, SL-inducible local auxin accumulation in the root elongation zone, which may lead to asymmetrical root growth and chemotropism, was suppressed in nutrient-rich conditions. These data suggest that SL-based chemotropism is tightly linked to environmental ammonium concentrations (Ogawa et al. 2022).

SLs as Signals for Detecting Neighboring Plants

Even autotrophic plants exploit SLs as signals to recognize their neighbors. Wheeldon et al. (2022) and Yoneyama et al. (2022) reported this function of SLs using pea and rice, respectively. When plants are densely planted in one pot, they recognize the neighboring plant-secreted SLs and compromise shoot branching and shoot growth (Wheeldon et al. 2022, Yoneyama et al. 2022). Densely planted plants synthesize and exude less SLs per plant, but interestingly, consistent per pot, suggesting that plants adjust SL secretion depending on environmental SLs to keep SL homeostasis (Wheeldon et al. 2022, Yoneyama et al. 2022). SL biosynthesis in neighboring plants and SL sensing in focal plants are required for the density-dependent adjustment of SL production and shoot growth (Wheeldon et al. 2022, Yoneyama et al. 2022). It is intriguing that despite reduced SL production, the plants showed less branching phenotype, a typical response to high SLs. However, the measured SLs in these studies were canonical SLs, which are most likely not responsible for shoot-branching suppression, but rather act as rhizosphere signals as suggested in recent reports (Ito et al. 2022). A bryophyte Physcomitrium patens, which does not form symbiotic interaction with AMF, produces SL-like compounds via CCD7- and CCD8-dependent pathways. Although the exact molecule has not been identified, Phy. patens may produce CL-derived non-canonical SLs because it lacks MAX1 homolog in the genome (Decker et al. 2017). Endogenous SL-like compounds in Phy. patens are secreted into the medium and repress caulonema filament growth of neighboring colonies (Proust et al. 2011, Decker et al. 2017). This suggests that the function of SLs as detectors of plant density may be established in an early lineage of plant evolution. Alternatively, mosses and vascular plants have independently acquired the SL function for neighbor detection, because no evidence of SL perception is reported in liverworts and hornworts, located in the basal plant lineage. Whether species-specific SL composition affects kin recognition is awaiting future studies.

SLs Act as Branching Factors for AMF

AMF are a group of obligate symbionts that inhabit plant roots in the rhizosphere. These fungi enter the root cortex and their hyphae branch to form arbuscules, surrounded by an expanded plant plasma membrane known as a periarbuscular membrane, from which plants exchange fixed carbons for nutrients such as phosphate, nitrate and ammonium (Parniske 2008, Luginbuehl and Oldroyd 2017, Wang et al. 2020). In addition to improving nutrient acquisition, arbuscular mycorrhizal (AM) symbiosis enhances plants’ responses to biotic and abiotic stress (Leigh et al. 2009, Campos-Soriano et al. 2010, Ma et al. 2022). AMF have been shown to increase their hyphal branching near host roots to increase the probability of the fungi encountering the host roots, a crucial first step for obligate symbiosis. One of the canonical SLs, 5DS, was isolated from the legume Lotus japonicus root exudates as a branching factor for AMF (Akiyama et al. 2005). Since then, SLs have been known to play a major role as host recognition signals in plants’ communication with AMF. SLs stimulate the metabolic activity of AMF, preparing them for the successive establishment of symbiosis (Besserer et al. 2006, 2008, Tsuzuki et al. 2016). Plants defective in genes regulating SL biosynthesis, including the biosynthetic enzymes (CCD7 and CCD8) and GRAS (GAI-RGA-SCR)-type transcriptional regulators NODULATION SIGNALING PATHWAY 1 (NSP1) and NSP2 of D27, as well as ABC transporter PLEIOTROPIC DRUG RESISTANCE 1 (PDR1) that exudes SL into the soil, show reduced colonization levels (Gomez-Roldan et al. 2008, Liu et al. 2011, Kretzschmar et al. 2012, Banasiak et al. 2020).

Both canonical and non-canonical SLs have been reported as hyphal branching factors for AMF, though the latter generally have weaker activities (Akiyama et al. 2005, 2010, Mori et al. 2016, Xie et al. 2019) (Fig. 2). Akiyama et al. (2010) previously found that intact ABC-ring and C-D part are necessary for high hyphal branching activity in canonical SLs. Later research showed that non-canonical SLs that lack BC-rings but retain A-ring and enol–ether bridge of C-D ring also have hyphal branching activity, with oxidation of C-19 being important for higher activity (Mori et al. 2016). Although CL only had weak activity, CLA, whose C-19 is oxidized, showed activity similar to that of natural SLs, strigol or sorgomol, or synthetic GR24^5DS (Mori et al. 2016). Methyl-esterified MeCLA has moderate activity 10 times less than that of CLA but higher than that of CL. Natural non-canonical SLs, lotuslactone, avenaol, heliolactone and zealactone show equal or lesser activity than that of MeCLA (Xie et al. 2019). Because these non-canonical SLs are active germination stimulants for various parasitic species, their abilities to induce AM hyphal branching and germination can be dissected. A sorghum line with a mutation in the LGS1 locus, showing a shift of SL production from 5DS to orobanchol, was colonized by AMF at a similar level to the control line (Gobena et al. 2017). Rice MAX1-900 KO lines, lacking the production of canonical SLs, showed delayed but otherwise normal AM colonization (Ito et al. 2022), suggesting that both non-canonical and canonical SLs contribute to AM colonization in rice. Notably, zealactol and zealactonoic acid, which are exudated from maize roots and do not awaken Striga seeds, have C-19 oxidation, a possible mark for AM hyphal branching activity of non-canonical SLs (Mori et al. 2016, Li et al. 2023). This may imply that plants possess parallel biosynthetic pathways for SLs with different biological activities.

Involvement of Plant KAI2 Receptors in Arbuscular Mycorrhizal Symbiosis

Regulation of AM colonization involves the D3 and D14L (KAI2) pathways, because impaired colonization was observed in the d3 and d14l (kai2) mutants in rice and in the kai2a mutant in Petunia (Yoshida et al. 2012, Gutjahr et al. 2015, Liu et al. 2019). Upon activation of the D3 and D14L signaling pathways, SMAX1, a negative regulator of SL biosynthesis and AM symbiosis is degraded, similar to the proteolysis of D53. Choi et al. (2020) proposed an endogenous KL that represses SMAX1 through D14L/D3 receptor complex. The D14L-dependent KL signaling may precede the evolution of SL signaling as KAI2 is present in all land plants and algae (Kyozuka et al. 2022). While KAI2 and D14 have distinct ligand-binding specificity, the sub-functionalization of KAI2 enables the recognition of various substrates including SLs and their derivatives (Tsuchiya et al. 2015, Xu et al. 2018, Carbonnel et al. 2020, de Saint Germain et al. 2021, Lopez-obando et al. 2021). Mutants for Petunia kai2a exhibit the phenotype of reduced hypodermal passage cells (HPCs), the non-suberized passage cells that facilitate nutrient uptake as well as SL exudation (Liu et al. 2019, Banasiak et al. 2020). The treatment of GR24^5DS but neither the enantiomer GR24^ent−5DS nor KARs increase the number of HPCs in wild type, but not in kai2a, suggesting that SLs function as regulators of cell-type formation in a KAI2-dependent manner, and that they possibly facilitate AMF penetration (Sharda and Koide 2008, Liu et al. 2019). By contrast, KAI2 and MAX2 are not involved in AM symbiosis in the liverwort Marchantia paleacea. There was no significant difference in AM colonization between the wild type and kai2 or max2 mutants of M. paleacea (Kodama et al. 2022). The regulation of AM symbiosis is therefore presumably not an ancient function of the KL pathway.

Origin of SLs and SL Perception Systems in Embryophytes

The synthesis of SLs is an evolutionarily ancient trait in embryophytes (Walker et al. 2019). While the origin of SLs in Charales is disputed due to conflicting evidence regarding their symbiotic role and detection (Delaux et al. 2012, Kyozuka et al. 2022), the discovery of bryosymbiol (BSB), a SL produced by M. paleacea, led to the proposal that SLs originated as signaling molecules to induce AM symbiosis in embryophytes (Kodama et al. 2022, Fig. 2D). BSB production is exclusive to AM symbiosis–forming species and is correlated with phosphate-dependent colonization of AMF. However, the biosynthesis of BSB was uncoupled from the endogenous SL signaling pathway in M. paleacea due to the lack of a BSB-perceivable receptor, presumably an ortholog of D14 in flowering plants (Kodama et al. 2022).

The ancient paralog of D14, KAI2, is not understood to react with endogenous SL. Two KAI2s (KAI2A and KAI2B) in M. paleacea had either weak or no interaction with rac-GR24 and rac-2ʹ-epi-GR24 (an equimolar mixture of GR24^4DO and GR24^ent−4DO), which is revealed by in vitro differential scanning fluorimetry (DSF) analysis. Moreover, ccd8 and max1 mutants of M. paleacea showed unaltered growth and morphology, suggesting a minor effect of endogenous BSB on plant development, if any (Kodama et al. 2022). In contrast, additional rac-GR24 caused elongated rhizoids of Marchantia spp. (Delaux et al. 2012). High concentration of rac-GR24, but not KARs, severely suppressed thalli growth of Marchantia polymorpha, although M. polymorpha was insensitive to exogenous (−)-2ʹ-epi-GR24 (GR24^4DO) and (+)-GR24 (GR24^5DS) (Mizuno et al. 2021, Kodama et al. 2022). A weak interaction of M. polymorpha KAI2 protein with GR24^ent−5DS measured by the DSF analysis suggests that the effects of rac-GR24 are likely through unnatural GR24^ent−5DS; however, the endogenous pathways reacting to rac-GR24 remain obscure, because Mpkai2akai2b double mutants only partially restored the phenotype. Phy. patens, a bryophyte that produces and perceives SL-like molecules that act as rhizosphere signals and regulate protonema branching, possesses 13 KAI2 homologs in the genome and two (KAI2L-G and KAI2L-J) of them are putative SL receptors (Lopez-obando et al. 2021). Although Phy. patens may have undergone convergent evolution, duplication and neofunctionalization of KAI2 genes, similar to root parasitic plants, it renders debatable prospect that SLs originated as mere rhizosphere signals. To further elucidate the existence of other unknown KAI2-dependent SL signaling pathway(s) in bryophytes, experiments involving different Marchantia species or stereochemistry of SLs are needed.

Role of SLs in Root Nodule Symbiosis

SLs also participate in symbiotic relationships between legumes and nitrogen-fixing bacteria, in addition to their role in fungi. In this process, legumes form nodules on their root, hosting nitrogen-fixing bacteria to exchange fixed nitrogen and carbon sources (Oldroyd 2013). The amount of SLs present in plants is finely tuned to regulate nodulation. For instance, the application of 0.1 μM rac-GR24 has been shown to increase the number of nodules in Medicago sativa and Medicago truncatula, while higher concentrations of rac-GR24 (0.2–5 μM) have been found to repress nodulation (Soto et al. 2010, De Cuyper et al. 2015). Moreover, impaired SL biosynthetic genes, including CCD7 or CCD8, in soybean, pea and L. japonicus resulted in the formation of fewer nodules (Foo and Davies 2011, Foo et al. 2013, Liu et al. 2013, Haq et al. 2017). Conversely, mutation of the SL signaling transducer MAX2 caused the opposite phenotype between soybean and pea, perhaps attributed to differences in the formation of indeterminate (pea) and determinate nodules (soybean) (Foo et al. 2013, Ahmad et al. 2020). Notably, in addition to affecting the number of nodules, ccd8 mutants of pea have also been shown to form fewer infection threads, highlighting the role of SLs during the stage of infection thread formation in root nodule symbiosis (Mcadam et al. 2017).

SLs Influence Rhizosphere Microbiome

SLs also participate in the interactions of plant roots with various microbes including both pathogenic and beneficial species underground. In particular, SLs have been found to play a positive role in plant defense against pathogenic bacteria, fungi and nematodes (López-Ráez et al. 2017, Aliche et al. 2020). Thus, as signaling molecules for both pathogenic and beneficial microbes, SLs are likely to play a broader role in shaping the overall rhizosphere microbiome. Advances in sequencing technology now enable us to capture the entire plant microbiome without culturing microbes, allowing us to shift from a single interaction between a plant and one microbe to exploring the inter-connections between plants and microbial communities (Bulgarelli et al. 2012, Bai et al. 2015, Edwards et al. 2015).

To better understand the functions of SLs in the rhizosphere microbial community, the microbiome profile has been investigated using genetic modification of the SL pathway in Arabidopsis, soybean and rice (Carvalhais et al. 2019, Nasir et al. 2019, Liu et al. 2020). Microbiome profiling in the rhizosphere of an Arabidopsis max4 mutant, which impairs in the SL biosynthetic gene CCD8, suggested that SLs affect the rhizosphere fungal community but not the bacterial community (Carvalhais et al. 2019). In this case, SLs might abolish some pathogenic fungi, such as the genus Fusarium, while attracting other pathogenic fungi, such as Epicoccum nigrum and the genus Mycosphaerella (Carvalhais et al. 2019). By contrast, a similar study using transgenic soybean lines overexpressing genes related to SL biosynthesis (MAX1d), perception (D14) and downstream signaling (MAX2a) suggested that SLs greatly affected the rhizosphere bacterial community, but not the fungal community, including the symbiotic arbuscular fungal family Glomeraceae in soybean (Liu et al. 2020). However, these overexpression lines led to an increase in the abundance of family Rhizobiaceae in the soybean rhizosphere, suggesting that SLs may selectively recruit specific microbes to their rhizosphere prior to nodule formation (Liu et al. 2020). Microbiome profiling using rice mutants impaired in SL biosynthesis (d17/ccd7) and perception (d14) suggested that SLs affect both the bacterial and fungal rhizosphere communities in rice (Nasir et al. 2019). The abundance of the beneficial bacterial family Nitrosomonadaceae and the genus Rhodanobacter was found to be depleted in both rice mutants (Nasir et al. 2019). Therefore, the assembly of the rhizosphere microbial community mediated by SLs appears to be highly host-species-specific, probably because each plant species secrete a unique blend of SLs with different amounts in their rhizosphere (Yoneyama et al. 2009, Xie et al. 2010).

Recently, Kim et al. (2022) conducted a comprehensive analysis of the root-associated microbiome using 16 rice genotypes, including 11 varieties with varying levels of SL production, including high (IAC 165, IAC 1246, Kinko and Gangweondo), intermediate (Dullo, Binagimbing and Sonkanoir) and low (TN1, Bhasmanik and Shuang-Chiang) SL producers, as well as five SL biosynthesis and perception mutant lines [d10 (CCD8), d17 (CCD7), d3, d14 and d27] (Kim et al. 2022). By integrating microbial community data with SL production levels, their regression model revealed that differences in the rhizosphere bacterial and fungal microbiome community profiles were dependent on SL levels in plant roots. Interestingly, certain microbes were found to be associated with specific SLs. For example, orobanchol was positively associated with the bacterial genera Burkholderia–Caballeronia–Paraburkholderia and Granulicella in the rhizosphere, and an AM fungal genus, Archaulospora, in roots. This correlation between AMF colonization and orobanchol levels is consistent with the fact that orobanchol is one of the most potent inducers of hyphal branching in AMF (Akiyama et al. 2010). In addition, orobanchol is negatively associated with the fungal genera Mortierella, Solicoccozyma, Saitozyma and Discosia, while 4DO is negatively associated with the fungal genus Clonostachys. Overall, these findings suggest that the production of different SLs has evolved to recruit specific microbes in the rhizosphere.

SLs are known to function both inside and outside the plant body at very low concentrations in the nano/picomolar range (Xie et al. 2015, Mashiguchi et al. 2021), suggesting their potential role as rhizosphere signals for cell–cell communications within the plant holobiont. This assemblage of plants and their associated microbiome can benefit from a sustainable approach that focuses on controlling the functions of SLs, instead of relying on chemical fertilizers, pesticides and frequent tillage. With recent technological advances in the multi-omics approach on the field scale (Ichihashi et al. 2020, Fujiwara et al. 2022), a better understanding of the precise functions of SLs as rhizosphere signals under the natural environment will be achieved. This approach has the potential to enable plants to enhance their rhizosphere microbiome and utilize the soil’s potential more effectively.

Conclusion

A comprehensive analysis of SL structures and their biosynthesis and signaling pathways have begun to reveal their multifaced roles ‘in planta’ and in the rhizosphere. Recent genetic evidence indicates that major germination stimulants for parasitic weeds are not responsible for the suppression of shoot branching. Instead, CLA-derived yet unknown products are likely responsible for this effect. Furthermore, some germination stimulants do not appear to activate AM symbiosis, although it should be noted that most of AM experiments have used only one or few AM species and the diversity of SL recognition by AMF remains unknown. This raises an old question—why do plants produce SLs that stimulate parasitic plant germination? One possibility is that SLs may have evolved to detect neighboring plants, similar to quorum sensing effects in mosses, rather than to attract parasitic plants (Proust et al. 2011). Alternatively, SLs may have evolved to resist pathogens, as defects in the SL biosynthesis pathway in moss showed susceptibility to pathogenic fungi (Decker et al. 2017). The diversification of SLs may have resulted from an evolutionary arms race between plants and pathogens, including parasitic weeds, to create an effective SL that attracts beneficial microbes and avoids pathogenic plants. Indeed, some SLs act as inhibitors against specific parasitic species (Nomura et al. 2013, Nelson 2021). The duplication of KAI2 receptors in parasitic plants is in line with this; parasitic plants may have evolved KAI2 receptors to recognize their best partners and to avoid inhibitory SLs. The diversification of SLs may have led to the exudation of SL mixtures from a single plant, but synergistic or antagonistic effects of mixed SLs remain to be elucidated. The further analysis of SLs in various plants and their effects on various organisms will reveal how SLs contribute to shaping the rhizosphere community.

Data Availability

No new datasets were generated or analyzed in this study.

Funding

Ministry of Education, Culture, Sports, Science and Technology KAKENHI grants (JP20H05909 and JP21H02506 to K.S. and S.Y., 17H06172 and 22H00364 to K.S. and JP19K22432 to S.Y.); JST PRESTO (JPMJPR194D to S.Y.); the Japan Society for the Promotion of Science (JSPS) KAKENHI Grants-in-Aid for JSPS Fellows (22KJ3127 to S.O.); Cabinet Office, Government of Japan, Moonshot Research and Development Program for Agriculture, Forestry and Fisheries (funding agency: Bio-oriented Technology Research Advancement Institution, JPJ009237 to Y.I.); MEXT scholarship to Y.J.K.

Author Contributions

All the authors wrote the manuscript and read and approved the final version of the manuscript.

Disclosures

The authors have no conflicts of interest to declare.

References

Author notes