Evidence for ‘silicon’ within the cell walls of suspension-cultured rice cells



  • Despite the ubiquity and beneficial role of silicon (Si) in plant biology, structural and chemical mechanisms operating at the single-cell level have not been extensively studied.
  • To obtain insights regarding the effect of Si on individual cells, we cultivated suspended rice (Oryza sativa) cells in the absence and presence of Si and analyzed single cells using a combination of physical techniques including atomic force microscopy (AFM).
  • Si is naturally present as a constituent of the cell walls, where it is firmly bound to the cell wall matrix rather than occurring within intra- or extracellular silica deposition, as determined by using inductively coupled plasma mass spectrometry (ICP-MS) and X-ray photoelectron spectroscopy (XPS). This species of Si, linked with the cell wall matrix, improves the structural stability of cell walls during their expansion and subsequent cell division. Maintaining cell shape is thereby enhanced, which may be crucial for the function and survival of cells.
  • This study provides further evidence that organosilicon is present in plant cell walls, which broadens our understanding of the chemical nature of ‘anomalous Si’ in plant biology.


Silicon (Si) is beneficial to plant growth (Richmond & Sussman, 2003), and intra- or extracellular silica in plants is useful for improving mechanical strength and alleviating biotic and abiotic stress (Wang et al., 2000, 2005; Yeo et al., 2002; Gao et al., 2004; Fauteux et al., 2006; Currie & Perry, 2007; Epstein, 2009). Rice (Oryza sativa) plants have a feature whereby they accumulate Si at proportions of up to 10% in dry weight of shoots for maintaining high and sustainable production (Savant et al., 1997), and it has been regarded as the prototype for controlled formation of silica in higher plants (Ma & Yamaji, 2006; Neethirajan et al., 2009). Si uptake by rice roots involves a type of transporter (Ma et al., 2006), and more than one transporter has been identified and is likely involved in the Si uptake system in plants. Following its uptake, Si is translocated to shoots as monomeric silicic acid and disilicic acid (Casey et al., 2003) and deposited in specialized cells (Kaufman et al., 1969, 1985) or on cell walls, forming amorphous silica-cuticle/cellulose double layers in the epidermis of leaves and on the surfaces of stems and hulls (Kaufman et al., 1985; Sangster et al., 2001).

Rice takes up silicic acid at a relatively high rate (Ma et al., 2006), raising the possibility that organosilicate complexes may subsequently be formed in the roots and shoots. Despite several attempts to look for evidence of organosilicate complexes (Kinrade et al., 2001a,b; Benner et al., 2003), most investigations to determine the protective roles played by silica in leaves have been carried out at the level of the whole plant, tissue or organ, and advances in the detection of roles of Si at the single-cell level remain limited (Prabagar et al., 2011). We used suspended rice single cells to perform a comprehensive investigation of the existence of Si at cell surfaces, demonstrating that a form of Si covalently bound to organic moieties in cell walls is present. This finding reveals a structural role of Si in individual rice cells, contributing to the stability of cell walls during their expansion and subsequent cell division. Cell shape is thereby maintained, which may be crucial for the function and survival of cells.

Materials and Methods

Cell culture

Suspension-cultured cell lines of rice (Oryza sativa L. cv Zhonghua 11) were established following the processes of Chu et al. (1975) and Thomas et al. (1989a,b): mature seeds were dehusked, sterilized with 75% ethanol for 1 min, 0.1% mercury chloride for 10 min and washed with sterilized water five times. Sterilized seeds were incubated at 28°C in the dark for 1 month in a modified N6 medium and subcultured three times in subculture medium for callus formation. The calli (50–100) were transferred to 125-ml plastic Erlenmeyer flasks containing 40 ml liquid AA medium in the presence of 2, 4-D at 1 mg l−1 and cultivated on a rotary shaker (110 rpm) at 28°C in the dark. The suspension cells were subcultured at 5-d intervals for 1–3 months by replacing old liquid medium using newly prepared nutrient solution every 5 d to supplement nutrients and the Si. To apply different culture conditions, rice cells were cultivated in the absence (−Si) or presence (+Si) of silicic acid at various concentrations in the liquid medium (pH 5.6). Specifically, freshly made silicate was added as a basic Na2SiO3 solution (Casey et al., 2003) for each subculture use. The solutions were then acidified to pH 5.6 with 1.0 M HCl, generating the corresponding concentration of NaCl which was supplemented in the control solutions. Only then were the other constituents added to dilute culture solutions to the final Si concentrations ranging from 0.15 to 2.0 mM. In order to produce constant Si percentages during the 5 d culture, only a limited number of cells remained in the plastic flask to minimize depletion of the solution (< 2% Si was depleted after the 5 d period) according to an established method (Chu et al., 1975). The Si content of the liquid medium was determined to be 0–2 mg kg−1 using an inductively coupled plasma mass spectrometer (ICP-MS, Elan DRC-e; Perkin-Elmer, Concord, Ontario, Canada). Cell samples were kept sealed during most manipulations to avoid Si contamination from dust. All solutions were prepared using ultra-high purity water, 18 MΩ-cm, from a two-step purification treatment including double distillation (YaR, SZ-93, Shanghai, China) and deionization (Milli-Q, Billerica, MA, USA). Total Si content of the ultra-high purity water was determined to be close to zero using ICP-MS. All chemicals were high purity (Aldrich-Sigma, unless otherwise indicated). On visualization of cells, we used a digital CCD camera (Nikon DS-Fi1, Tokyo, Japan) coupled to a light microscope (Olympus BX51, Tokyo, Japan), and measured cell sizes using NIS-Elements F 3.0 software. For the solution compositions of different culture media, see Supporting Information Table S1.

Protoplast culture

Rice protoplasts were cultured from rice suspension cells that were added to a 10-ml enzyme solution containing 2.5% cellulase and 1% macerozyme R-10 and 0.4 M mannitol (Yamada et al., 1986). For a thorough description of culture methods for protoplasts of rice, see Methods S1.

Cell wall isolation

Cells were collected and washed five times by repeated centrifugation at 10°C, 1700 g for 5 min and re-suspended in ultra-high purity water (the final concentration of Si was < 2 mg kg−1). Then they were ground to a fine powder in liquid N2 with a teflon mortar and homogenized in 1% sodium dodecyl sulfate (SDS) buffer (50 μM Tris-HCl, 1% SDS, pH 7.2). The samples were kept in an 80°C water bath for 15 min and centrifuged at 1700 g for 5 min. The supernatant was removed with an aspirator and washed five times with hot water to remove all SDS. The residual cell walls were washed three times with warm 80% ethanol at 65°C and two times with warm 50% ethanol at 65°C, and then washed three times with ultra-high purity water at 65°C. Finally, the samples of cell walls were lyophilized (Carpita et al., 1985; Thomas et al., 1989a,b).

Cell viability assay

The viability of suspension cells was determined by fluorescein diacetate (FDA)-propidium iodide (PI) staining. FDA can pass through living cell membranes and is converted to a green fluorescent dye by intracellular esterases found exclusively in living cells. By contrast, PI passes through dead cell membranes and generates a red fluorescence by forming a PI–nucleic acid conjugate. The FDA-PI solution was made fresh each time by adding 1 μl PI and 1 μl FDA to 98 ml ultra-high purity water. Ten microliters of FDA-PI solution was added to 90 μl of solution containing cells and incubated for 2 min at room temperature. Stained cells with FDA-PI were observed with a light/fluorescence microscope (Olympus BX51). The experiment was conducted three times.

Total Si concentration measurement

Three millilitres of concentrated HNO3 and 2 ml H2O2 were added to 0.05 g of the samples (lyophilized cells or isolated cell walls or protoplasts) in a teflon bottle and left overnight. The teflon bottle was put into a high-temperature digestion tank to prevent nitric acid volatilization. The digestion tank was then placed in an oven at a temperature of 80°C for 3 h followed by 160°C for 5 h. The digestion tank was heated on a hot plate at a temperature of 150°C, and evaporated to dryness. The dried material was made up to a volume of 10 ml with 2% HNO3. An ICP-MS was used for Si concentration measurement. All experiments were repeated three times, and their mean values ± SD are presented.

Stability of Si binding in cell walls

Experiments to liberate total Si using 0.1 M KOH, 0.1 M HNO3 or enzymes were conducted. Isolated cell walls were weighed and placed in a teflon bottle. Three millilitres of 0.1 M KOH or 0.1 M HNO3 was added to the sample of cell walls at room temperature (24°C) for 2 and 24 h, respectively. One millilitre of supernatant was taken after 2 or 24 h treatment for determination of free, unbound Si using ICP-MS. For enzymatic hydrolysis on bound Si, 10 mg isolated cell walls were incubated in a 5-ml 2-(N-morpholino) ethanesulfonic acid (MES) buffer solution (10 mM, pH 5.6) in the presence of 1 mM MgCl2 and either 0.5% cellulase, 1% hemicellulase or 0.1% pectinase. After 6 h of digestion at 25°C, the suspension was centrifuged at 1062 g and the supernatant was discarded. The residue was washed once with ultra-high purity water and then centrifuged. The undissolved residual in weak acid or alkaline or enzyme solution was digested using 3 ml concentrated HNO3 and 2 ml H2O2 following the procedure for the determination of total Si using ICP-MS.

X-ray photoelectron spectroscopy (XPS) studies

XPS has proved to be an ideal tool for the characterization of the nanometric layer (c. 200 nm of the wall thickness) as well as to probe the elemental depth distribution and their bonding states. Before spectroscopy measurement, all samples were dried under vacuum for at least 8 h. Lyophilized cells and the walls were ground using a teflon mortar. A powder sample was placed on an aluminum (Al) platform and homogenized with a spatula in order to obtain a relatively smooth surface. Field-grown leaves of rice were cut into small fragments and placed on the Al platform for XPS measurements (VG multilab 2000 equipment ThermoVG scientific, East Grinstead, West Sussex, UK) using the Al Kα X-ray line of 1486.6 eV excitation energy at 300 W. To correct for sample charging, high-resolution spectra were used as a reference by setting the C 1s hydrocarbon peak to 284.6 eV. The background was linearly subtracted. Data analysis was performed with the Thermal Advantage software (http://www.tainstruments.com). The ratios of atomic concentrations were calculated using the peak areas normalized on the basis of acquisition parameters and sensitivity factors proposed by the manufacturer. XPS experiments were repeated at least five times to ensure the reproducibility of results.

Mechanical properties of cells by atomic force microscopy (AFM)

Using a pipette, suspension cells were placed on an 18 × 18 mm glass slide that was treated with 0.02% Poly-L-Lysine (PLL) for 15 min. The sample was washed carefully with ultra-high purity water to remove any unfixed cells. AFM images were obtained in tapping mode at room temperature (25°C) in water using an Agilent 5500 atomic force microscope (AFM) (Si3N4 cantilevers of 0.058 nN nm−1 spring constants). The imaging force was kept as low as possible (c. 250 pN) to minimize sample damage. All measurements for cell mechanical properties were repeated (> 10 cells for each measurement of three independent experiments). All data with their mean values ± SD of at least three independent sets of experiments are presented. Statistical differences were determined with a Student's t (ST) test.


The concentration of Si in the whole cells of rice, that were incubated for 3 months in the presence of silicic acid ranging from 0.15 to 2.0 mM, increased with the concentration of silicic acid in the culture medium and reached plateaus of uptake, at c. 113 mg kg−1 DW of cells (Fig. 1a). A trace amount of Si as determined by energy dispersive X-ray spectroscopy (EDX) was evident at the cell wall surface after cultivation for 3 months (Figs S1, S2). In addition, Si concentrations in the isolated cell walls from suspension cells were usually 2.5–3 times higher than those in whole cells under the cultivated conditions (Fig. 1a). For individual cells, the cell wall makes up c. 30% of the dry weight of the cell, indicating that most Si accumulated in the cell walls. Almost no Si was detected inside the cells by determining the concentration of Si in rice protoplasts (Figs S3, S4). In the absence of silicic acid in the culture medium, the cells accumulated almost undetectable amounts of Si (< 10 mg kg−1) in their walls. Only a very small quantity of Si (< 10% of total Si) was released by extended treatments by KOH or HNO3 from 2 to 24 h (Fig. 1b). After enzymatic hydrolysis of cellulose, hemicellulose or pectin by corresponding enzymes only 7.4%, 7.6% or 21.2% (respectively) of the Si initially bound to the wall matrix was liberated; the main portion was present as bound Si because the corresponding enzyme just breaks down the organic parts in the walls. This shows that the Si in cell walls was less readily liberated by weak alkali or acid or even enzyme and that it occurred firmly bound to the polysaccharide matrix of cell walls, not as free silicate/silicic acid or silica.

Figure 1.

(a) Uptake and accumulation of silicon in suspension-cultured cells of rice (Oryza sativa) in the presence of 0.15, 1.0 or 2.0 mM silicic acid for 3 months after transplanting to solution medium from solid culture medium. Whole cells, black squares; isolated cell walls, red circles. (b) Stability of Si binding in cell walls. Values are mean ± SD of three independent sets of experiments.

The Si2p core-level XPS spectra of suspension cells and isolated walls cultivated in the presence of 1 mM silicic acid showed an obvious peak at 101.3 ± 0.2 eV (Figs 2a, S5), which is neither the elemental Si (Si0 at c. 99.2 eV), silicate/(poly)silicic acid (c. 102.1 eV) nor silica (fully oxidized silicon Si4+ at c. 103.2 eV) (Tesson et al., 2009). For both the cell and wall samples cultivated in the absence of 1 mM silicic acid, the concentration of Si was not measurable, that is, it was below the detection limit of XPS for the experimental conditions used. However, for field-grown leaves of rice, the Si2p peak was decomposed into three components with the same FWHM (full width at half maximum of 2 eV). A peak at c. 101.3 eV was also observed although its content (5.1 in at%) was much lower than the components at 103.4 (76.6 in at%, silica) and 102.1 eV (18.3 in at%, silicate or (oligo/poly)silicic acid), respectively (Fig. 2b). Despite a considerably smaller amount of this unknown component at 101.3 eV, it is somewhat hidden because this form is buried under relatively large amounts of silica and silicate formed at the leaf surfaces, suggesting that this new component may have previously been overlooked because of its scarcity.

Figure 2.

X-ray photoelectron spectroscopy (XPS) characterization of the chemical composition for (a) Si2p, (c) C1s and (d) O1s of the suspension cell surfaces and for (b) Si2p of leaves of rice (Oryza sativa) (as a control). Suspension cells were cultivated at 1 mM silicic acid for 3 months and leaves were grown in the field for over 3 months. Si2p intensity for cells and the walls in the absence of 1 mM silicic acid was not measureable or below the detection limit for the experimental conditions used.

The Si2p peak at 100.1–101.3 eV has been thought to be due to a Si-O-C or O-Si-C chemical environment from simple chemical systems such as Si-organic monolayers (Boukherroub et al., 2000). Specifically, 100.9 and 101.6 eV (Si-O 2p3/2 and Si-O 2p1/2 respectively) indicate covalent bonding of Si with a long-chain aldehyde (Effenberger et al., 1998). The peak positions of XPS shift to the higher-energy from 100.1 to 101.3 eV, depending on the average coordination number of O atoms with Si atoms in Si–O–C bonds by organic ligands (Nakayama & Hata, 2006). At present we have no direct evidence for the nature of the interaction between Si and the exact wall macromolecules (either in the form of polysaccharides or glycoproteins) identified at the molecular level, and thus it is difficult to directly relate the XPS signal arising from the Si-O-C or O-Si-C linkage to an exact organic ligand in the walls. However, we suggest, based on our present data, that a similar chemical interaction may occur between Si-OH and the surfaces of organic polysaccharides, resulting in ‘-O-Si-C or -Si-O-C-’ bonds for crosslinking the cell wall. This wall-bound form of Si was tentatively assigned to Si-O-C because the possibility of the chemical reaction for the formation of the Si-O-C is greater than that of the Si-C. The component of silicate or silica at the cell surfaces has been ruled out based on the XPS data and the assessment of stability of Si binding in cell walls in a weak acid or alkaline solution. The exact Si ligands of the organic compounds need further detailed investigation. The cumulative results represent direct evidence that Si is naturally present as a constituent of the cell walls, where it is firmly bound to the polysaccharide matrix (possibly to pectin or hemicellulose) rather than occurring within intra- or extra-cellular silica deposits.

The carbon region (C 1s) consists of four components (Fig. 2c). The first component at 283.3 eV (referred to as C−1) has been assigned to the Si-C bond in a simple chemical system (Condorelli et al., 2004). The content of this component was relatively low compared to the other three components. The second at 284.6 eV suggested that the carbon only bound to carbon and hydrogen (C-(C, H) (referred to as C0) and represented aliphatic and aromatic hydrocarbons. The third one with oxidized carbon species was centered at 286.2 eV (referred to as C+1), and it can be attributed to carbon singly bound to oxygen or nitrogen, C-(O,N), including ether, alcohol, amine and amide (Dufrene et al., 1997). It has been reported that this feature could arise from polysaccharides and sugar moieties and can be attributed to carbon bonded to one oxygen for the formation of the Si-O-C (Beamson & Briggs, 1992; Wallart et al., 2005; Condorelli et al., 2006). The last component at 287.9 eV was due to carbon making two single bonds (O-C-O, referred to as C+2) or one double bond with oxygen (C=O), attributable to acetal, hemiacetal, amide, carbonyl, carboxylate and ester (Tesson et al., 2009).

Similar to the C 1s peaks, the O 1s peak was decomposed into three overlapping signals (Fig. 2d). The first one at 532.6 eV was assigned to oxygen making single bonds with hydrogen or carbon (C-OH of acetal, alcohol and carboxyl, C-O-C of acetal and hemiacetal). The second one c. 531.3 eV was attributed to oxygen making a double bond with carbon (O=C), including carboxylic acid, carboxylate, ester, carboxyl or amide. The third one near 530.6 eV was assigned to the O-C bond in the Si-O-C component (Kitoh et al., 1996). Moreover, the O 1s at 531.3 and 532.6 eV have also been attributed to oxygen in Si-O bonds (Condorelli et al., 2004; Jedlicka et al., 2007; Tesson et al., 2009).

It is relatively difficult to detect this unknown wall-bound form of Si using FTIR or Raman spectroscopy due to its trace amounts which are below the detection limits of both techniques. Furthermore, the components of cellulose, hemicellulose or pectins with complex stoichiometry are partly overlapped with the components of Si-containing species (Currie & Perry, 2009) in the FTIR/Raman range of 400–1400 cm−1 (Bellamy, 1975; Falcone et al., 2010). Therefore, the spectral evidence for bonding between Si and carbohydrates to form organosilicon compounds was not conclusive. It may be possible to further characterize and provide support for strong bonding between the wall polysaccharides and Si using isolated/concentrated wall components such as cellulose, hemicellulose or pectin through spectroscopic analyses. Gierlinger et al. have reported that a Raman peak near 956 cm−1 was possibly attributed to a covalent bonding of Si to carbohydrates (the SiOC form) in Equisetum hyemale (Gierlinger et al., 2008). However, due to the uncertainty of the chemistry of the pectic and hemicellulose components, their spectral contributions cannot be completely excluded (Gierlinger et al., 2008). Currently, it is not yet known how to explain exactly the chemical nature of the XPS peak at c. 101.3 eV because no XPS data are reported in the literature for such complex compounds. The complex interactions between the wall polysaccharides/glycoproteins/phenols and Si to form an organosilicon component that differs from silica or silicate or polysilicic acid deserves a separate study.

In the presence of silicic acid, > 80% of the cells were viable (Fig. 3a,c), whereas c. 70% of the cells were dead in its absence (Figs 3b,c, S6). These living cells cultivated with silicic acid underwent normal division (from the exponential growth stage to the stationary phase) to create a population of cells with the appropriate size range and regular spherical shape (20 ± 5 μm) (Fig. 3a). With an increase in culture time to > 100 d, control cells (–Si) underwent significant elongation without division (Fig. 4d–g), regardless of whether they were living (Fig. 4d,e) or dead (Fig. 4f,g), with a statistical difference in length to width of the cells at P ≤ 0.05. By contrast, cells cultured with Si (+Si) retained their normal spherical shape, and only minor changes in cell shape appeared even after a longer period of suspension culture (Fig. 4a–c,h). Collectively, these results demonstrate that incorporating Si into cell walls endows cells with higher viability/better growth and stability against division-correlated transformations in cell shape. At the level of the whole plant and tissue, it has also been shown that Si accumulated in the cell wall of rice seedlings in a low-silica rice mutant, lsi1, thereby maintaining leaf tissue rigidity and enhancing growth in the absence of silica bodies formed in the motor cells (Isa et al., 2010). This suggests that complexation of Si with pectins and/or hemicelluloses may play an important role in increasing wall rigidity and tissue growth that is different from the function of silica bodies in rice leaves (Isa et al., 2010). During the division stage, cellulose microfibrils with hemicelluloses and pectins constrain turgor-driven cell expansion in one preferential direction. They control the shape and ulitmately viability of the plant cells themselves (Höfte, 2001). A very recent result showed that a specifically patterned and integrated cell wall is a determinant of plant cell shape (Oda & Fukuda, 2012).

Figure 3.

Viability of rice (Oryza sativa) suspension cells cultured for 3 months in the absence and presence of 1 mM silicic acid using live-dead viability probes. (a) Optical image (left) identifies normal and mitotic cells (division shown by an arrow) cultivated with silicic acid, and fluorescence image (right) of suspension cells shows those that remained viable (green). (b) Optical (left) and fluorescence image (right) of suspension cells cultivated without silicic acid show cells that died (red). (c) Cells cultivated without silicic acid (−Si) show a very low viability at c. 30%, whereas cells with silicic acid (+Si) preserve cell viability at >80%. Data are means ± SD (n > 200, cells) for three independent observations. Different letters (a, b in (c)) show significant difference at P ≤ 0.05.

Figure 4.

Changes in rice (Oryza sativa) cell shape after a long-term culture (over 100 d) in the presence (a–c) and absence (d–g) of 1 mM silicic acid. (a, d, f) Optical, (b, e, g) fluorescence and (c) scanning electron microscopy (SEM) images. (h) Changes in cell shape were defined as changes in the ratio of cell length over width; data are means ± SD (n = 170, cells) for three independent observations. Different letters (a, b in (h)) show significant difference at P ≤ 0.05. The arrows in (c) and (f) indicate the dividing locations of cells.

A possible linking of Si with the wall polysaccharides/glycoproteins may influence the mechanical strength and structural stability of the cell wall. We directly examined the local cellular nanomechanical properties using in situ atomic force microscopy (AFM). The sensitivity and ability to apply only minimal forces on a single, living cell makes the AFM a useful nondestructive tool to study cellular nanomechanics without the use of chemical immobilizers under physiological conditions in fluid (Pelling et al., 2004). Force–displacement curves were measured on living cells which were not dividing rapidly and were near stationary phase. The cells cultivated with 1 mM Si displayed a considerably higher kcell than those cultivated without 1 mM Si (Fig. 5; local nanomechanical spring constants (kcell) can be determined from the slope of the linear portion of the curves), showing a statistical difference at  ≤ 0.05. The +Si cells exhibited local kcell values of 0.0435 ± 0.015 N m−1 (n = 10), whereas corresponding values for –Si cells were much lower (0.024 ± 0.01 N m−1 (n = 12)). AFM results may suggest that the Si improved wall stability through enhancing organized assembly in relation to cell function (i.e. cell division), which accounts for the local stiffness of cell walls (Touhami et al., 2003). The value of kcell also reflects the internal pressure generating a certain degree of turgor, which is balanced and maintained by an equal and opposite inward force exerted by the cell wall against the plasma membrane (Vella et al., 2012). Furthermore, Arnoldi et al. theoretically demonstrated that the kcell of bacteria cells increases with its turgor pressure, and the mechanical behavior of bacteria under high loading forces, typically 0.5 nN, could then be correlated with their turgor pressure from the spring constant (Arnoldi et al., 2000). Thus, the mechanical surface properties could be interpreted in terms of the presence of a specific and organized polymeric structure of the cell wall (Gaboriaud et al., 2005).

Figure 5.

Representative force–distance curves for rice (Oryza sativa) cells cultivated for 3 months in the absence (−Si) and presence (+Si) of silicic acid by monitoring the atomic force microscopy (AFM) deflection of the cantilever. Both extension (approach) and retraction lines are usually overlapping, thus retraction lines were not drawn for clarity. More than 10 individual cells were examined for each measurement of three independent AFM experiments.


Growing plant cells are shaped by an extensible wall that is a complex organization of cellulose microfibrils bonded noncovalently to a matrix of hemicelluloses, pectins and structural proteins (Cosgrove, 1997). Although an individual plant cell may expand its volume by nearly 20 000 times, its cell wall must maintain a uniform thickness and structure to prevent hemorrhaging of the cell through local defects (Perrin et al., 1999). The wall must grow proportionately, while remaining strong enough to counter up to 5 atmospheres of pressure (turgor pressure) from inside the cell (Strauss, 1998). Newly synthesized wall materials are deposited in a relaxed state and become load bearing only after they become integrated into the wall network and elastically stretched by wall enlargement so that they begin to resist further stretching. At this point, the wall formation and its organized assembly must precede continued cell division (Szymanski & Cosgrove, 2009). Otherwise, cells undergo maximal elongation along the long axis of the cell and cannot divide (Marga et al., 2005).

The cell walls of the monocots including rice and other Poales possess a different kind of primary wall, a ‘Type II’ wall, which is different from the ‘Type I’ wall of the dicot plants. ‘Type I’ wall is composed of a cellulose-xyloglucan framework embedded in a pectin gel (Carpita, 1996). Amongst inorganic elements in cell walls, both boron (B) and Si maintain the physical strength of wall structures (Miwa et al., 2009). Boron in dicots mainly contributes to retaining the integrity of cell walls, where it covalently crosslinks the pectic polysaccharide rhamonogalacturonan II (RG-II); deficiency in B will result in structurally abnormal walls and influence wall expansion (O'Neil et al., 2001). Silicon may be required for cross-linking of cell wall components in monocots. Some recent studies have shown that both Equisetum and the Poales (including rice plants) are heavily silicified in their epidermis (Currie & Perry, 2007) which are rich in mixed-linkage (1→3,1→4)-β-d-glucan (MLG) in cell walls, raising the possibility that MLG serves as a key ligand that plays an important role in the poorly understood mechanisms of cell wall silicification (Fry et al., 2008; Sørensen et al., 2008). Currie and Perry provided chemical evidence for intrinsic ‘Si’ within Equisetum cell walls (albeit not rice) and also proposed the role of the ‘Si’ in some form to rigidify the cell wall by crosslinking, although they were not able to define the precise chemical connection of the Si to the plant cell wall polymers (Currie & Perry, 2009). Moreover, large amounts of bound Si were found to be present in citrus pectin (2580 ppm) that was resistant to dialysis, 8 M urea and hydrolysis by weak acids and alkalis. Also, alginic acid from horsetail kelp contained Si (451 ppm) in a bound form (Schwarz, 1973). Very recently, callose and other similar carbohydrates have been suggested to serve as key molecules in biosilicification in horsetail (Law & Exley, 2011).

There is increasing awareness that silicate–sugar complexes can be spontaneously formed by the formose reaction (Lambert et al., 2010). The rationale is similar to that proposed for a borate-mediated formose reaction (Ricardo et al., 2004), with the advantage of the much wider availability of silicate minerals and hence readily available silicate ions. The addition of aliphatic polyols (sugar-like molecules) to aqueous silicate solutions has been shown to yield high concentrations of stable polyolate complexes containing five- or six-coordinated Si (Kinrade et al., 1999). Benner et al. (2003) have demonstrated a crystallized diolatosilicate by hydrogen-bonded sugar-alcohol trimers as hexadentate silicon chelators in aqueous solution, and suggested that the Si-O-C linkage is stable towards hydrolysis in special diolato and alditolato ligands. This study shows the significance of specific patterns of stabilizing secondary interactions which have their origin in the unique polyfunctionality of the carbohydrates. Silicon complexation by carbohydrates depends on the discovery of ligands that combine the principles: the stability range of complexes around neutral pH may be broadened by using ligands that are free of strain, that give complexes that can be further stabilized by secondary interactions (Benner et al., 2003).

Many naturally occurring organic molecules in cell wall surfaces contain groups structurally similar to the simple polyols as ligands. Limited evidence exists on how Si is isolated, transported and deposited or incorporated by plants, although complexing by sugars has long been suspected (Iler, 1979; Frausto da Silva & Williams, 1991; Birchall, 1995). Kinrade et al.'s (1999) observation of the ease by which Si binds to polyols (cis-diols in sugars) to form such hypervalent silicate complexes would provide fundamental solution chemistry of Si in relation to biochemical function in cell walls that contain functional groups may indeed play a vital role in isolating silica and other Si species containing Si-O-C bonds in nature.


Combining the prior evidence for Si within Equisetum cell walls (Currie & Perry, 2009), Si in some form in the cell walls of suspension-cultured rice, like B, could crosslink plant cell wall polysaccharides or other matrix molecules. Si is naturally present as a constituent of the cell walls of individual rice cells rather than occurring within intra- or extracellular silica deposition. This wall-bound form of Si exhibits a significance in maintaining cellular integrity and structure of the walls, thereby highlighting the biochemical and structural role of Si at the trace level in the preservation of cell shape and mechanical properties to ensure subsequent wall expansion and cell division, which may be crucial for the function and survival of cells. The ability of Si to assemble a cell wall surrounding an individual cell may provide the foundation for understanding the nature of ‘anomalous Si’ in plant biology.


This work was supported by the National Natural Science Foundation of China (Grant No. 31172027) (to L.J.W.); a startup grant from the Huazhong Agricultural University (52204-09008) (to L.J.W.); and the Fundamental Research Funds for the Central Universities (2011PY150) (to F.S.X.). We thank Dr. Richard Gordon for careful reading of the manuscript, and Dr. Richard Bélanger for constructive suggestions for the protoplast culture of rice. Lijun Wang also dedicates this publication to the memory of Prof. Li Min for her lasting contributions to the advancement of understanding of the role of silicon in turfgrass growth.