The relationship between aluminium and silicon accumulation in leaves of Faramea marginata (Rubiaceae)


  • Ricardo M. Britez,

    1. Department of Soils, Federal University of Paraná, C. Postal 2959, 89.035-050, Curitiba/PR, Brazil;
    Search for more papers by this author
  • Toshihiro Watanabe,

    1. Graduate School of Agriculture, Hokkaido University, Kita 9, Nishi 9, Kitaku, Sapporo 060-8589, Japan;
    2. Present address: Japan International Research Center for Agricultural Sciences, 1–1 Ohwashi, Tsukuba 305-8686, Japan;
    Search for more papers by this author
  • Steven Jansen,

    1. Laboratory of Plant Systematics, Institute of Botany and Microbiology, K. U. Leuven, Kasteelpark Arenberg 31, B−3001 Leuven, Belgium;
    2. Present address: Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3DS, UK
    Search for more papers by this author
  • Carlos B. Reissmann,

    1. Department of Soils, Federal University of Paraná, C. Postal 2959, 89.035-050, Curitiba/PR, Brazil;
    Search for more papers by this author
  • Mitsuru Osaki

    Corresponding author
    1. Graduate School of Agriculture, Hokkaido University, Kita 9, Nishi 9, Kitaku, Sapporo 060-8589, Japan;
      Author for correspondence:Mitsuru Osaki Tel: +81 11 7064170 Fax: +81 11 7064170 Email:
    Search for more papers by this author

Author for correspondence:Mitsuru Osaki Tel: +81 11 7064170 Fax: +81 11 7064170 Email:


  • •   The relationship between high aluminium (Al) and silicon (Si) levels in Faramea marginata was investigated and the hypothesis tested that the coexisting accumulation of these elements is associated.
  • •   Mineral concentrations of Al, Si and calcium (Ca) were analysed in 30 samples by atomic absorption spectrophotometry and the spectrophotometric molybdenum blue method. Extraction patterns of Al and Si from leaves were compared with Melastoma malabathricum, rice (Oryza sativa), aluminium silicate, and silicon dioxide. The localization of Al and Si was studied using pyrocatechol violet staining of sections and fluorescent X-ray analytical microscopy.
  • •   A positive correlation occurred between the Al and Si levels and both elements showed a similar distribution in leaf and stem tissues. The Al and Si elution patterns were similar to those of aluminium silicate.
  • •   These results suggest the formation of an Al–Si complex in the shoot tissues of F. marginata, which may substantially contribute to the internal detoxification of Al.


Aluminium (Al) toxicity is a very serious problem for the cultivation of many crop plants in acid soils all over the world (Foy et al., 1978). Excess of soluble and bioavailable Al especially inhibits the root growth of Al-sensitive plants such as wheat (Triticum aestivum), maize (Zea mays) and soybean (Glycine max). Because the normal metabolic activities of root cells can be hindered by apoplastic as well as symplastic Al effects, the growth of the whole plant can be seriously reduced. Several mechanisms dealing with the toxic properties of Al have been proposed (Kochian, 1995; Matsumoto, 2000; Barceló & Poschenrieder, 2002).

The majority of plant species avoids Al stress by excluding Al from the roots and can therefore be termed Al excluders. At present, exudation of organic acids from the roots is considered as the most common Al exclusion mechanism (Ma et al., 2001a). Some organic acids (e.g. citrate) have a high chelating affinity to Al and this may substantially reduce the uptake and toxicity of Al. Not only organic acids, but also some inorganic anions, such as fluorine (F), silicon (Si) or sulfate (SO4) can make stable complexes with Al and reduce the harmful Al effects (Tanaka et al., 1987; Hodson & Evans, 1995). Among these anions, the effect of Si application on Al toxicity has been well studied. For example, in the case of maize, the concentration of Al3+ in culture solution is strongly reduced when Si is applied (Ma et al., 1997a). Silicon has also been demonstrated to inhibit Al penetration into the root cortex of Sorghum bicolor, indicating that Si makes a complex with Al in the medium and/or roots and contributes to the detoxification of Al (Hodson & Sangster, 1993).

While many plants exclude Al from the roots, some plant species accumulate huge amounts of Al in both roots and shoots (Haridasan, 1982; Geoghegan & Sprent, 1996; Osaki et al., 1998; Jansen et al., 2002a,b; Watanabe & Osaki, 2002). Generally, plant species are classified as Al accumulators if they accumulate at least 1000 mg kg−1 in their leaves (Chenery, 1948). The Al accumulators often exhibit a high Al tolerance despite the high Al concentration in their above-ground tissues (Haridasan, 1988; Osaki et al., 1997; Watanabe et al., 1997). Several researchers considered that Al-accumulating species are able to detoxify Al, for example, by binding to the cell walls, compartmentalization in the vacuole or formation of Al chelates. Examples that have demonstrated the occurrence of chelated Al forms in plant leaves are Al-catechin in the tea bush (Camellia sinensis), Al-citrate in Hydrangea macrophylla, and Al-oxalate in buckwheat (Fagopyrum esculentum) and Melastoma malabathricum (Nagata et al., 1992, 1993; Ma et al., 1997b,c; Watanabe et al., 1998a). Chelation of Al by inorganic ligands has not yet been reported in Al accumulators, although Al–F was considered as the translocation form in the tea bush (Nagata et al., 1993; Liang et al., 1996). Moreover, Masunaga et al. (1998) found a positive correlation between Al and Si levels in leaves of various Al accumulators (3000 mg Al kg−1) growing in a tropical rain forest in Indonesia, which may contribute to the alleviation of Al toxicity in several Al accumulators.

Al accumulators are not randomly distributed among the flowering plants, but frequently characterize families such as Rubiaceae, Melastomaceae or Vochysiaceae (Chenery, 1948; Haridasan, 1982; Jansen et al., 2002a,b). Especially in Rubioideae, one of the three subfamilies within Rubiaceae, strong Al accumulators have been reported (Jansen et al., 2000a,b). We have previously reported on the chemical characterization of plant species from the swamp forest in the Ilha do Mel, Brazil (Britez et al., 1997). These preliminary results indicated that Faramea marginata, a woody member of the Rubiaceae family that grows on mainly acid soils, is not only an Al accumulator, but also appears to accumulate Si in its leaves. Since Si can form stable complexes with Al, we hypothesize that an Al–Si complex is formed in F. marginata and that this may contribute to an increased Al tolerance in its above-ground tissues. Therefore, this study investigates in detail the relationship between Al and Si accumulation in the leaves of this species.

Materials and Methods

Sampling and mineral analysis of leaves

Leaves of F. marginata Cham. were sampled at the swamp forest in the Ilha do Mel, Brazil (25°30′ S; 48°20′ W, soil pH (CaCl2) 3.4, A layer). The soil was classified as a Podzol (Podosol) with sandy texture, developed on the quaternary coastal plains and was considered as relatively homogeneous. A total of 30 samples was collected from 10 plants, including 10 young, 10 intermediate and 10 mature leaves. All 30 leaf samples were dried at 80°C for 48 h. Mineral concentrations in the leaves were analysed after wet digestion with an acid mixture (HNO3–HClO4–H2SO4, 5 : 2 : 1). The concentrations of Al and Ca were determined by atomic absorption spectrophotometry (AA-6200; Shimadzu, Kyoto, Japan). The concentrations of Si were determined by atomic absorption spectrophotometry for acid-soluble Si and the spectrophotometric molybdenum blue method for acid-insoluble Si. Therefore, acid-insoluble Si was isolated from the digested solution (0.9 n H2SO4) by filtering through a filter paper (No. 5C; Advantec Toyo, Tokyo, Japan) and extracted with 20 ml of 5% NaOH at 80°C for 4 h. The extract was made up to 250 ml. An adequate amount of the extract (< 0.06 mg Si) was diluted with deionized water to approx. 15 ml and neutralized with 0.5 n HCl. Then 0.6 ml of 0.5 n HCl and 1.25 ml of 10% ammonium molybdate solution were added to the sample solution and allowed to stand for 3 min. Finally, 2.5 ml of 17% sodium sulphite solution was added to the sample and made up to 25 ml. The absorbance at 650 nm was measured after 10 min.

Extraction patterns of Al and Si from leaves

Extraction with ethanol Mature leaves of eight plants of F. marginata were lyophilized. The leaves were well ground, extracted with ethanol (sample–ethanol, 1 : 100; pH was not adjusted or adjusted to 1.0), and shaken for 1 h. The extract was filtered with filter paper (No. 5C; Advantec) and the concentration of Al and Si in the extract was determined by atomic absorption spectrophotometry. For comparison, the leaves of M. malabathricum L., which is an Al accumulator but not a Si accumulator, were similarly extracted with ethanol (Watanabe et al., 1997). Each extraction was repeated three times.

Extraction with water Lyophilized mature leaves were extracted with deionized water with different pH (sample–water, 1 : 100), and shaken for 1 h. The pH of deionized water was adjusted to 1, 4, and 12 with NaOH or HCl, or not adjusted. The extract was filtered with a membrane filter (pore size 0.45 µm), and the concentration of Al and Si was determined by atomic absorption spectrophotometry. The concentration of organic acids in the extract of the leaves was also determined by capillary electrophoresis (Watanabe et al., 1998a). In addition, shoot samples of rice (O. sativa L.), which is a typical Si accumulator, and the chemical reagent of aluminium silicate (Wako Chemicals, Osaka, Japan; approximately Al2O3.3SiO2 = 282.21) and SiO2 (Wako Chemicals, Japan) were extracted with deionized water with different pH. The concentration of Al and Si was determined as described above. Samples of rice were obtained from the field at the Japan International Research Center for Agricultural Sciences (Tsukuba, Japan). Each extraction was carried out with three replicates.

Al and Si localization in the shoot

The Al localization in a fresh mature leaf and a young stem was determined by the PCV (pyrocatechol violet) staining method (Watanabe et al., 1998a). Transverse leaf and twig sections were obtained by hand-sectioning with a razor blade. The sections were placed on a slide and stained with 0.02% PCV in 2.5% hexamine–NH4OH buffer (pH 6.2) for 15 min, washed with 2.5% hexamine–NH4OH buffer (pH 6.2) and observed under a light microscope.

Fresh samples of a whole leaf and stem cross-sections were obtained from seedlings of F. marginata growing on a pot containing 12 l of soil in a greenhouse at the Hokkaido University, Japan. The samples were dried using an ironing press and the Al and Si distribution was determined by a fluorescent X-ray analytical microscope (XGT-2000W; Horiba, Kyoto, Japan). The analytical parameters were as follows: X-ray tube target = Rh, tube voltage = 15 kV, tube current = 1.0 mA.


Interaction between Al and Si concentrations in leaves

The Al and Si concentrations in all leaves of F. marginata analysed ranged from 13 000 to 23 000 mg kg−1 and from 6500 to 14 000 mg kg−1, respectively (Fig. 1). The mean values of Al and Si concentrations were 18 100 and 8480 mg kg−1, respectively. There was a significant positive correlation between Al and Si concentrations in leaves (P < 0.001, Fig. 1). The mole ratio of Si : Al was approximately 0.5, which was calculated by a slope of the regression line between Al and Si concentrations (Fig. 1a). However, considerable variation was found in the mole ratio of Si : Al. The Si concentration showed a significant positive correlation with the mole ratio of Si : Al, while Al concentrations did not (P < 0.001, Fig. 1b,c).

Figure 1.

The relationship between (a) aluminium (Al) and silicon (Si) concentrations, (b) mole ratio of Si : Al and Al concentration, and (c) mole ratio of Si : Al and Si concentration in various aged leaves of Faramea marginata growing in a Brazilian swamp forest. a, Regression equation in mmol kg−1; ***, significant at P < 0.001.

The mean values of Al, Si and Ca concentrations in young, intermediate, and mature leaves were summarized in Table 1. The concentrations of all three elements increased in mature leaves, but the differences between young and mature leaves were not very large.

Table 1.  The mean values of aluminium (Al), silicon (Si) and calcium (Ca) concentrations, and mole ratios of Si : Al, Al : Ca and Si : Ca in young, intermediate, and mature leaves of Faramea marginata.
  1. Values are means of 10 samples ± SE. The data are included in Fig. 1.

Al (mg kg−1)16 555 ± 50317 341 ± 82520 540 ± 537
Si (mg kg−1)  7 479 ± 215  7 762 ± 30910 194 ± 591
Ca (mg kg−1)  1 695 ± 71  1 951 ± 104  2 644 ± 159
Si : Al    0.44 ± 0.02    0.43 ± 0.02    0.48 ± 0.02
Al : Ca  14.65 ± 0.50  13.37 ± 0.55  11.89 ± 0.70
Si : Ca    6.35 ± 0.19    5.76 ± 0.20    5.67 ± 0.44

Extraction patterns of Al and Si from leaves

Aluminium in the leaves of F. marginata could not be extracted with ethanol, even at pH 1.0 (Table 2). However, approx. 40% of the total Al content in leaves of M. malabathricum, in which soluble Al mainly consisted of monomeric Al ions and an Al–oxalate complex, was extracted with ethanol (pH not adjusted, Table 2). When organic acids in leaves of F. marginata were extracted with water at pH 1.0, the mole amounts of extracted citrate and oxalate, which were the primary organic acids in the extract, were less than 10% of extracted Al (Table 2).

Table 2.  Concentrations of Al, Si, oxalate and citrate (mm) in the ethanol- or water-extract of leaves from Melastoma malabathricum and Faramea marginata.
  1. Values are means of three replicates ± SE. 1Percentage of total Al in the leaves; n.d., not determined.

F. marginata
Ethanol (pH not adjusted)TraceTracen.d.n.d.
Ethanol (pH 1)TraceTracen.d.n.d.
Water (pH not adjusted)3.03 ± 0.02 (39)10.04 ± 0.010.20 ± 0.010.16 ± 0.03
M. malabathricum
Ethanol (pH not adjusted)1.53 ± 0.14 (42)1n.d.0.56 ± 0.010.13 ± 0.07
Water (pH not adjusted)2.70 ± 0.06 (74)1n.d.4.26 ± 0.090.33 ± 0.03

The Al and Si in leaves of F. marginata could be extracted with water, especially at low pH (Fig. 2). Near pH 1, 90% of the total Al and 35% of the total Si content was detected in the extract. Extraction patterns of Al and Si in leaves of F. marginata were similar to those of aluminium silicate, although they did not completely correspond with each other (Figs 2 and 3). By contrast, Si in shoots of rice, in which SiO2 was the main form of Si, was only soluble at high pH and extraction patterns of Si in leaves of rice completely corresponded with those of SiO2 (Ma et al., 2001b; Fig. 2). Since the leaves and chemical reagents were well ground and mixed, the standard errors of Al, Si and pH were 7.7, 7.2 and 4.3% at maximum cases (except for the data near zero).

Figure 2.

Extraction patterns of aluminium (Al) from leaves of Faramea marginata and aluminium silicate. *, Percentage of total Al in leaves. The lines were fitted by hand.

Figure 3.

Extraction patterns of silicon (Si) from leaves of Faramea marginata, aluminium silicate, rice shoots, and SiO2. *, Percentage of total Si in leaves. The lines were fitted by hand.

Al and Si localization in the shoot

The PCV formed a blue colour by making a chelating complex with Al. Although PCV would also chelate with other metal cations, such as Fe, the disturbance by these elements could be ignored because the concentrations of these elements were much lower compared with that of Al (Britez et al., 1997). The PCV staining indicated high Al levels in both the upper and lower epidermis, and also showed Al accumulation in the cell walls of the palisade and the spongy mesophyll (Fig. 4a). The transverse section of the midrib illustrated that Al is localized in the cell walls of the upper and lower epidermis, parenchyma and collenchyma, but absent in sclerenchyma, xylem and phloem (Fig. 4b). The phloem and part of the ground tissue under the upper epidermis showed a dark colour, possibly because these areas included tanniniferous cells or secretory cells with unidentified contents. As this dark colour was also observed in sections that were not stained with PCV, these cells did not appear to react with PCV and therefore provided no evidence for Al localization. The transverse stem section clearly illustrated that the epidermis and cortex cells (collenchyma and parenchyma) were strongly stained with PCV, while the cuticle, sclerenchyma, phloem, xylem, and pith parenchyma did not stain blue (Fig. 4c).

Figure 4.

Transverse sections of Faramea marginata after pyrocatechol violet staining, the blue-coloured tissues indicate the localization of aluminium (Al): (a) leaf lamina; (b) leaf lamina with midrib; (c) stem section; e, epidermis; c, cortex; le, lower epidermis; m, mesophyll; p/c, parenchyma/collenchyma; pa, palisade parenchyma; ph, phloem; pi, pith parenchyma; s, sclerenchyma; ue, upper epidermis; x, xylem.

The images obtained using the fluorescent X-ray analytical microscope showed a similar distribution regarding Al and Si in a whole leaf (Fig. 5a,b) and stem section (Fig. 5c,d). Both Al and Si were found all over the leaf and mainly occurred in the cortex of the stem sections (Fig. 5c,d) which corresponded with the PCV staining (Fig. 4c).

Figure 5.

The localization of aluminium (Al) and silicon (Si) in Faramea marginata as determined by fluorescent X-ray analysis: (a,b) whole leaf (c,d) stem section.


Since the mean Al concentration in all 30 leaf samples of F. marginata analysed is 18 100 mg kg−1, this species is indisputably an Al accumulator. Even the lowest Al level obtained is much higher than the criterion of 1000 mg kg−1 used to define Al accumulators. Moreover, the Al levels found in F. marginata are in agreement with previous studies that reported Al concentrations ranging from 16 000 to 36 900 mg kg−1 in several Faramea species (Chenery, 1946). We suggest that probably all Faramea species are strong Al accumulators, since there are no records of nonaccumulating specimens in this genus (Chenery, 1946, 1948; Jansen et al., 2000a). Aluminium accumulation was also reported in all wood samples of Faramea that were tested using the chrome azurol-S test (Kukachka & Miller, 1980; Jansen et al., 2000b).

It is well known that chelation is one of the most important mechanisms to detoxify Al in the shoot of Al accumulators (Ma et al., 2001b; Watanabe & Osaki, 2002). The Al forms in shoots of several Al accumulators have been identified to be Al complexes with organic ligands (Nagata et al., 1992, 1993; Ma et al., 1997b,c; Watanabe et al., 1998a). In the present study, however, the results obtained from the extraction with ethanol and the analysis of organic acids in the water extract indicate that neither Al–organic acid complexes nor monomeric Al occur in the leaves of F. marginata.

Silicon is generally considered to interact with Al at the root level, and the formation of an Al–Si complex in above-ground tissues has not yet been well demonstrated. Si accumulation is common in the shoot of several monocots, pteridophytes and bryophytes, but this feature is relatively rare in dicots (Ma & Takahashi, 2002). Silicon accumulators are defined as plants with Si levels above 10 000 mg kg−1 and a Si : Ca ratio higher than 1, because plants that accumulate Si tend to have a low Ca concentration (Ma et al., 2001b; Ma & Takahashi, 2002). The Si : Ca mole ratio in the leaves of F. marginata is higher than 1, but Si levels are above 10 000 mg kg−1 in only five out of the 30 samples analysed . This suggests that F. marginata can be no more than an intermediate or weak Si accumulator (Fig. 1, Table 1).

The most important conclusion from this study is that there is strong evidence to suggest that an Al–Si complex may be formed in shoots of F. marginata. This hypothesis is based on the following results: (1) the striking positive correlation between Al and Si in various aged leaves (Fig. 1); (2) the extraction patterns of leaves are similar to those of aluminium silicate, not to Si in rice or SiO2 (Figs 2,3), and (3) the two elements show a similar localization in both leaves and stems (Figs 4 and 5). It has been documented that Al and Si(OH)4 form hydroxyaluminosilicate (HAS) with a Si : Al ratio of approx. 0.5 in acidic solutions (Exley et al., 2002). Since the mole ratio of Si : Al in the leaves of F. marginata estimated from the regression line in Fig. 1a is approximately 0.5, the formation of HAS may occur in the shoot of this species. However, the variations in the Si : Al ratio are not negligible (0.34–0.60, see Fig. 1b,c), which may indicate that more than one Al–Si compound occurs. This can be supported by the slight differences found in the extraction patterns of Al and Si between F. marginata and aluminium silicate. However, these differences may be affected by the chemical composition of the aluminium silicate used, because it is not known whether this reagent contains only a single or more than one Al–Si compound. Interestingly, the mole ratio of Si : Al shows a significant positive correlation as a function of Si concentrations (Fig. 1c). Therefore, Si seems to regulate the Al accumulation in leaves of F. marginata. Moreover, data on the Si : Al ratio in other Al-accumulating Rubiaceae are found to differ widely (from 0.23 in Alberta minor Baillon to 17.4 in Emmeorhiza umbellata K. Schum.), and there appear to be different possibilities for Al–Si associations (S. Jansen et al., unpublished). An alternative possibility may be that Al–Si compounds occur as solid compounds such as aluminosilicate (AS) or phytoliths (plant opal) that mainly consist of hydrated silica. It has been reported that 60–95% of the total Al content in leaf tissues of plants growing on acid soils is incorporated into phytolith structures (Bartoli & Wilding, 1980). Klinowski et al. (1998) found that Al in bamboo species (Melocanna bambusoides Trin.) was incorporated into the silicate network. However, this hypothesis is unlikely in F. marginata because phytoliths are absent in this species (Fig. 3), as well as in all other Al-accumulating Rubiaceae. Further research is needed to identify the actual chemical forms of the Al–Si compounds. Since HAS and solid Al–Si compounds are considered to be nontoxic or less toxic than monomeric Al, F. marginata may be able to grow successfully in acid soils despite high concentrations of Al in its shoots.

Compartmentalization, especially localization of Al in the cell wall or vacuole so that Al3+ cannot interfere with the cytoplasmic activities, has been suggested by several authors (Ma et al., 2001a; Watanabe & Osaki, 2002). Observations on the localization of Al in the shoot of Rubiaceae have, as far as we know, only been investigated in three species of Palicourea (Haridasan et al., 1986; Cuenca & Herrera, 1987). As suggested for several other Al accumulators, Al occurs mainly in the epidermis of the leaves in Palicourea and F. marginata (Matsumoto et al., 1976; Haridasan et al., 1986; Cuenca et al., 1991; Watanabe et al., 1998a). This may suggest that Al accumulators avoid Al toxicity in their leaves by localizing Al in the epidermal cells, which do not directly participate in photosynthesis in general. Accumulation in the epidermis of leaves has also been reported for nickel (Ni) and zinc (Zn) in heavy metal hyperaccumulators (Krämer et al., 1997; Küpper et al., 1999; Frey et al., 2000) and with respect to Ca and chlorine (Cl) in Hordeum vulgare (Leigh & Storey, 1993). Karley et al. (2000) suggested that ions that are not preferentially taken up at the bundle sheath plasma membrane move from the leaf vascular tissue to the epidermis via vein extensions. Localization of Al and other elements in the epidermis of the leaves may be explained by this hypothesis.

In addition to the epidermis, Al occurs in the mesophyll of the leaves in F. marginata (Fig. 3a), as was also reported for the Al accumulator M. malabathricum (Watanabe et al., 1998a). Since fluorescent induction of the leaves of M. malabathricum changed with increasing Al concentration in the leaves, this may indicate that Al affects the photosynthetic reaction in the mesophyll cells (Watanabe et al., 1998b). By contrast, the photosynthetic performance of two mistletoes species (Phthirusa ovata Eichler and Phoradendron crassifolium Nutt.) was found to be independent of whether they parasitized Al-accumulating or nonaccumulating host species (Lüttge et al., 1998). It may be worth studying in more detail the effect of Al on the photosynthesis in Al accumulators.

In general, the concentration of an element with low mobility in the plant body is higher in mature leaves than in young leaves. The mature : young leaf ratios found for Al, Si and Ca in F. marginata are 1.24, 1.36 and 1.56, respectively. This may indicate that the mobility is slightly higher for Al and Si than for Ca. With respect to Al, a significant difference between mature and young leaves has been observed in Al nonaccumulators and in at least few Al accumulators (e.g. the tea bush, F. esculentum) (Matsumoto et al., 1976; Liang et al., 1996; Osaki et al., 1997; Shen & Ma, 2001). Therefore, Al can be suggested to have a rather low mobility in these species. In most Al accumulators, however, Al concentrations do not appear to show considerable differences between young and mature leaves and this is also observed in F. marginata (de Medeiros & Haridasan, 1985; Osaki et al., 1997; Watanabe et al., 1997, 1998a,b; Masunaga et al., 1998). Hence, Al seems to behave as an element with high mobility in the majority of Al accumulators including F. marginata. Although Al localization is not observed in phloem in the present study (Fig. 4b,c), further research is required because Al has been suggested to be transported through the phloem elements in several Al accumulators (Haridasan et al., 1986).

In conclusion, this study strongly suggests that Si may not only alleviate Al toxicity near roots or in the rhizosphere, but also in shoots of some Al accumulators, such as F. marginata. We believe this is the first study providing evidence for the formation of an Al complex with an inorganic ligand in the shoot of an Al accumulator. Not only F. marginata, but also many species in Rubiaceae tend to accumulate Al and Si in their shoots (Golley et al., 1980; S. Jansen et al. unpublished). Further research is required to determine whether all Al-accumulating Rubiaceae form Al–Si complexes in their shoot tissues.


We thank M. Okamoto for the rice samples and two anonymous reviewers for valuable comments. TW is a Domestic Research Fellow of the Japan Society for the Promotion of Sciences (JSPS) and SJ is a postdoctoral fellow of the Fund for Scientic Research–Flanders (Belgium) (FWO–Vlaanderen).