The D3 F-box protein is a key component in host strigolactone responses essential for arbuscular mycorrhizal symbiosis


Author for correspondence:

Ken Shirasu

Tel: +81 45 503 9574



  • Arbuscular mycorrhiza (AM) represents an ancient endosymbiosis between plant roots and Glomeromycota fungi. Strigolactones (SLs), plant-derived terpenoid lactones, activate hyphal branching of AM fungi before physical contact. Lack of SL biosynthesis results in lower colonization of AM fungi.
  • The F-box protein, DWARF3 (D3), and the hydrolase family protein DWARF14 (D14) are crucial for SL responses in rice. Here we conducted AM fungal colonization assays with the SL-insensitive d3 and d14 mutants.
  • The d3 mutant exhibited strong defects in AM fungal colonization, whereas the d14 mutant showed higher AM fungal colonization. As D14 has a homologous protein, D14-LIKE, we generated D14-LIKE knockdown lines by RNA interference in the wildtype and d14 background. D14 and D14-LIKE double knockdown lines exhibited similar colonization rates as those of the d14-1 mutant.
  • D3 is crucial for establishing AM symbiosis in rice, whereas D14 and D14-LIKE are not. Our results suggest distinct roles for these SL-related components in AM symbiosis.


The arbuscular mycorrhiza (AM) represents an ancient, highly evolved endosymbiotic association between fungi in the phylum Glomeromycota and the roots of c. 80% of plant species (Harrison, 2005; Parniske, 2008). The AM fungi are obligate biotrophs that are not able to complete their life cycle without symbiotic partners (Requena et al., 2007). The AM fungi produce lipochito-oligosaccharides as a signal (Maillet et al., 2011) and are recognized by potential host plants in which a characteristic calcium spiking occurs before direct fungal contact (Kosuta et al., 2008). Fungal hyphae then attach to the plant root surface and form a hyphopodium (also referred to as an appressorium). Several hours after the formation of a hyphopodium, an intracellular cytoplasmic assembly called the prepenetration apparatus appears and directs intracellular hyphal growth (Genre et al., 2005, 2008). The fungal hyphae extend through the outer cell layers and enter cortex cells where the fungi develop highly branched membrane-rich arbuscules. Many glomeromycotan fungi also establish vesicles that are lipid-filled storage structures.

Plant genes required for AM symbiosis have been identified by both forward- and reverse-genetic approaches (Parniske, 2008; Oldroyd et al., 2009). Many of the AM symbiosis-impaired mutants identified from legumes are also defective in root-nodule symbiosis and are therefore designated as common symbiosis (sym) mutants (Parniske, 2008). The AM phenotype of common sym mutants is characterized by early blockage of fungal penetration in outer cell layers. These common SYM genes were originally identified in legumes, and orthologs are functionally conserved in nonnodulating but mycorrhiza-forming plants such as rice (Banba et al., 2008; Gutjahr et al., 2008). Rice mutants or transgenic plants that lack functional common SYM orthologs exhibit an impaired AM phenotype at early infection stages (Banba et al., 2008; Gutjahr et al., 2008). Knockdown lines of a Vapyrin-encoding gene in Medicago truncatula or the pam1 mutant in petunia exhibit defects in intracellular AM hyphal progression and arbuscule morphogenesis (Reddy et al., 2007; Feddermann et al., 2010; Pumplin et al., 2010). Apart from the common sym mutants, several genes important for AM symbiosis have been identified. The half-size ABC transporters STR1 and STR2 are also crucial for arbuscule development in M. truncatula and rice (Zhang et al., 2010; Gutjahr et al., 2012). In addition, a phosphate transporter (Javot et al., 2007) and apoplastic subtilases (Takeda et al., 2009) are essential for arbuscule formation.

Strigolactones (SLs), a group of terpenoid lactones, are key components involved in the initiation stages of the plant–AM fungi interaction. Host roots secrete SLs, which then stimulate hyphal growth and branching through activation of fungal mitochondrial metabolism (Akiyama et al., 2005; Besserer et al., 2006, 2008). SLs were originally identified as stimulants of parasitic plant germination (Cook et al., 1966) and later demonstrated to be plant hormones inhibiting outgrowth of axillary buds or tillers (Gomez-Roldan et al., 2008; Umehara et al., 2008). Recently, SLs were also shown to control leaf senescence and root architecture (Yan et al., 2007; Hu et al., 2010; Koltai, 2011).

A number of mutants in SL biosynthesis have been identified (reviewed by Beveridge & Kyozuka, 2010; Xie et al., 2010). In rice, increased-tiller-outgrowth mutants dwarf10 (d10), dwarf17 (d17 ), and dwarf 27 (d27 ) have mutations in SL biosynthesis enzymes. D27 encodes β-carotene isomerase, which converts all-trans-β carotene into 9-cis-β-carotene. Subsequently, carotenoid-cleaving dioxygenase 7 (CCD7), encoded by D17, cleaves 9-cis-β-carotene into the 9-cis-configured aldehyde. D10 encodes carotenoid-cleaving dioxygenase 8 (CCD8), which produces carlactone, a compound structurally similar to SL (Alder et al., 2012). Consistently, mutants of these genes contain undetectable concentrations of SLs, and the mutant phenotype disappears upon exogenous SL application (Gomez-Roldan et al., 2008; Umehara et al., 2008; Lin et al., 2009). Mutants or transgenic plants having reduced CCD7 or CCD8 expression exhibit a decreased AM colonization phenotype (Gomez-Roldan et al., 2008; Koltai et al., 2010; Vogel et al., 2010; Gutjahr et al., 2012). The NSP1 and NSP2 GRAS-type transcription factors regulate D27expression, and the nsp1nsp2 double mutant results in lower AM colonization (Liu et al., 2011). In addition, the Petunia hybrida ABC transporter PDR1 functions as an SL exporter, and pdr1 mutants are less colonized with AM fungi (Kretzschmar et al., 2012). Importantly, although these mutants show less AM colonization, no abnormal infection processes are observed. Therefore, the AM phenotypes in these mutants are likely the result of reduced fungal activities.

Mutants insensitive to exogenously applied SLs have been identified from various plant species. The causal genes of SL-insensitive mutants, rice dwarf3 (d3), Arabidopsis max2, and pea rms4, encode an F-box protein that is probably a component of the Skp1-Cul1/Cdc53-F-box (SCF) E3 ubiquitin ligase complex for facilitating the ubiquitination of a target protein(s) for degradation by the 26S proteasome (Stirnberg et al., 2002, 2007; Turnbull et al., 2002; Ishikawa et al., 2005; Petroski & Deshaies, 2005). Rice dwarf14 (d14), another SL-insensitive mutant, has defects in a gene for an α/β-hydrolase family protein (Arite et al., 2009). These SL-insensitive mutants have phenotypes similar to SL-deficient mutants, such as increased tiller outgrowth, delayed leaf senescence, and mesocotyl elongation (Yan et al., 2007; Gomez-Roldan et al., 2008; Umehara et al., 2008; Hu et al., 2010), but these phenotypes are not reversed by exogenously applied SLs. In addition to the SL-related phenotype, max2 promotes photomorphogenesis and less seed germination (Shen et al., 2007, 2012; Tsuchiya et al., 2010; Nelson et al., 2011). Interestingly, MAX2 is required for responses to karrikins, which are butenolides derived from wildfire smoke (Flematti et al., 2004; Nelson et al., 2011, 2012). Karrikins stimulate seed germination of many angiosperms and enhance the light response in Arabidopsis (Nelson et al., 2009). Another protein required for SL response, AtD14, does not have a major role in karrikin response in Arabidopsis (Waters et al., 2012). However, the Arabidopsis kai2 mutants that lack the functional D14 orthologous protein, D14-LIKE, are insensitive to karrikins but only weakly responsive to SLs (Waters et al., 2012). Therefore, D14 and D14-LIKE have partially overlapping but distinct functions for SL and karrikin responses (Waters et al., 2012). Notably, karrikin is not an endogenous compound, and therefore these proteins may regulate the signaling pathway(s) of as yet undiscovered plant compounds. In rice, D3 mutants exhibit a more severe mesocotyl elongation phenotype compared with SL-biosynthesis mutants (Hu et al., 2010), suggesting that D3 may have a function(s) unrelated to SL response.

Here, we report a novel phenotype of rice d3 mutants showing strong defects in AM symbiosis. Further analysis of D14 and D14-LIKE double knockdown transgenic lines confirmed that the AM phenotype was specific to d3 and that D14 and D14-LIKE were unlikely to be involved in D3-dependent regulation of AM symbiosis. Our results suggest a novel function for D3 F-BOX protein in AM symbiosis.

Materials and Methods

Plant materials

Rice (Oryza sativa L. ssp. japonica) mutants used in this study were reported previously: d3-1 (Ishikawa et al., 2005), d10-1 (Arite et al., 2007), d14-1 (Arite et al., 2009), and d17-1 (Umehara et al., 2008) are in the cv Shiokari background. The deletion mutants d3-2 and d3-3 in the cv Nipponbare background were identified from Tos17-inserted mutant lines for high tillering phenotype (Miyao et al., 2007) and the mutation positions were determined by sequencing. The seeds were provided by the National Institute of Agrobiological Sciences, Japan.

AM fungal strains and cocultivation conditions

Spores of Glomus intraradices and Gigaspora margarita were obtained from Premier Tech (Quebec, Canada) and Central Glass Co. Ltd (Tokyo, Japan), respectively. Rice seeds were sterilized with 10% (v/v) bleach (containing c. 6% sodium hypochlorite; Kao Corporation, Tokyo, Japan) for 15 min followed by extensive washing with water; seeds were germinated on moist filter paper at 26°C for 1 wk. Rice seedlings were transferred to 50 ml plastic tubes filled with vermiculite : clay soil or sand (1 : 1) containing c. 1000 G. intraradices spores. Plants and fungi were cocultivated in a glasshouse with a 12 h photoperiod at 28 : 20°C day : night temperatures. For G. margarita inoculation, rice seeds were sterilized with 70% ethanol for 3 min, then 1% sodium hypochlorite for 10 min, and germinated on 1% agar plates for 3 d in the dark. Rice seedlings were transferred to 20 × 250 mm glass tubes filled with sterile vermiculite containing 100 spores of G. margarita and grown in a 16 h photoperiod light at 25°C. Plants were given half-strength Hoagland solution containing 100 μM phosphate twice a wk. When indicated, 10 nM (+)-strigol for G. intraradices or 100 nM (±)-2′-epi-5-deoxystrigol for G. margarita was applied together with nutrient solution.

Hyphal branching of G. margarita near rice roots was observed using the membrane-sandwich method (Gutjahr et al., 2009). Two spores were sandwiched with nitrocellulose membranes, and 10-d-old rice roots were placed between spore-containing membranes and a filter paper. The samples were placed in sterile vermiculite for 7 d, and hyphal branching was observed using a microscope.

Quantification of AM staining and colonization

Roots with AM fungi were stained with trypan blue as previously described (Banba et al., 2008) and observed using a light microscope. For G. intraradices interactions, colonization percentages of intraradical and extraradical hyphae, arbuscules, and vesicles were evaluated using the line-intersect method (Giovannetti & Mosse, 1980) from at least 100 intersecting points per plant. Total colonization was counted as a percentage of the intersecting points where hyphae, arbuscules, or vesicles were observed. For the G. margarita or G. intraradices interactions shown in Fig. 5, arbuscule colonization or total colonization rates, including both hyphal and arbuscule colonization, were calculated using the magnified intersections method (McGonigle et al., 1990) from at least 200 intersection points. To quantify hyphopodia formation, the AM-colonized roots were stained, mounted on a slide, and observed using light microscopy. The slides with root samples were scanned, and the root lengths were calculated using ImageJ software (

Fluorescence microscopy

Alexa Fluor 488-conjugated wheat germ agglutinin (WGA) from Life Technologies (Carlsbad, CA, USA) was used to label fungal cell walls fluorescently (Takeda et al., 2009). Rice roots were heated at 90°C in 10% KOH for 15 min and then washed three times with excess phosphate-buffered saline. The roots were soaked in 3 μM WGA Alexa Fluor 488 and 10 μg ml−1 propidium iodide (Sigma-Aldrich, St Louis, MO, USA) for 15 min under vacuum and then washed three times with phosphate-buffered saline. The stained roots were embedded in 5% (w/v) agarose and sectioned using a Linear Slicer PRO-7 (Dosaka EM , Kyoto, Japan) to 200–300 μm thickness. The specimens were observed using a confocal laser scanning microscope system (LSM510, META; Carl Zeiss, Overkochen, Germany) with 488 nm excitation and a 505–530 nm bandpass filter for Alexa Fluor 488, and 543 nm excitation and a 560 nm longpass filter for propidium iodide. Acquired images were analyzed using Zeiss LSM Image Browser software.

Plasmid construction and transgenic rice production

A partial cDNA of D14-LIKE containing a portion of the open reading frame and 3′-untranslated region was amplified using the primer set 5′-CACCACATAGTCATCCCTGT-3′ and 5′-CATAAAACAGAGTTGCAGCTCG-3′. The PCR fragment was cloned into vector pENTR/D-TOPO (Life Technologies). The fragment was transferred by recombination reaction into vector pANDA (Miki & Shimamoto, 2004; Miki et al., 2005) using LR clonase (Life Technologies) according to the manufacturer's instructions. The recombinant plasmid was transformed into the wildtype (WT) rice of cv Nipponbare and the d14-1 mutant (cv Shiokari). Homozygous T2 generations were used for subsequent experiments.

Measurement of plant height and tiller number

Rice plants were grown in a growth incubator under a 16 : 8 h, light : dark, 28 : 24°C regime for 2 wk and then transferred to a glasshouse with a 13 : 11 h, light : dark, 28 : 24°C regime. Plant height and tiller number were measured when the ninth leaves were fully expanded.

Quantitative reverse transcription-PCR

Reverse transcription polymerase chain reaction (RT-PCR) primers for AM-inducible markers in rice have been described by Gutjahr et al. (2008). CYCLOPHILIN2 (CYC2) primers were used as internal controls. Total RNA was extracted from rice roots using an RNeasy kit (Qiagen, Hilden, Germany) with on-column DNase I digestion, and 500 ng of total RNA was used for reverse transcription. Quantitative RT-PCR (qRT-PCR) was performed using Superscript III transcriptase (Life Technologies) and THUNDERBIRD SYBR qPCR Mix (Toyobo, Osaka, Japan) according to the manufacturer's instructions using the Mx3000p (Agilent Technologies, Santa Clara, CA, USA) fluorescence-detection thermocycler. The expression levels of each gene were calculated with a standard curve method, and normalized by CYC2 quantity. To quantify expression of D14-LIKE, sterilized rice seeds were germinated on a filter paper for 1 d in the dark and grown on solidified hydroponic culture media (Umehara et al., 2008) for 7 d in an incubator (16 : 8 h, light: dark at 28 : 24°C). Whole root tissues were harvested for total RNA extraction. For quantification of D10 expression, the basal portions of shoots, including the shoot apical meristem, were harvested from rice plants grown in an incubator (16 : 8 h, light: dark at 28 : 24°C) for 2 wk. Total RNA was extracted using a Plant RNA Isolation mini kit (Agilent Technologies). After DNase I treatment, first-strand cDNA was synthesized using SuperScript III reverse transcriptase. The primer sets used to amplify the transcripts and internal controls were as follows: for D14-LIKE, 5′-TGGAGGATTTGAGCAGGAG-3′ and 5′-CACCAGGCCTTGTAGTTTGAC-3′; for D10, 5′-CGTGGCGATATCGATGGT-3′ and 5′-CGACCTCCTCGAACGTCTT-3′; for ubiquitin, 5′-AGAAGGAGTCCACCCTCCACC-3′ and 5′-GCATCCAGCACAGTAAAACACG-3′. PCR was performed with SYBR green I using a Light Cycler 480 System II (Roche Applied Science, Penzberg, Germany).

Quantification of fungal genomic DNA content

Genomic DNA of AM-colonized roots was extracted with Nucleon Phytopure plant DNA extraction kit (GE Healthcare UK Ltd, Little Chalfont, England). DNA (1 μg) was used for quantitative PCR (qPCR) using THUNDERBIRD SYBR qPCR Mix (Toyobo) with the Mx3000p fluorescence-detection thermocycler (Agilent Technologies). To amplify fungal and plant genomic DNAs, GiITS primers and CYC2 primers were used, respectively (Gutjahr et al., 2008). The quantity of fungal genomic DNA was calculated as the GiITS quantity divided by the CYC2 quantity.


AM fungal colonization in rice SL-insensitive mutants

To investigate the involvement of SL-related genes in AM symbiosis, rice SL-insensitive mutants d3-1 and d14-1 (cv Shiokari background) were coincubated with spores of the AM fungus G. intraradices for 4 wk. The frequencies of arbuscule and vesicle formation were significantly reduced in the d3-1 mutant (Fig. 1a). Similarly, d3-2 and d3-3 alleles in cv Nipponbare background showed significantly reduced arbuscule and vesicle formation (Supporting Information, Fig. S1a,b). These results indicated that D3 is crucial for AM symbiosis. By contrast, AM colonization rates in d14-1 were higher than in the WT (Fig. 1a). The SL biosynthesis mutant d17-1 formed fewer AM symbiotic structures, as recently reported (Gutjahr et al., 2012).

Figure 1.

Arbuscular mycorrhiza (AM) fungal colonization in rice strigolactone (SL)-insensitive mutants. Relative colonization frequency in rice wildtype (cv Shiokari) and d3-1, d14-1, and d17-1 mutants at 4 wpi with Glomus intraradices (a) and at 5 wpi with Gigaspora margarita (b). Bars represent mean values ± standard errors (n = 10). Asterisks indicate significant differences compared with the wildtype using the Mann–Whitney U-test (P < 0.05). IRH, intraradical hyphae; ERH, extraradical hyphae; arb, arbuscules; vesi, vesicles. All experiments were repeated at least three times.

Arbuscular mycorrhiza fungi have a broad host range, and their hyphal growth patterns, network formation, and branching frequencies differ significantly (de la Providencia et al., 2005; Parniske, 2008). Hence, we investigated whether the AM phenotype of d3 also occurs with other mycorrhizal fungal species. Spores of G. margarita (Gigasporaceae) were cocultivated with WT or mutant plants, and AM phenotypes were evaluated at 5 wk postinoculation (wpi). Similar to the results for G. intraradices, the colonization rates in d3-1 were significantly lower than in the WT (Fig. 1b). The AM colonization rates of d14-1 tended to be higher than the WT, but the difference was not statistically significant.

The d3 and d14 mutants are insensitive to SL and produce more SLs than the WT owing to the lack of feedback suppression (Umehara et al., 2008; Arite et al., 2009). Consistently, hyphae of G. margarita had additional branches when grown in the vicinity of roots of d3 and d14 mutants compared with the WT (Fig. S2). Hyphae had up to four branches near WT roots but had extensive branches with d3 or d14 roots (Fig. S2). No apparent differences between d3-1 and d14-1 were observed.

To investigate whether exogenous application of SLs influenced the AM colonization rate, WT and mutants d3, d10, d14, and d17 were coincubated with G. intraradices or G. margarita in the presence of exogenously applied SL (Fig. S3). Although SL application yielded a higher AM colonization rate in SL-deficient d10 and d17 mutants than in nontreated plants, this treatment did not alter the rates in WT, d3-1, and d14-1 (Fig. S3), suggesting that additional SLs did not affect the AM colonization phenotype in these lines, at least under our experimental conditions.

Mutant d3 has defects in the early stages of AM symbiosis

To find out which infection stages were affected by the d3 mutation, the AM fungal colonization phenotype in d3-1 mutants was observed under bright-field microscopy. The number of hyphopodia was counted at 2 and 4 wpi. Although slightly fewer hyphopodia formed in the d3-1 mutant than in the WT (Fig. 2a), no significant difference was detected. At 4 wpi, WT roots were colonized by AM fungi with concomitant formation of arbuscules and vesicles (Fig. 2b,c). In d3-1 roots, extraradical hyphae grew on the root surface, and hyphopodia formed. In many cases, however, there were no arbuscules or vesicles formed near infection sites (Fig. 2e). The infection attempts in the d3 mutant were often confined to the epidermal cells, and several septa formed in hyphae (Fig. 2f). Arbuscules rarely formed in d3 mutants, and in those cases the arbuscule shape was similar to that observed in the WT (Fig. 2g). No obvious difference in AM colonization phenotype was observed in d14-1 compared with the WT (Fig. 2d).

Figure 2.

Arbuscular mycorrhiza (AM) colonization phenotype of rice strigolactone (SL)-insensitive mutant. (a) Number of hyphopodia per root length at the indicated time points after coincubation with Glomus intraradices. Bars represent mean values ± standard errors (n = 3–5). (b–g) Wildtype (b, c), d14-1 (d), and d3-1 (e–g) mutants colonized by Gintraradices at 4 wpi. (b) In a wildtype root, AM colonization structures such as intraradical hyphae (IRH), extraradical hyphae (ERH), arbuscules (A), and vesicles (V) were observed. (c) A wildtype arbuscule. (d) A d14 mutant root colonized by AM fungi looked similar to the wildtype. (e) A d3-1 mutant root had several aborted infection attempts (arrows). (f) A typical hyphopodium in a d3 root showing arrest of hyphal growth in the epidermis. Numerous septa (arrowheads) were observed in hyphae. The arrow indicates an aborted infection attempt. (g) An arbuscule (A) formed in a d3-1 mutant root. Bars: 50 μm (b, d, e); 20 μm (c, f, g).

To observe the intracellular morphology of AM fungi, the fungal structures in AM-colonized roots were stained by Alex Fluor 488-conjugated WGA, and cross-sections were observed using confocal laser-scanning microscopy. In WT roots, the G. intraradices hyphae formed hyphopodia at epidermal cells, entered into the exodermal and cortical cell layers, and then formed arbuscules (Fig. 3a,c,e). In the d3 mutant, the hyphopodia formed at the epidermal cells, but hyphal growth was restricted. The aborted hyphopodia were often unusually large or had a branched structure (Fig. 3b). Cross-section images showed that G. intraradices hyphae successfully entered into epidermal cells but failed to enter subepidermal cells (Fig. 3d,f).

Figure 3.

Confocal microscopy images of Glomus intraradices infection. Fluorescence images of wildtype (a, c, e) and d3-1 mutants (b, d, f) colonized by Gintraradices at 4 wpi. Roots were double-stained with Alexa Fluor 488-conjugated wheat germ agglutinin (WGA) and propidium iodide to observe fungal structure (green) and root cell shape (red), respectively. (a, b) Merged projection images of hyphopodia (arrowhead) formed in wildtype (a) and d3-1 (b) roots. (a) In the wildtype, intraradical hyphae and arbuscules were observed. (b) In a d3-1 root, a hyphopodium had arrested growth and an aberrant shape. (c–f) Cross-sections of the arbuscular mycorrhiza (AM)-infected wildtype (c, e) and d3-1 (d, f) roots. In the wildtype, hyphae successfully penetrated the exodermis (e). In d3-1 roots, hyphae penetrated the epidermis but not the exodermis (f). arb, arbuscule; ep, epidermal cell; ex, exodermal cell. The total thickness and number of confocal laser scanning microscopy images used to make the projection image were as follows: (a) 24 μm, 17 images; (b) 19 μm, 20 images; (c) 10.5 μm, 11 images; (d) 13 μm, 14 images; (e) single scan; (f) 11.55 μm, 12 images. Bars, 20 μm.

AM marker gene expression

A set of AM-inducible genes has been identified from rice (Gutjahr et al., 2008). To investigate the role of D3 genes in AM symbiosis, we exploited four AM-inducible marker genes (AM1, AM3, AM11, PT11). AM1, AM3, and AM11 are all activated at a relatively early time point, and the induction of the latter two requires functional common sym genes, whereas AM1 induction does not (Gutjahr et al., 2008). PT11 is specifically expressed in arbuscule-containing cells (Gutjahr et al., 2008; Kobae & Hata, 2010). The expression of AM marker genes was measured by qRT-PCR in WT and mutant roots at 4 wpi with G. intraradices. In the WT, the expression of all four markers was increased by cocultivation with AM fungi, whereas the expression of AM genes was nearly abolished in the d3-1 mutant (Fig. 4a), suggesting that molecular responses to AM fungi are impaired in d3 mutants. The marker gene expression in the d14-1 mutant was seemingly higher than in the WT, but the difference was not statistically significant.

Figure 4.

Expression of arbuscular mycorrhiza (AM)-inducible genes in rice d mutants. (a) Expression of rice AM-inducible genes AM1, AM3, AM11, and PT11 was quantified by quantitative reverse transcription polymerase chain reaction (qRT-PCR) at 4 wpi with Glomus intraradices (AM). Control samples were harvested from roots grown in the same condition as AM but without fungal spores. Expression values were normalized with internal control CYCLOPHILIN2 and are presented as relative values to the wildtype (WT) AM sample. Bars represent the mean value ± standard error of individual experiments (n = 4). Each experiment used six to 10 plants. Asterisks indicate significant differences compared with the wildtype by the Mann–Whitney U-test (P < 0.05). (b) Fungal genomic DNA content in root samples was calculated by qPCR using GiITS primer sets. Asterisks indicate significant differences compared with the wildtype by the Mann–Whitney U-test (P < 0.05). (c) Expression levels of AM1, AM3, AM11, and PT11 at 4 wpi in indicated genotypes, normalized to fungal DNA content quantified by qPCR of GiITS. Expression values are presented as relative values to the wildtype (WT) sample. Bars represent the mean value ± standard error of individual experiments (n = 4). Each experiment used six to 10 plants. No significant difference (P < 0.05) was detected by Student's t-test or Mann–Whitney U-test between the wildtype and each genotype.

The defects in AM gene expression in d3 may be the result of abolishment of early signaling of AM colonization or, alternatively, correlated with the amount of AM fungi around the root samples. To test this hypothesis, fungal genomic DNA content was estimated by quantitative PCR (qPCR) using GiITS primer sets, and AM marker gene expression was then normalized to fungal DNA content. The quantity of fungal DNA in d3 was significantly lower than in the WT (Fig. 4b). Even though expression levels normalized using fungal DNA were quite variable because of the low quantity of fungal DNA, we did not find significant differences between the WT and the mutants (Fig. 4c).

Phenotype of D14-LIKE knockdown plants

In Arabidopsis, Atd14 and kai2/Atd14-like mutant phenotypes have similarities with the max2 mutants (Waters et al., 2012). Because the reduced AM colonization phenotype was observed only in d3 but not in d14, we hypothesized that the AM symbiosis may also be regulated by D14-LIKE. To investigate this possibility, D14-LIKE RNA interference (RNAi) constructs were transformed into WT and d14-1 mutant backgrounds. Two selected lines, Ri8 and Ri43, of D14-LIKE RNAi roots showed reduced expression (< 10%) of D14-LIKE (Fig. 5a). The D14/D14-LIKE double knockdown line (d14-1 Ri), in which the D14-LIKE RNAi construct was introduced into the d14-1 background, showed < 10% expression of D14-LIKE compared with the WT (Fig. 5b). Unlike the d14-1 mutant, the d14-like RNAi lines did not show dwarf or increased-tiller-outgrowth phenotypes (Fig. S4), suggesting distinct functional roles for D14 and D14-LIKE in rice. The double knockdown plants had a similar phenotype to the d14-1 single mutant, indicating that D14-LIKE was not functional in tiller outgrowth and plant height control (Fig. S4). Similarly, D10 expression was not affected by the introduction of the D14-LIKE RNAi construct into the WT or d14-1 mutant background (Fig. S5). These results suggest that D14-LIKE does not have a major function in regulating SL responses in rice.

Figure 5.

Arbuscular mycorrhiza (AM) colonization phenotype in D14-LIKE RNAi lines. (a, b) Quantitative reverse transcription polymerase chain reaction (qRT-PCR) quantification of D14-LIKE expression in independent D14-LIKE RNAi transgenic lines in either a rice wildtype cv Nipponbare (NBP) background (a, Ri8, Ri43) or in a d14-1 background in cv Shiokari (b, d14-1Ri). Expression values were normalized with an internal Ubiquitin control and are presented as relative values to the wildtype (cv Nipponbare or Shiokari) sample. Bars represent the mean value ± standard error of individual experiments (n = 3). (c, d) Glomus intraradices colonization frequency (± standard errors; n = 4–11) in D14-LIKE RNAi transgenic lines (Ri8, Ri43, d14-1Ri), d3 alleles, d14-1, and the wildtype (cv Nipponbare or Shiokari). Colonization was measured at 5 wpi. Asterisks indicate a significant difference (P < 0.05) compared with the wildtype according to Student's t-test.

We performed AM colonization assays using D14-LIKE RNAi lines and double knockdown lines. The D14-LIKE single RNAi plants showed AM colonization frequencies similar to that of the WT. The double knockdown line showed AM colonization rates higher than the WT but similar to that of d14-1 (Fig. 5c,d). Based on these results, we conclude that neither D14 nor D14-LIKE is essential for AM colonization.


In this report, we present a novel phenotype of the rice SL-insensitive mutant d3 in AM symbiosis. Because SLs stimulate the growth of AM fungi, the amount of SL production generally correlates with AM colonization. Mutants or transgenic plants that produce fewer SLs show reduced AM fungal colonization (Gomez-Roldan et al., 2008; Koltai et al., 2010; Liu et al., 2011; Gutjahr et al., 2012; Kretzschmar et al., 2012). By contrast, G. margarita hyphal branches are activated near roots of d3 and d14 mutants that produce higher amounts of SLs (Umehara et al., 2008; Arite et al., 2009; Fig. S2). The d14 mutant had more AM colonization, possibly reflecting higher AM fungal activities because of increased SL production in the roots, although we cannot exclude the possibility that D14 negatively regulates AM colonization. However, the d3 mutants also produce more SLs and showed strong defects in AM colonization. It is unlikely that higher concentrations of SLs inhibit AM colonization because d3 and d14 similarly produce higher amounts of SLs. In addition, the exogenous application of SLs did not affect AM colonization.

In the d3 mutants, G. intraradices hyphae were often unable to extend beyond the epidermal cell layer (Figs 2, 3). Progressive internal hyphal growth and arbuscule and vesicle formation were all significantly reduced. Once penetration was achieved, however, arbuscule shapes were similar to those of WT. The phenotype of abolished penetration is reminiscent of those found in the common sym mutants of rice (Banba et al., 2008; Gutjahr et al., 2008). However, intracellular hyphal progression is abolished in common sym mutants, whereas arbuscule formation is occasionally seen in d3 mutants. The expression of four AM marker genes, including the common sym-independent marker AM1, was almost abolished in the d3 mutants. Thus the molecular responses to AM fungi are dramatically reduced in the d3 mutants correlating to low fungal colonization rate. The d3 mutant hyphopodia phenotype is similar to that of Vapyrin RNAi lines in M. truncatula and pam1/vapyrin mutants in petunia (Reddy et al., 2007; Pumplin et al., 2010). In pam1 mutants, hyphae penetrate epidermal cells and develop complex hyphopodia with numerous septa, and in many cases penetration does not proceed further (Reddy et al., 2007). Similar aborted hyphopodia with complex shapes and hyphae with septa were also observed in the d3 mutants. Further analysis will be required to clarify the functional relationship between D3 and Vapyrin.

The D3 gene encodes an F-box protein, which is a component of the SCF complex that mediates target protein degradation, and some F-box proteins form coreceptor complexes for plant hormones such as GA, auxin, and jasmonic acid (Somers & Fujiwara, 2009). Thus, based on the SL-insensitive phenotype of the d3 mutant, it is also possible that D3 is a part of the SL receptor. Recent studies revealed that Arabidopsis MAX2, a D3 ortholog, is also required for karrikin responses, suggesting that D3/MAX2 may be a part of a karrikin receptor complex (Waters et al., 2012). In Arabidopsis, D14 and D14-LIKE are essential for SL and karrikin responses, respectively, suggesting that these proteins may form distinct complexes with MAX2 that recognize these molecules (Waters et al., 2012). It remains unclear whether the AM phenotypes observed in d3 mutants are the result of defects in SL or karrikin perception/signaling. However, our analysis of a double knockdown line showed that D14 and D14-LIKE are not involved in AM colonization, indicating a distinct pathway for AM colonization. AM symbiosis may be regulated by SL or karrikin but independent of D14/D14-LIKE, or D3 may be involved in recognizing an as yet unidentified chemical(s) that is required for AM symbiosis. Regardless of SL dependency, our results indicate that the physiological role of D3 is distinct from that of the D14 or D14-LIKE-dependent response. Although it remains to be elucidated how D3 regulates AM symbiosis, our results suggest its essential role in plant–microbe interactions.


We thank Professor K. Mori for providing synthetic (+)-strigol and Mrs Akiko Ueno for technical support. This work was funded by grants from KAKENHI (23128513 and 23657044 to S.Yo., 24228008 to K.S. and 23370025 and 24114010 to S.Ya.) and Program for Promotion of Basic and Applied Researches for Innovations in Bio-oriented Industry (PROBRAIN) to K.A., J.K. and S.Ya.