OsIPD3, an ortholog of the Medicago truncatula DMI3 interacting protein IPD3, is required for mycorrhizal symbiosis in rice


Author for correspondence:
Hongyan Zhu
Tel: +1 859 257 3647
Fax: +1 859 323 1077
Email: hzhu4@uky.edu


  • • Medicago truncatula IPD3 (MtIPD3) is an interacting protein of DMI3 (does not make infections 3), a Ca2+/calmodulin-dependent protein kinase (CCaMK) essential for both arbuscular mycorrhizal (AM) and rhizobial symbioses. However, the function of MtIPD3 in root symbioses has not been demonstrated in M. truncatula, because of a lack of knockout mutants for functional analysis. In this study, the availability of IPD3 knockout mutants in rice (Oryza sativa) was exploited to test the function of OsIPD3 in AM symbiosis.
  • • Three independent retrotransposon Tos17 insertion lines of OsIPD3 were selected and the phenotypes characterized upon inoculation with the AM fungus Glomus intraradices.
  • • Phenotypic and genetic analyses revealed that the Osipd3 mutants were unable to establish a symbiotic association with G. intraradices.
  • • In conclusion, IPD3 represents a novel gene required for root symbiosis with AM fungi in plants.


Arbuscular mycorrhizal (AM) and rhizobial symbioses are two important mutually beneficial associations formed between plants and microbes. The AM symbiosis is widespread, occurring in > 80% of land plants (Remy et al., 1994; Heckman et al., 2001), whereas the rhizobial symbiosis is almost completely restricted to leguminous plants (Soltis et al., 1999). Through the AM symbiosis, AM fungi assist the plant in assimilating mineral nutrients, particularly inorganic phosphate, from the soil, whereas the legume–rhizobium symbiosis results in the formation of the root nodule in which the bacteria fix atmospheric nitrogen for use by the plant. Thus, the two symbioses are of crucial importance in sustainable agriculture.

In recent years, numerous plant genes required for root symbioses have been cloned from the two model legumes Medicago truncatula and Lotus japonicus (reviewed by Stacey et al., 2006). In addition to the genes that are specific to the nodulation (Nod) factor signaling pathway, at least seven genes, so called the common symbiosis genes, have been identified that are essential for both AM and nodulation symbioses (Kistner et al., 2005). Among the common signaling components are the Ca2+/calmodulin-dependent protein kinase MtDMI3/L. japonicus CCaMK/Pisum sativum SYM9 (Lévy et al., 2004; Mitra et al., 2004; Tirichine et al., 2006). MtDMI3 acts downstream of Nod factor-induced calcium spiking (Ehrhardt et al., 1996; Oldroyd & Downie, 2004, 2006, 2008), and thus presumably functions to interpret the calcium spiking signal, which subsequently activates the downstream transcription factors leading to transcriptional reprogramming of the host symbiotic genes (Schauser et al., 1999; Kalóet al., 2005; Smit et al., 2005; Middleton et al., 2007). Despite the identification of a number of nodulation-specific signaling components (NFP, NSP1, NSP2, ERN, etc.; Madsen et al., 2003; Radutoiu et al., 2003; Kalóet al., 2005; Smit et al., 2005; Arrighi et al., 2006; Middleton et al., 2007), mycorrhiza-specific signaling components are largely unknown (Harrison, 2005).

Biochemical approaches have been used to identify additional players involved in symbiotic signaling. For example, yeast-two-hybrid (Y2H) screens and immunoprecipitations have identified 3-hydroxy-3-methylglutaryl CoA reductase 1 (MtHMGR1) as an interactor of MtDMI2 (Kevei et al., 2007). Knockdown of MtHMGR1 expression by RNA interference (RNAi) in M. truncatula revealed a crucial role of MtHMGR1 in nodule development. Recently, an interacting protein of MtDMI3, named MtIPD3, has also been identified in M. truncatula (Messinese et al., 2007). Like other symbiotic genes, the IPD3 orthologs are universally present and highly conserved in angiosperms that have the ability to establish AM symbioses (Zhu et al., 2006). However, the function of MtIPD3 in root symbioses has not been demonstrated, because of a lack of null Mtipd3 mutants for functional analysis and the absence of obvious nodulation phenotypes for the MtIPD3 knockdown roots inoculated with Sinorhizobium meliloti (Messinese et al., 2007). Here we take advantage of the availability of IPD3 knockout mutants in rice (Oryza sativa) and demonstrate that OsIPD3 is required for AM symbiosis in rice. Thus, IPD3 represents a novel gene required for root symbiosis with AM fungi in plants.

Materials and Methods

Isolation and characterization of homozygous Osipd3 mutant lines

The rice (Oryza sativa L.) Tos17 insertion lines (NC0263, NC2713, and NC2794) in the ‘Nipponbare’ background were provided by the Rice Genome Resource Center of the National Institute of Agrobiological Sciences (RGRC-NIAS), Ibaraki, Japan. The accession numbers of the 3′-flanking OsIPD3 sequences of Tos17 insertion in NC0263, NC2713, and NC2794 are AG021800, AG023104, and AG023190, respectively. Because seeds of the mutant lines from the provider were from a mixed progeny of a primary tissue culture-derived plant, we carried out PCR experiments to isolate homozygous mutants. The PCR experiments were performed by adding three primers in a PCR reaction mixture, including a primer pair flanking the insertion sites, and a Tos17-specific primer. The PCR conditions were set to allow amplification of only the wild-type allele and not the insertion alleles by the flanking primer pairs. The positions of these primers are indicated in Fig. 1(a). Primer sequences were as follows: F1: 5′-ATGCAGCAGAAAAGGGAATG-3′; R1: 5′-ATGCTGTACCAAGCCAAACC-3′; F2: 5′-GGTTTGGCTTGGTACAGCAT-3′; R2: 5′-TCCTTGTTGGGTTTACCTGC-3′; and Tos17-specific primer T1: 5′-ATTGTTAGGTTGCAAGTTAGTTAAGA-3′.

Figure 1.

Isolation and characterization of Tos17 insertion mutants of Oryza sativa OsIPD3. (a) Gene structure of OsIPD3, and the Tos17 insertion sites. The exons and introns are indicated by boxes and lines, respectively. Insertion sites of Tos17 are indicated. (b) OsIPD3 expression levels in roots, stems, leaves, panicles, and mycorrhizal roots. Relative transcript abundance was determined by quantitative reverse transcription (RT)-PCR and normalized against ubiquitin 1 (OsUBI1). Error bars represent ± SD from three independent biological replications. (c) Identification of homozygous (−/−) insertion mutants by PCR. The bands with a larger size represent the wild-type allele amplified from a primer pair flanking the Tos17 insertion sites. The bands with a smaller size represent the mutant allele amplified from a pair of Tos17-specific and IPD3-specific primers. WT, the wild-type genotype Nipponbare; WT*, the wild-type genotype for the OsIPD3 segregated from a heterozygous plant. Het, heterozygous; Homo, homozygous. (d) RT-PCR analysis of OsIPD3 expression in the wild-type and mutant plants. ACT, actin.

Inoculation of rice roots with Glomus intraradices

The AM fungus G. intraradices was ordered from Premier Tech Biotechnologies (Canada). The inoculation method was as described by Chen et al. (2007). Roots were harvested at 7 wk after inoculation. A random sample of the root tissues was used for phenotypic analysis, and the remaining tissues were used for RNA isolation.

Mycorrhizal colonization was assessed by means of WGA-AlexaFluor 488 staining (Molecular Probes, Eugene, OR, USA). The cleaned roots were first fixed in 50% (v/v) ethanol. The fixed roots were then incubated at 90°C in 5% KOH for 20 min. After rinsing thoroughly with distilled water, the roots were soaked in 0.1 M HCl at room temperature overnight. The roots were then rinsed with phosphate-buffered saline (PBS) buffer and placed in PBS-WGA staining solution at room temperature. The final concentration of WGA was 0.2 µg ml−1. The stained roots were examined using a fluorescent microscope (Axioplan2; Carl Zeiss Optical, Chester, VA, USA) and images were captured by a microscope digital camera system (AxioCam MRc5; Carl Zeiss). Quantification of root colonization was based on an assessment of the percentage of colonized root segments in a sample of root segments c. 1.0 cm in length.

Analysis of gene expression

Total RNA was isolated using the Qiagen Plant RNeasy kit (Qiagen, Valencia, CA, USA). RNA (2.0 µg) was used to perform reverse transcription (RT) reactions using murine Moloney leukaemia virus (MMLV) reverse transcriptase (Invitrogen) in a 20-µl reaction mixture. Two microliters of the RT reaction was used as a template in a 20-µl PCR reaction solution. The PCR primers were as follows: actin (OsACT), 5′-GCGATAATGGAACTGGTATG-3′ and 5′-CTCCATTTCCTGGTCATAGTC-3′; OsIPD3, 5′-ATGCAGCAGAAAAGGGAATG-3′ and 5′-TCCTTTGCTTCTGCCATCTT-3′; OsPT11, 5′-ATGGCTCGACGGACAGTAAG-3′ and 5′-GATCAGCTGGATCATGTACCT-3′. Quantitative RT-PCR was performed on a StepOne Real-time PCR System (Applied Biosystems, Foster City, CA, USA) using the SYBR Green I detection kit (BioRad, Hercules, CA, USA). The ubiquitin (OsUBI) gene was selected as a constitutive internal control. PCR primers used for the real-time PCR experiments were: OsUBI, 5′-TGCACCCTAGGGCTGTCAAC-3′ and 5′-TGACGCTCTAGTTCTTGATCTTCTTC-3′; OsIPD3, 5′-GGTTTGGCTTGGTACAGCATCT-3′ and 5′-GGGAGGCAGGTCATCACAA-3′.


Characterization of OsIPD3

MtIPD3 orthologs are universally conserved in angiosperms, except for Arabidopsis which lacks the ability to establish symbiotic associations with AM fungi (Messinese et al., 2007). OsIPD3 was identified as Os06g02520, a single-copy gene located on chromosome 6 of the rice genome (‘Nipponbare’). Alignment of the full-length cDNA sequence (AK243034) with the genomic sequence revealed a gene structure of 11 exons (Fig. 1a), which is conserved relative to MtIPD3 (Messinese et al., 2007) and its other legume and nonlegume counterparts (data not shown). The putative OsIPD3 protein consists of 506 amino acid residues with a domain structure identical to that of MtIPD3. OsIPD3 and MtIPD3 share c. 45% global identity and up to 86% identity over the C-terminal 90 amino acids (coiled-coil domain). In silico analysis of the rice massively parallel signature sequencing (MPSS) database indicated that the expression of OsIPD3 was only detectable in the rice root (Nobuta et al., 2007). Quantitative RT-PCR analysis performed in this study confirmed that OsIPD3 is predominantly expressed in rice roots, with an expression level c. 25-fold higher than that of stems and panicles, and c. 10-fold higher than that of leaves (Fig. 1b). The expression level of OsIPD3 in mycorrhizal roots appears to be c. 2-fold higher than that of control plants. A similar expression pattern was also observed for MtIPD3 in M. truncatula (Messinese et al., 2007).

Isolation and characterization of Osipd3 mutants in rice

To examine the function of OsIPD3 in root symbioses, we searched the rice mutant databases for putative Tos17 insertion lines (Miyao et al., 2007). From multiple Tos17 insertion lines available for OsIPD3, we selected three independent insertion alleles, from the tissue culture-derived lines NC0263, NC2713, and NC2794, for further analysis. For all three lines, the retrotransposon Tos17 was inserted into the sixth exon of OsIPD3 (Fig. 1a). From progeny of each of the primary mutant lines, positive and homozygous Tos17 insertion plants were identified by PCR analysis (Fig. 1c). As shown in Fig. 1(d), for all three mutants, the expression of OsIPD3 was disrupted based on RT-PCR analyses. As Tos17 mutant lines comprise multiple insertion sites in a single genome, wild-type plants (for the OsIPD3 locus) segregated from the progeny of the heterozygous mutant lines were used as additional controls for the experiments described below.

Osipd3 mutants are unable to form mycorrhizal symbiosis

To determine whether OsIPD3 is essential to form AM symbiosis in rice, we inoculated the mutant and wild-type rice roots with the mycorrhizal fungus G. intraradices. At 7 wk after inoculation, wild-type plants were normally colonized by G. intraradices, revealing the range of symbiotic structures typical of a functional symbiosis, including intercellular and intracellular hyphae, vesicles, and arbuscules (Fig. 2a; Table 1). An average of > 80% of the total root length was colonized by fungal hyphae and arbuscules (n = 40 plants). An equivalent level of colonization was also observed for wild-type segregants derived from heterozygous mutant parental plants (n = 16 from each of the three lines). In contrast, arbuscules and vesicles were not detected on roots of a total of 32 NC0263, 16 NC2713, and 32 NC2794 homozygous mutant plants. For homozygous mutant plants, extraradical hyphal growth and the numbers of appressoria were significantly reduced on the root surface (Table 1). Despite the presence of fungal hyphae and appressoria on the root surface (Fig. 2b), the fungus failed to penetrate the root epidermis. Only in one root segment of the NC2713 root samples did we detect aborted intracellular fungal hyphae. Overall, the observed phenotypes of Osipd3 mutants were similar to those of the Osdmi3 mutants (Chen et al., 2007). As a molecular correlate of symbiotic function, we examined expression of the rice mycorrhiza-specific phosphate transporter OsPT11 (Paszkowski et al., 2002). As shown in Fig. 2(c), OsPT11 transcript was readily detected in wild-type roots inoculated with G. intraradices but not in roots homozygous for the OsIPD3 mutation. Taken together, these data indicate that OsIPD3 is required for AM symbiosis in rice.

Figure 2.

Phenotypic and molecular analysis of Tos17 insertion mutants of Oryza sativa (rice) OsIPD3 upon inoculation with Glomus intraradices. (a) Roots of wild-type rice plants formed arbuscules upon inoculation with G. intraradices. (b) Osipd3 mutants failed to form arbuscular mycorrhizal (AM) symbiosis despite the presence of fungal hyphae on the root surface. (c) Expression of OsPT11 in the wild-type and mutant roots under inoculated and noninoculated conditions. ACT, actin.

Table 1.  Fungal colonization in wild-type (WT) and Osipd3 rice (Oryza sativa) mutants at 7 wk after inoculation
 Extraradical hyphaeIntraradical hyphaeAppressoriaArbusculesVesiclesAborted
  1. The calculation was based on the percentage of colonized root segments in a sample of root segments c. 1.0 cm in length. Arbuscules and vesicles were also counted as intraradical hyphae. For wild-type roots, when intraradical hyphae were detected, the presence of appressoria was presumed even if the structure of appressoria was not observed.

WT85.5 ± 3.552.0 ± 1.061.0 ± 2.335.0 ± 2.120.0 ± 5.30
NC02635.7 ± 4.703.5 ± 4.9000
NC271316.0 ± 2.81.2 ± 1.79.8 ± 0.7001.2 ± 1.7
NC27947.5 ± 4.002.6 ± 3.6000


MtIPD3 was identified as an interacting protein of DMI3 in M. truncatula (Messinese et al., 2007). Both DMI3 and IPD3 orthologs are conserved in leguminous and nonleguminous plants, with the notable exception of the nonmycorrhizal plant Arabidopsis in which both proteins are absent. These findings suggest that IPD3 is likely to play a crucial role in root symbioses. However, the involvement of IPD3 in root symbioses has not been demonstrated. Knockdown of IPD3 expression in M. truncatula roots did not result in obvious nodulation phenotypes (Messinese et al., 2007) and knockout mutants of IPD3 in M. truncatula are currently unavailable. Here, we used rice IPD3 knockout mutants to demonstrate that OsIPD3 is required for AM symbiosis in rice. Work is in progress to screen Mtipd3 knockout lines from the Noble Foundation (Ardmore, OK, USA) mutant populations and to determine whether MtIPD3 is also required for root nodule symbiosis in M. truncatula.

Nearly all identified legume genes essential for root symbioses have their orthologs in nonlegumes such as rice (Zhu et al., 2006), providing an opportunity to study symbiosis signaling pathways in nonlegumes. We and others have demonstrated that the rice orthologs of the legume common symbiosis genes LjCASTOR, LjPOLLUX, MtDMI2/LjSYMRK, and MtDMI3 are all required for AM symbiosis (Chen et al., 2007; Markmann et al., 2008; C. Chen & H. Zhu, unpublished data), indicating that the mycorrhizal symbiosis pathways are highly conserved among plant families. Thus, parallel and comparative analysis of mycorrhizal signaling in legumes and nonlegumes will take advantage of the resources available in other model plants such as rice. Such analyses are likely to reveal signaling components that separate the mycorrhizal and rhizobial signaling downstream of the common signaling pathways and provide further insights into the evolution of root symbioses in plants.


We thank the Rice Genome Resource Center (RGRC, Japan) for providing the rice Tos17 mutant lines. This project was supported by the US National Science Foundation (grant no. 0640197 to HZ) and by the US Department of Agriculture (USDA-CSREES 2005-16202 to JMA).