Characterization of the Silicon Uptake System and Molecular Mapping of the Silicon Transporter Gene in Rice

Abstract

Rice (Oryza sativa L. cv Oochikara) is a typical silicon-accumulating plant, but the mechanism responsible for the high silicon uptake by the roots is poorly understood. We characterized the silicon uptake system in rice roots by using a low-silicon rice mutant (lsi1) and wild-type rice. A kinetic study showed that the concentration of silicon in the root symplastic solution increased with increasing silicon concentrations in the external solution but saturated at a higher concentration in both lines. There were no differences in the silicon concentration of the symplastic solution between the wild-type rice and the mutant. The form of soluble silicon in the root, xylem, and leaf identified by ²⁹Si-NMR was also the same in the two lines. However, the concentration of silicon in the xylem sap was much higher in the wild type than in the mutant. These results indicate that at least two transporters are involved in silicon transport from the external solution to the xylem and that the low-silicon rice mutant is defective in loading silicon into xylem rather than silicon uptake from external solution to cortical cells. To map the responsible gene, we performed a bulked segregant analysis by using both microsatellite and expressed sequence tag-based PCR markers. As a result, the gene was mapped to chromosome 2, flanked by microsatellite marker RM5303 and expressed sequence tag-based PCR marker E60168.

Rice (Oryza sativa L. cv Oochikara) requires high silicon for healthy growth and stable and high productivity (Savant et al., 1997; Ma and Takahashi, 2002). Silicon at up to 10% of dry weight is accumulated in the shoot, and more than 90% of silicon is present in the form of silica gel (Ma and Takahashi, 2002). Silica gel is deposited on the cell wall of epidermal cells of leaves, stems, and hulls, forming a silica-cuticle double layer and a silica-cellulose double layer (Yoshida, 1965; Raven, 2003). Silicon is also deposited on the bulliform cells, dumbbell cells, and long and short cells on the surface of leaves and hulls. The deposition of silicon enhances the strength and rigidity of cell walls and thus increases the resistance of rice to diseases, pests, and lodging, improving light-receiving plant form in a community and decreasing transpiration (Epstein, 1994, 1999; Ma and Takahashi, 2002; Ma, 2003). Thus, silicon plays an important role in enhancing the resistance of rice to multiple stresses, including biotic and abiotic stresses (Ma, 2004). Silicate fertilizer has been applied to paddy soils to increase rice productivity.

High accumulation of silicon in rice has been attributed to the ability of the roots to take up silicon (Takahashi et al., 1990; Richmond and Sussman, 2003). Recently, a rice mutant that has a low silicon concentration in the shoots was isolated from sodium azide-treated M₂ seeds of rice (Ma et al., 2002). This mutant (low silicon rice 1, lsi1, formerly GR1) had a plant type similar to the wild type except that the leaf blade of lsi1 remained droopy when silicon was supplied. The silicon concentration of the tops was much lower in the mutant than in the wild type, while that of the roots was similar. A short-term uptake experiment showed that the silicon uptake by the mutant was significantly lower than that by the wild type, while there was no difference in the uptake of other nutrients such as phosphorus and potassium. Further, silicon uptake by the wild-type rice was inhibited by metabolic inhibitors, including NaCN and 2,4-dinitrophenol, and by low temperature, whereas silicon uptake by lsi1 was not inhibited by these agents. The silicon concentration in the xylem sap of the wild-type rice was also much higher than that of lsi1. These results suggest that an active transport system for silicon uptake is disrupted in the mutant (Ma et al., 2002).

Generally, the uptake of a mineral by the roots includes at least two processes: radial transport of the mineral from external solution to cortical cells and release of the mineral from cortical cells into the xylem (xylem loading; Marschner, 1995). However, these processes for silicon uptake in rice roots remain poorly understood. In this study, we characterized the silicon uptake system in rice roots by using the mutant (lsi1) in comparison to wild-type rice. The gene responsible for the low silicon uptake was mapped by using both expressed sequence tag (EST)-based PCR and microsatellite markers.

RESULTS AND DISCUSSION

Characterization of the Silicon Uptake System in Rice Roots

For decades, rice has been known as the most effective silicon-accumulating species, although the mechanisms involved in high silicon uptake are poorly understood. One of the reasons is, unlike other minerals, the genotypic difference in silicon concentration of rice is too small to be utilized for comparative study on silicon uptake by rice roots. An alternative approach is to use a rice mutant (lsi1) with a low shoot silicon concentration (Ma et al., 2002). This mutant provides a useful tool for studying the silicon uptake mechanism and cloning genes for silicon uptake. In this study, the silicon uptake system was characterized in terms of radial transport of silicon from external solution to the cortical cells and of xylem loading of silicon by comparing the mutant and wild-type rice. To study the former process, we obtained the symplastic solution of roots by centrifugation after removing the apoplastic solution and then freeze-thawing. A time-course experiment showed that silicon concentration in the root-cell symplast from the plants cultured in 0.5 mm silicon solution increased with time in both lines and reached a plateau after 8 h (Fig. 1). There was no significant difference in the silicon concentration of the symplast between the wild type and the mutant. The silicon concentration in the apoplastic solution was similar to that in the external solution (Fig. 1), but the silicon concentration in the symplastic solution was 3- to 7-fold higher than that in the external solution. This result suggests that silicic acid is transported against a concentration gradient from the external solution to the cortical cells. This is in agreement with previous findings that the silicon uptake is inhibited by metabolic inhibitors (Okuda and Takahashi, 1962) and low temperature (Ma et al., 2002), although the whole uptake was measured in these experiments.

Figure 1.

Concentrations of silicon in apoplastic and symplastic solutions of rice roots exposed to silicon solution for various durations. Apoplastic and symplastic solutions were extracted by centrifugation from the root tips of wild-type (WT) and mutant (lsi1) rice cultured in half-strength Kimura nutrient solution containing 0.5 mm silicon as silicic acid. Values are means ± sd of three replicates.

A kinetic study showed that the silicon concentration in the root-cell symplast increased with increasing silicon concentration in external solution but saturated at a higher silicon concentration in both lines (Fig. 2). Again, there was no significant difference in the silicon concentration of symplastic solution between the wild type and the mutant. These results suggest that silicon transport from the external solution to the root cortical cells is mediated by a type of transporter and that the transporter of the mutant is identical to that of the wild type. Based on the curve in Figure 2, the K_m value was estimated to be 0.15 mm silicon. This K_m value for silicon is much higher than that for other minerals such as phosphorus (Kochian, 2000), but taking the silicon concentration in the soil solution into consideration, which ranges from 0.1 to 0.6 mm (Epstein, 1994), this value seems reasonable.

Figure 2.

Concentration of silicon in the symplastic solution of rice root tips exposed to silicon solution with various silicon concentrations. Symplastic solution was extracted from root tips of wild-type (WT) and mutant (lsi1) rice cultured in half-strength Kimura nutrient solution containing various concentrations of silicon for 8 h. Values are means ± sd of three replicates.

A kinetic study on xylem loading of silicon was then conducted in the wild-type and mutant rice. In contrast to the silicon concentration in the root cortical cell symplast, the silicon concentration in the xylem sap was much higher in the wild type than in the mutant (Fig. 3). In the mutant, the silicon concentration in the xylem sap increased gradually with increasing silicon concentration in the external solution without saturation. In the wild-type rice, the silicon concentration in the xylem sap also increased with increasing silicon concentration in the external solution, but it was saturated at a higher concentration. The silicon concentration in the xylem sap of the wild type was higher than 30 mm at 0.9 mm silicon supply (Fig. 3). This concentration was much higher than that in root-cell symplast (Fig. 2), suggesting that silicon is transported from the root cells to the xylem also against a concentration gradient. The curve of Figure 3 also suggests that the release of silicon into the xylem is mediated by a type of transporter in the wild type. As there has been no way to determine the silicon concentration in the xylem parenchyma cells, the K_m value for silicon xylem loading could not be estimated. However, taking the K_m value (0.32 mm) for the whole uptake into consideration (Tamai and Ma, 2003), it seems that the transporter for xylem loading is different from that for the transport from external solution to the root cells (Fig. 2).

Figure 3.

Concentration of silicon in the xylem sap of rice cultured in silicon solution at various concentrations. Seedlings (26 d old) of wild-type (WT) and mutant (lsi1) rice were cultured in half-strength Kimura nutrient solution containing various concentrations of silicon. The stem was severed after 8 h, and the xylem sap was collected for 30 min. Values are means ± sd of three replicates.

The silicon form in the xylem sap has been recently identified to be monosilicic acid, while the existence of monosilicic acid at a high concentration in the xylem was transient (J.F. Ma, unpublished data). The silicon form in the roots, xylem, and leaves was compared by using ²⁹Si-NMR technique to examine whether there are any other differences between the wild type and the mutant. Only one signal in all samples was observed at a chemical shift of −72.5 ppm (Fig. 4). This chemical shift is consistent with that of monosilicic acid, suggesting that soluble silicon in the roots, xylem, and leaves is present in the form of monosilicic acid and that there is no difference in the silicon form between the wild type and the mutant. The peak intensity in the NMR spectra of root-cell sap was nearly the same between the two lines, but that of xylem sap was higher in the wild type than in the mutant (Fig. 4), confirming the results shown in Figures 1 to 3.

Figure 4.

The ²⁹Si-NMR spectra of leaf-cell sap, root-cell sap, and xylem sap of rice. After seedlings of wild-type (WT) and mutant (lsi1) rice were cultured in ²⁹Si-enriched 0.5 mm silicon solution for 4 h, the xylem sap was collected for 30 min, and leaf and root-cell saps were extracted from frozen-thawed samples. Spectra were measured at 99.4 MHz for ²⁹Si resonance frequency.

Above results suggest that at least two transporters are involved in the silicon uptake by rice roots (Fig. 5). One is located on the plasma membrane of root cortical cells (SIT1, silicon transporter 1), which transport silicon from external solution to the root cortical cells. The other is located on the plasma membrane of xylem parenchyma cells (SIT2, silicon transporter 2), which is responsible for releasing silicon into the xylem. These transporters may have different affinities for silicic acid (Figs. 2 and 3). Our results also clearly showed that the mutant is defective in xylem loading of silicon rather than transport of silicon from the external solution to the root cell.

Figure 5.

Schematic representation of the silicon uptake system in rice roots. SIT1, Silicon transporter from external solution to cortical cells. SIT2, Silicon transporter for xylem loading.

Molecular Mapping of the Gene for Xylem Loading of Silicon

As discussed above, silicon uptake by rice roots is regulated by at least two different transporters. However, neither the gene encoding the silicon transporter nor the transporter protein itself has been isolated from rice and other higher plants. In a marine diatom (Cylindrotheca fusiformis) that requires silicon as an essential element, a gene family encoding a silicon transporter has been identified (Hildebrand et al., 1993, 1997). Furthermore, a similar silicic acid transporter was found in all of the diatoms that they examined. However, similar genes were not found in rice in a sequence homology search. Further, when the silicon transporter gene from diatoms was transferred into tobacco (Nicotiana tabacum), increased silicon uptake was not observed in transgenic tobacco (J.F. Ma, unpublished data). These facts suggest that the silicon uptake system is different between diatoms and higher plants. In this study, molecular mapping of genes for silicon uptake by rice roots was performed as an initial step toward cloning silicon transporter genes. F₁ hybrids were obtained by crossing the mutant (lsi1) characterized above and Kasalath, an indica variety of rice, and F₂ seeds were then generated from the F₁. Measurement of silicon uptake showed that F₂ populations were segregated in the silicon uptake. Of 105 F₂ plants, 80 plants showed a high silicon uptake similar to Kasalath (1.75 ± 0.25 mg silicon g⁻¹ root fresh weight 6 h⁻¹), and 25 plants showed a low uptake similar to the mutant (0.14 ± 0.04 mg silicon g⁻¹ root fresh weight 6 h⁻¹; Fig. 6). This segregation ratio fitted to a 3:1 ratio (χ² = 3.41, 0.05 < P < 0.10), confirming a previous conclusion from an F₂ population derived from a cross between the mutant and its wild-type cultivar (Oochikara) that low silicon uptake by the mutant is controlled by a single recessive gene (Ma et al., 2002).

Figure 6.

Frequency distributions for silicon uptake in a progeny resulting from genetic crosses between Kasalath and a mutant (lsi1). Silicon uptake by each seedling in a nutrient solution containing 0.6 mm silicon was determined during 6 h.

To map the gene (Lsi1), which is associated with silicon uptake, we first performed bulk segregant analysis by pooling equal amounts of DNA from 10 low silicon uptake F₂ plants or 10 high silicon uptake F₂ plants based on the results shown in Figure 6. A total of 128 EST-based PCR markers that are scattered on whole chromosomes were selected to examine polymorphism among the low silicon uptake and high silicon uptake bulk, the mutant, and Kasalath. As a result, markers (C53493–C63223) on chromosome 2 were linked to the gene for low silicon uptake genes. To further map the gene for silicon uptake, we selected additional EST-based PCR markers and microsatellite markers around the above markers on chromosome 2 and tested them for polymorphism in the low silicon uptake F₂ plants. The results showed that the flanking markers were microsatellite markers RM5303 and RM5631 and EST-based marker E60168. Further investigation of the polymorphism using these three markers in 105 F₂ plants revealed that the target gene (Lsi1) and markers were located between RM5303 and E60168 on chromosome 2 (Fig. 7). As the mutant is defective in xylem loading of silicon as characterized above, this gene encodes the SIT2 (Fig. 5).

Figure 7.

Linkage relationship between EST-based PCR and microsatellite markers and a gene for silicon xylem loading (Lsi1) on chromosome 2 in rice.

Recently, a gene responsible for xylem loading of boron has been isolated in Arabidopsis (Takano et al., 2002). Like silicon, boron is also taken up in the form of an undissociated molecule. However, physiological evidence has shown that transporters for boron and silicon are different at least in rice (Tamai and Ma, 2003). Currently, a map-based cloning of Lsi1 is in progress in our laboratory.

MATERIALS AND METHODS

Extraction of Symplastic Sap

Seeds of both wild-type rice (Oryza sativa L. cv Oochikara) and a low-silicon rice mutant, lsi1 (formerly called GR1), were soaked in water overnight at 25°C in the dark. The seeds were then transferred to a net floated on 0.5 mm CaCl₂ solution in a plastic container. On day 6, the seedlings were transferred to a 1.5-L plastic pot containing one-half-strength Kimura B solution (pH 5.6) with silicic acid. The composition of Kimura B solution was previously reported (Ma et al., 2001). Silicic acid was prepared by passing potassium silicate through a cation-exchange resin (Amberlite IR-120B, H⁺ form; Organo, Tokyo; Ma et al., 2001). Apoplastic and symplastic solutions were extracted by centrifugation with slight modifications of the methods of Yu et al. (1999). For a time-course experiment, the root tips (0–1.5 cm) were excised at time points indicated in Figure 1 after exposure to 0.5 mm silicon solution at 25°C. For each sample, 30 roots were used. The cut ends were washed in distilled water quickly and blotted dry. The tips were placed in a 0.45 μm filter unit (ULTRAFREER-MC; Millipore, Bedford, MA) with the cut ends facing down and centrifuged at 2,000g for 15 min at 4°C to obtain the apoplastic solution. After centrifugation, root segments were frozen at −80°C for 2 h and then thawed at room temperature. The symplastic solution was prepared from frozen-thawed tissues by centrifugation at 2,000g for 15 min at 4°C. The silicon concentration in the apoplastic and symplastic solutions was determined immediately as described below. In a preliminary experiment, we confirmed that the freeze-thawing process did not affect the silicon concentration.

For a kinetic study, seedlings (6-d-old) prepared as above were cultured in one-half-strength Kimura B solution (pH 5.6) containing various silicon concentrations. After an 8-h culture, the apoplastic and symplastic solutions were extracted as described above. All experiments were conducted with three replicates.

To check the purity of apoplastic solution, we determined the activity of malic dehydrogenase in apoplastic and symplastic solution according to Bergmeyer and Bernt (1974). The activity of malic dehydrogenase in apoplastic solution was below one-twentieth and approximately one-fortieth of symplastic solution.

Kinetic Study of Silicon Xylem Loading

Seedlings (6 d old) prepared as described above were transferred to a 3-L plastic pot containing one-half-strength Kimura B solution (pH 5.6). The solution was renewed every 2 d. On day 20, the seedlings were transplanted to a 250-mL plastic bottle (4 seedlings per pot) containing nutrient solution with various silicon concentrations. After the culture for 8 h, the stem was severed at 1 cm above the roots, and the xylem sap was collected for 30 min with a micropipette. The silicon concentration in the xylem sap and external solution before and after the 8 h culture was determined immediately. The root weight and water loss were also recorded.

Measurement of ²⁹Si-NMR

Seedlings (40 d old) were cultured in the nutrient solution containing 0.5 mm ²⁹Si-enriched silicon as silicic acid. ²⁹Si-enriched ²⁹SiO₂ (98.7% enrichment) was purchased from Shoko (Tokyo). A portion (0.1 g) of ²⁹SiO₂ was dissolved in 2 n NaOH with a microwave, then diluted with distilled water and passed through cation-exchange resin as described above before use. After the culture for 4 h, xylem sap was collected as described above. The roots and leaves were frozen at −80°C for 2 h. Cell sap was obtained from the roots and leaves by squeezing the frozen sample with a plastic syringe before it was completely thawed at room temperature. Xylem sap, root-cell sap, and leaf-cell sap collected were immediately subjected to ²⁹Si-NMR measurement using 5-mm NMR tube.

The ²⁹Si-NMR spectra were obtained at 99.36 MHz (DMX-500 spectrometer; Bruker BioSpin GMBH, Karlsruhe, Germany). The observation parameters for ²⁹Si-NMR were as follows: frequency range, 27.8 kHz; data point, 64 k; acquisition time, 1.18 s; relaxation delay, 3.2 s; number of scans, 512. The deuterated chloroform solution containing 1% tetramethylsilane was used as an external reference for calibration of the chemical shift (0 ppm). The ²⁹Si-NMR spectrum of a 2 mm ²⁹Si-enriched silicic acid was also recorded.

Determination of Silicon Concentration

The concentration of silicon in the symplastic and apoplastic solutions and in xylem sap was determined by the colorimetric molybdenum blue method immediately after the collection. The method is briefly described as follows. The solution of 0.01 mL of sample was added to 1.15 mL of water, followed by 0.6 mL of 0.26 n HCl, 0.08 mL of 10% (NH₄)₆Mo₇O₂₄, 0.08 mL of 20% tartaric acid, and 0.08 mL of reducing agent. The reducing agent was prepared by dissolving 1 g Na₂SO₃, 0.5 g 1-amino-2-naphthol-4-sulfonic acid, and 30 g NaHSO₃ in 200 mL of water. After 1 h, the absorbance was measured with a spectrophotometer (Jasco, Tokyo) at 600 nm.

Linkage Mapping of Lsi1

Seedlings (4 weeks old) of F₂ populations derived from the mutant (lsi1) and Kasalath were used for silicon uptake determination. A total of 105 F₂ seedlings each in a 50-mL plastic bottle were cultured in the nutrient solution containing 0.6 mm silicon for 6 h at 25°C. The concentration of silicon in the nutrient solution before and after the culture was determined as described above. Water loss and root fresh weights were also recorded.

For extraction of DNA, leaf was sampled after the uptake experiment and stored at −80°C until use. DNA was extracted according to Komatsuda et al. (1998). Briefly, leaf (approximately 50 mg) was placed in a tube and homogenized in a buffer solution (300 μL) containing 100 mm Tris-HCl, 50 mm EDTA, 500 mm NaCl, and 10 mm 2-mercaptoethanol at pH 8.0, followed by adding 40 μL of 10% SDS and heated at 65°C for 15 min with occasional shaking. Then, 100 μL of 5 m potassium acetate was added, vortexed briefly, and cooled in ice for 15 to 30 min. After addition of 500 μL of phenol/chloroform and vortexing, the tubes were centrifuged at 12,000 rpm for 5 min at 4°C. The supernatant (350 μL) was transferred to a new tube and mixed with an equal amount of isopropanol and mixed well. The tubes were centrifuged at 12,000 rpm for 30 s, and the supernatant was discarded. The pellet was washed with 70% ethanol and centrifuged again. Finally, the pellet was dissolved in 100 μL of 0.1× Tris-EDTA.

To identify the molecular markers linked with Lsi1, we performed bulk segregant analysis by pooling equal amounts of DNA from 10 low silicon uptake plants and 10 high silicon uptake plants from the F₂ population, based on the result of the silicon uptake experiment described above. A total of 128 EST-based PCR markers scattered on whole chromosomes were selected to examine polymorphism between the low silicon uptake and high silicon uptake bulks as well as Kasalath and the mutant. The details of these markers for bulked segregant analysis are available at http://rgp.dna.affrc.go.jp.

The chromosomal location of Lsi1 was confirmed by investigating the individual low silicon uptake line by using flanking markers (C53493–C63223). For further mapping of the Lsi1 gene, EST-based PCR markers and microsatellite markers on chromosome 2 were used (McCouch et al., 2002; Wu et al., 2002). Finally, the genotypes of 105 F₂ plants were investigated using RM5303, RM5631, and E60168.

PCR amplifications were performed in 10-μL aliquots, each containing 10 ng μL⁻¹ DNA, 0.2 μm of both primers, 200 μm dNTPs, 1.5 mm MgCl₂, and 1× Ampli Tag gold. Thermocycling programs used were a preliminary step of 2 min at 94°C, 30 cycles of 30 s at 94°C, 1 min at 60°C, 1 min at 72°C, with a final step of 5 min at 72°C for EST-based PCR markers and a preliminary step of 2 min at 94°C, 35 cycles of 1 min at 94°C, 1 min at 55°C, and 2 min at 72°C with a final step of 5 min at 72°C for microsatellite markers. The PCR products were analyzed on 3% agarose gels stained with ethidium bromide. The segregation data of silicon uptake and polymorphism were used for the linkage analysis performed by MAPMARKER/EXP version 3.0b (Lander et al., 1987).