Silica uptake in aquatic and wetland macrophytes: a strategic choice between silica, lignin and cellulose?


  • Jonas Schoelynck,

    1. Department of Biology, Ecosystem Management Research Group, University of Antwerp, Universiteitsplein 1C, B–2610 Wilrijk, Belgium
    Search for more papers by this author
  • Kris Bal,

    1. Department of Biology, Ecosystem Management Research Group, University of Antwerp, Universiteitsplein 1C, B–2610 Wilrijk, Belgium
    Search for more papers by this author
  • Hans Backx,

    1. Department of Biology, Ecosystem Management Research Group, University of Antwerp, Universiteitsplein 1C, B–2610 Wilrijk, Belgium
    Search for more papers by this author
  • Tomasz Okruszko,

    1. Department of Hydraulic Engineering and Environmental Reclamation, Warsaw Agricultural University, ul. Nowoursynowska 166, PL–02-787 Warszawa, Poland
    Search for more papers by this author
  • Patrick Meire,

    1. Department of Biology, Ecosystem Management Research Group, University of Antwerp, Universiteitsplein 1C, B–2610 Wilrijk, Belgium
    Search for more papers by this author
  • Eric Struyf

    1. Department of Biology, Ecosystem Management Research Group, University of Antwerp, Universiteitsplein 1C, B–2610 Wilrijk, Belgium
    Search for more papers by this author

Author for correspondence:
Jonas Schoelynck
Tel: +32 3 2652252


  • Although silica (Si) is not an essential element for plant growth in the classical sense, evidence points towards its functionality for a better resistance against (a)biotic stress. Recently, it was shown that wetland vegetation has a considerable impact on silica biogeochemistry. However, detailed information on Si uptake in aquatic macrophytes is lacking.
  • We investigated the biogenic silica (BSi), cellulose and lignin content of 16 aquatic/wetland species along the Biebrza river (Poland) in June 2006 and 2007. The BSi data were correlated with cellulose and lignin concentrations.
  • Our results show that macrophytes contain significant amounts of BSi: between 2 and 28 mg BSi g−1. This is in the same order of magnitude as wetland species (especially grasses). Significant antagonistic correlations were found between lignin, cellulose and BSi content. Interestingly, observed patterns were opposite for wetland macrophytes and true aquatic macrophytes.
  • We conclude that macrophytes have an overlooked but potentially vast storage capacity for Si. Study of their role as temporal silica sinks along the land–ocean continuum is needed. This will further understanding of the role of ecosystems on land ocean transport of this essential nutrient.


Vegetation is a vast sink for biogenic silica (BSi; Conley, 2002). More and more evidence points towards strong control of silica biogeochemistry by plant communities (Derry et al., 2005; Struyf & Conley, 2009). Silica is not an essential element in the growth of higher vascular plants in the classical sense (Raven, 2003): most plants (except for Equisetaceae) can grow in silica deprived mediums. However, dissolved silica (DSi) uptake increases the rigidity and the survival chances of plants in dynamic environments. Silica is usually found in plant bodies named phytoliths or hydrogen bonded to cellulose molecules in the cell wall as a silica gel, SiO2.nH2O (Kaufman et al., 1999). In general three types of beneficial effects are recognized: structural, physiological and protective (Sangster et al., 2001; Street-Perrott & Barker, 2008). The presence of Si in plants has been found to alleviate abiotic and biotic stresses (including resistance to pathogens, and alleviation of metal toxicity, and of drought and salinity stress) leading to the incorporation of silicates into many fertilizers (Currie & Perry, 2007). Harrison (1996) showed that silica in plants is closely connected to cell wall proteins, and suggested that in the formation of biogenic silica a combination of biomolecules is involved.

Because river vegetation experiences constant hydrodynamic forces up to 25 times higher than terrestrial vegetation (Denny & Gaylord, 2002), increasing structural rigidity benefits macrophytes. They often accomplish structural rigidity with energetically expensive compounds such as lignin (6957 kcal kg−1) and cellulose (4000 kcal kg−1) (Jung et al., 1999). BSi could be a cheap alternative to these molecules for riverine vegetation, as it has a similar effect on the cell wall but a 10–20-fold lower energetic cost (Raven, 1983). Silica is an energetically cheap stiffening material, promoting upright stature and resistance to lodging (flattening by wind or rain). DSi is available in relatively high amounts in freshwater systems: the concentration in freshwater systems is on average 200 μm (Meybeck, 1979).

It is clear that aquatic and wetland species could benefit from BSi incorporation. For certain wetland species, like Phragmites australis (reed), it is known that they play an important role in the accumulation and release of Si (Struyf et al., 2007). However, detailed information on BSi accumulation for most common macrophyte species is currently lacking, as are potential links to the metabolism for other structural components. This is a gap in our understanding of both the individual plants’ strategy to withstand high flow forces and in our understanding of riverine Si biogeochemistry. A strong seasonal uptake, as observed for diatoms (Conley, 1997), could substantially alter riverine Si fluxes during the growth season and could potentially change the eventual fate of the transformed Si.

In this study we present intricate correlations between BSi, cellulose and lignin metabolism in macrophytes. We show that aquatic vegetation is a potentially large reservoir of Si. Our results suggest that wetland vegetation and true aquatic vegetation differ in their respective uptake dynamics of Si and production of cellulose and lignin. We show that aquatic vegetation should be considered in future studies of silica biogeochemistry, and that Si metabolism in macrophytes could be more intricate than commonly assumed.

Materials and Methods

Study site

Plant species were collected along the Biebrza river, a tributary of the Narew river in north-eastern Poland. The Biebrza river is c. 160 km long and drains 7000 km2. Almost the entire valley of the Biebrza river lies within the borders of the Biebrza National Park. Owing to the low fertilization character of the local agricultural activities the impact on nutrient cycles is low and pristine vegetation communities occur. For a detailed description of the Biebrza river valley, see Wassen et al. (2006).

Vegetation sampling and BSi analysis

In June 2006 and June 2007 five individuals of 16 plant species were randomly collected by hand on each of 10 locations along the length of the Biebrza river (if present) (Fig. 1). The only prerequisite for collection was that plants were standing in the river water: true aquatic as well as near-riverbank wetland species were included (Table 1).

Figure 1.

 A plan view of the Biebrza National Park in north-east Poland. Vegetation samples were taken along a downstream gradient in the river. Sample locations, from upper towards lower basin, were situated nearby the following villages: 1, Rogozyn Stare; 2, Rogozynec; 3, Lipsk; 4, Ostrowie; 5, Sztabin; 6, Jagłowo; 7, Dolistowo Stare; 8, Osowiec; 9, Mścichy; 10, Burzyn.

Table 1.   Summary of 16 collected species, divided into two groups: aquatic species (true aquatic) and wetland species (near-riverbank)
  1. Scientific names after De Langhe et al. (1995).

Ceratophyllum demersum L.Berula erecta (Huds.) Coville
Hydrocharis morsus-ranae L.Glyceria maxima (Hartm.) Holmb.
Myriophyllum spicatum L.Mentha aquatica L.
Nuphar lutea (L.) Sm.Phalaris arundinacea L.
Potamogeton crispus L.Phragmites australis (Cav.) Trin. ex Steud
Potamogeton lucens L.Rorippa amphibia (L.) Besser
Potamogeton natans L. 
Potamogeton perfoliatus L. 
Sagittaria sagittifolia L. 
Sparganium emersum Rehmann 

The collected plants were thoroughly rinsed and cleaned to completely remove sediments, epiphytic algae and macro-invertebrates. Subsequently, they were dried at 70°C for 48 h. All five individuals for each species were pooled per location, ground with a mill and homogenized. Biogenic silica was then extracted from 25 mg of dry plant material by incubation in a 0.1 m Na2CO3 mixture at 80°C for 4 h (DeMaster, 1981). The extracted and dissolved Si was spectrophotometrically analysed on a Thermo IRIS inductively coupled plasmaspectrophotometer (ICP; Thermo Fisher, Franklin, MA, USA). Extraction in 0.1 m Na2CO3 at 80°C has been well established and tested; it is capable of fully dissolving the BSi from plant phytoliths at the solid–solution ratios and extraction time we applied (Saccone et al., 2007).

Lignin and cellulose analysis

α-Cellulose and hemicellulose are structural polysaccharides, often referred to as holocellulose. α-Cellulose is a relatively pure compound of glucose units. Lignin, however, does not have a precise chemical formula, but is a complex molecule composed of repeating phenylpropane units composed of an aromatic ring with three carbon side-chains. The phenolic groups vary between plants, further confounding the definition of lignin (Palm & Rowland, 1997). Basically, two major techniques are often used for lignin analysis: ‘Klason lignin’ determination for forestry products (Effland, 1977) and ‘Acid-detergent fibre (ADF)-lignin’ for assessing animal forage quality (Van Soest, 1963). The latter is often referred to as more precise and more reproducable (Ryan et al., 1990; Rowland & Roberts, 1994). Moreover, the Van Soest method (Van Soest, 1963) also enables the determination of α-cellulose seperately from hemicellulose, in contrast to the Effland method (Effland, 1977).

We used the Van Soest method (Van Soest, 1963) to analyse the plant samples for α-cellulose and ADF-lignin. In short, between 0.5 and 1 g of dry plant material was treated with cetyltrimethylammonium-bromide to dissolve and remove proteins. The remaining material was treated with a 72% sulfuric acid. After weighing and drying at 105°C, the α-cellulose content was calculated by subtracting the mass before and after the sulfuric acid treatment and dividing this value by the initial material mass. In a third step the remaining material was ashed at 550°C: the ADF-lignin content was then calculated by subtracting the mass before and after ashing and dividing it by the initial material mass. To determine the laboratory error, we analysed an independent Potamogeton natans sample 10 times. This resulted in an average amount (± SE) of 273 ± 2 mg g−1 DM α-cellulose and 116 ± 3 mg g−1 DM ADF-lignin, yielding reproducabilities of 2.7% and 8.6%, respectively. We will hereafter refer to ADF-lignin and α-cellulose as lignin and cellulose.


A two-way ANOVA test was performed to test for differences in Si, lignin and cellulose content between years and between locations. No significant differences were found. A Spearman correlation test per species between BSi concentration and distance along the river (km) was also performed: no significant correlations were found. The BSi concentration varied widely for both plant groups (wetland and aquatic species): between 7 and 28 mg BSi g−1 DM for true aquatic species and between 2 and 14 mg BSi g−1 DM for wetland species (Fig. 2).

Figure 2.

 Mean biogenic silica (BSi) concentrations (mean ± SE) of 16 macrophyte species from a 2-yr sampling period (2006–07). Aquatic species: Ceratophyllum demersum (n = 5), Hydrocharis morsus-ranae (n = 3), Myriophyllum spicatum (n = 8), Nuphar lutea (n = 20), Potamogeton crispus (n = 3), Potamogeton lucens (n = 14), Potamogeton natans (n = 12), Potamogeton perfoliatus (n = 7), Sagittaria sagittifolia (n = 17), Sparganium emersum (n = 15). Wetland species: Berula erecta (n = 7), Glyceria maxima (n = 9), Mentha aquatica (n = 7), Phalaris arundinacea (n = 6), Phragmites australis (n = 6), Rorippa amphibia (n = 15). n = number of samples within a 2-yr period; every sample contains five replicate individuals.

Aquatic species had variable cellulose concentrations: between 104 and 387 mg g−1 DM and on average (± SE) 222 ± 6 mg g−1 DM (= 107). The cellulose concentrations of wetland species ranged between 174 and 388 mg g−1 DM with an average (± SE) of 279 ± 8 mg g−1 DM (= 50).

The BSi and cellulose concentrations were related (across the different species) in both aquatic and wetland species: the correlation was opposite for both groups of species (Fig. 3). For aquatic species, lower cellulose concentrations were typically associated with increasing BSi accumulation (Pearson correlation test, < 0.0001). In wetland species, an opposite correlation was observed, as cellulose concentration increased with increasing BSi concentration (Pearson correlation test, < 0.0001). For both correlations, a linear fit was found to be the best fit.

Figure 3.

 Antagonistic correlations between the biogenic silica (BSi) concentration and the cellulose concentration for (a) aquatic species (= 107), (b) wetland species (= 50). Each data point represents the concentrations of an = 5 mixture of replica’s occurring on the same location along a downstream gradient. Both correlations are significant with a linear fit as best fit: aquatic species, = −3.5x + 269, R² = 0.29, < 0.0001); wetland species, = 6.1x + 234, R²=0.40, < 0.0001). Aquatic species: Ceratophyllum demersum (⋄), Hydrocharis morsus-ranae (△), Myriophyllum spicatum (inline image), Nuphar lutea (•), Potamogeton crispus (○), Potamogeton lucens (bsl00066), Potamogeton natans (bsl00001), Potamogeton perfoliatus (□), Sagittaria sagittifolia (+), Sparganium emersum (♦). Wetland species: Berula erecta (△), Glyceria maxima (♦), Mentha aquatica (□), Phalaris arundinacea (⋄), Phragmites australis (bsl00001), Rorippa amphibia (inline image).

Lignin concentrations ranged between 3 and 192 mg g−1 DM for aquatic species and between 24 and 144 mg g−1 DM for wetland species. The averages (± SE) were 66 ± 4 mg g−1 DM (n = 107) and 66 ± 4 mg g−1 DM (= 50), respectively.

In aquatic species, there was a tendency towards an opposite effect compared with cellulose (increasing BSi content was mainly associated with increasing lignin content), however, this correlation was not significant (Fig. 4a,b). Similarly, the BSi–lignin correlation for wetland species was opposite to the correlation found for cellulose. The logarithmic function (best fit) indicated lower lignin concentrations with increasing BSi concentrations (Pearson correlation test, < 0.05).

Figure 4.

 Antagonistic correlations between the biogenic silica (BSi) concentration and the lignin concentration for (a) aquatic species (= 107), (b) wetland species (= 50). Each data point represents the concentrations of an = 5 mixture of replica’s occurring on the same location along a downstream gradient. For the aquatic species, only a tendency for a positive correlation could be observed. The correlation for the wetland species is significant with a logarithmic fit as best fit: = −11.8loge(x) + 85, R² = 0.19, < 0.05). Aquatic species: Ceratophyllum demersum (⋄), Hydrocharis morsus-ranae (△), Myriophyllum spicatum (inline image), Nuphar lutea (•), Potamogeton crispus (○), Potamogeton lucens (bsl00066), Potamogeton natans (bsl00001), Potamogeton perfoliatus (), Sagittaria sagittifolia (+), Sparganium emersum (♦). Wetland species: Berula erecta (△), Glyceria maxima (♦), Mentha aquatica (□), Phalaris arundinacea (⋄), Phragmites australis (bsl00001), Rorippa amphibia (inline image).


The structural components BSi, lignin and cellulose have, to our knowledge, never been studied together in detail for aquatic and wetland plant species. Yet, these components play an important role for vegetation growing in dynamic environments. We have found antagonistic correlations between these structural components, across species, in both wetland and aquatic macrophytes. Although our observations do not enable us to define a within-species trade-off between structural components (more replicas per species would be needed), they do show that plant metabolism of BSi, lignin and cellulose is intricately intertwined. Species that invest more in deposition of biogenic Si appear to differ in lignin and cellulose metabolism compared with species that exclude Si. Moreover, the large amounts of BSi observed in the riverine vegetation show the vegetation could affect riverine Si fluxes in a way similar to summer diatom blooms.

Impact of macrophytes on riverine Si biogeochemistry

From our results it is clear that the biogenic silica content of macrophytes is highly variable, ranging between 2 and 28 mg BSi g−1 DM. This is comparable with the values found for several Poaceae and Cyperaceae (Lanning & Eleuterius, 1983; Hodson et al., 2005; Struyf & Conley, 2009). The high BSi contents of Ceratophyllum demersum, Myriophyllum spicatum and Sagittaria sagittifolia are quite remarkable and are similar to concentrations in wetland grasses such as P. australis, a species well-known for its capacity to accumulate BSi (Struyf et al., 2007). High BSi content together with high biomass production of macrophytes (up to 1100 g DM m−2) (Champion & Tanner, 2000), results in a large storage capacity for BSi. Silica storage capacity is temporary owing to the annual decomposition of macrophytes in autumn and winter. The resulting effect could be a seasonal cycle similar to that observed for spring and summer Si uptake by diatoms. A large temporal macrophyte sink for BSi could retain Si within the system and alter downstream fluxes during the growing season. Subsequent transport downstream will depend on the dissolution rate of dead macrophyte BSi (transport as DSi). Virtually nothing is known about this subject. Only one study in P. australis in tidal marshes has studied macrophyte BSi dissolution in detail (Struyf et al., 2007): here a quick dissolution was observed with over 50% dissolved within 14 d. Another possibility for downstream transport is in the form of BSi, in suspended plant biomass particles.

Macrophytes and diatoms

Silica is an important element in diatom growth (Paasche, 1980) and silica fluxes towards estuaries and oceans are strongly influenced by diatom production. In freshwater systems diatom silica content is one order of magnitude higher than in marine diatoms (Conley & Kilham, 1989) probably because of the higher availability of Si in freshwater systems (Meybeck, 1979). Freshwater diatoms also are significantly smaller (Litchman et al., 2009) rendering them more efficient in acquiring limiting nutrients as a result of their high surface area to volume ratio (Raven, 1998).

Macrophytes can temporarily store a significant amount of silica, rendering it unavailable for diatoms, with a potential negative impact on this community. Conversely, macrophytes will increase riverine friction (Bal & Meire, 2009) resulting in longer water retention times, which can stimulate and increase diatom production (Muylaert et al., 2001; Struyf et al., 2004). In general, speciation of Si in freshwater riverine ecosystems, both by diatom and macrophyte uptake, has been strongly neglected compared with estuaries, lakes and the marine environment. In these last three environments, diatom uptake has been found to exert essential control on silica biogeochemistry (Conley and Kilham, 1989; Treguer et al., 1995; Humborg et al., 2000). Yet, the retention and transformation processes for Si in upstream ecosystems directly affect downstream processes, and more research is certainly recommended.

A trade-off between structural components

From our results it is clear that interspecific differences in BSi concentrations are large, and that this is potentially linked to other structural components. Association between BSi and structural biomolecules was previously suggested (Harrison, 1996). Cells are embedded in a matrix mostly built up out of hemicellulose (Lodish et al., 2003). In intercellular space lignin can be incorporated to provide additional rigidity and compressive strength to cell walls. In addition to strength provided by lignin and cellulose, biogenic silica can provide support to the shoot (Kaufman et al., 1999). It is classically taken up as DSi via the roots, but for submerged species shoot uptake from the water column can also be important (similar to other nutrients; Marschner, 1997; Madsen & Cedergreen, 2002). Once it is transported to stems and leaves where transpiration is occurring, it irreversibly polymerizes as amorphous silica gel, SiO2.nH2O in cell walls and phytoliths. It has a similar effect on plant stiffness as lignin, but at a 10–20-fold lower energetic cost (Raven, 1983).

The clearest correlation between BSi, lignin and cellulose incorporation was observed in wetland species. Here it is clear that the species incorporating most BSi are all Poaceae. Wetland grasses are well known for their capacity to accumulate BSi (Struyf & Conley, 2009). Their presence in wetlands enhances biological fixation of Si, which in turn enhances BSi content of soils and increases porewater DSi concentrations (Struyf et al., 2009). It has been suggested that enhanced DSi availability in wetland soils, engineered by grasses, could in turn enhance their competitiveness and partly explain their dominant position in these systems (Struyf et al., 2009). Grasses have abundant DSi available for uptake from wetland soils, allowing them to incorporate it as BSi at a low energetic cost. Biogenic Si reduces the palatability of grasses for herbivores (Massey et al., 2007) and increases their strength. It also alleviates the effect of a wide range of possible plant stress factors, such as metal toxicity, salinity and drought, although the mechanisms have yet to be fully elucidated (Currie & Perry, 2007). Our observations seem to suggest a further competitive advantage of BSi accumulation by grasses: they spend less energy on expensive lignin metabolism, which they might in turn use for increased cellulose synthesis, further increasing their internal strength and competitive abilities.

We observed a different correlation between the structural components for aquatic species. It is important to note that no grasses were among the aquatic species. Here we observed that plants that invest in BSi, often also incorporate lignin. Increasing BSi and lignin content is compensated for by decreasing cellulose content. This could partly be explained by a trade-off in relative contents: increased content of one component automatically decreases the relative importance of other components. However, we hypothesize that the trade-off between cellulose, lignin and BSi could also relate to different choices of how to cope with external environmental forces.

Wetland species and true aquatic species face different environmental forces (Fig. 5). Wetland species need to overcome gravity ‘on their own’ and this is accomplished by the plants’ stiffness. This is achieved through the rigidity of the plant cell walls, which are made up out of cellulose. The potential of wetland grasses to acquire BSi reduces the cost for lignin incorporation, allowing them to invest more in cellulose, as discussed above.

Figure 5.

 Schematic overview of the forces a 14.7-g fresh mass Potamogeton natans individual with a projected frontal area of 0.04 m² endures in (a,b) water and (c) air. In (a) the plant is able to bend, in (b,c) the plant was assumed to be rigid. Horizontal forces are hydro- or aerodynamic forces (F = 0.5 × Cd × ρ × A × uβ; Denny, 1994). In (a) F is based on data from (K. Bal et al. unpublished data), Cd is derived from that force; in (b,c) the forces are formula based. Vertical forces are a lift force (L = 0.5 × ρ × u² × A × CL) and gravity (F = mg) and are formula-based in all panels. Cd, drag coefficient; ρ, viscosity (1000 kg m−3 for water; 1.293 kg m−3 for air at sea level); A, projected frontal area (m²); u, velocity (mean June stream velocity in Biebrza = 0.1 m s−1); β, 2 for rigid objects and < 2 for flexible objects that reconfigure by bending; m, mass (g); g, gravitational constant (9.81 m s−2); CL, lift coefficient at the desired angle of attack with CL = 2 × π × α with α is 40° at 0.1 m s−1 (K. Bal et al., unpublished). Note that in (b) the hydrodynamic force is up to 4.5-times the amount of force endured when bending was possible (a) and that gravity and the hydrodynamic force endured when bending are close to the same order of magnitude. Nutrient uptake through shoots is proven for many macrophyte species (Marschner, 1997; Madsen & Cedergreen, 2002) so we assumed the same for silica (Si). In situations (a,b) the plant can have a continuous dissolved silica (DSi) supply from the surrounding water and from the soil. In situation (c) the plant can only rely on an environmental determined soil supply.

Aquatic species are supported by the water body, and floating leaves help them to achieve upright stature. The larger the plant, the more hydrodynamic forces it endures for which structural molecules need to be incorporated. Our results suggest that aquatic plants invest mostly either in cellulose or in BSi and lignin. The plants need to find a balance between cellulose, lignin and Si incorporation to be strong and rigid but still flexible. In addition, it needs to be established as economically as possible. According to Kaufman et al. (1999) the use of lignin diminishes flexibility. The correlations we observe might result from a choice between a growth strategy that reduces bending possibilities (BSi and lignin) but enhances rigidity, or a choice for a strategy that allows bending (cellulose). This is only a hypothesis, and we have no proof. Other factors, such as herbivory might play a role and this warrants further investigation.


Our results show that aquatic macrophytes can contain significant amounts of BSi. This was well known before for wetland species (especially grasses) but not for true aquatic species. Variability in BSi among species is high. The effect of Si accumulation in macrophytes is not well studied but it could affect Si availability both at the growth location and more downstream.

Our results also suggest macrophyte Si metabolism is intricately intertwined with cellulose and lignin metabolism. Although our results do not allow any definitive conclusions on this subject, we would strongly encourage both experimental and observational studies looking at this subject in more detail, and testing the hypotheses we put forward.

We can conclude that macrophytes have a so far overlooked but potentially vast storage capacity for Si. Studying their role as temporal silica sinks along the land–ocean continuum is needed. This will further understanding of the role of ecosystems on land–ocean transport of this essential nutrient.


We thank the Board of Biebrza National Park for enabling us to work in this area. J.S. thanks IWT (Institute for the Promotion of Innovation by Science and Technology in Flanders) for personal research funding. E.S. thanks FWO (Research Foundation Flanders) for funding for both personal postdoctoral research and for the project ‘Tracking the biological control on Si mobilization in upland ecosystems’ (Project no. G014609N). We would like to acknowledge Belgian Science Policy (BELSPO) for funding projects LUSi (Land use changes and silica fluxes in the Scheldt river basin) and MANUDYN (Macrophytes and nutrient dynamics: process and field studies in the upper reaches of river basins). Special thanks go to Wout Opdekamp for the map and to Koen Wyckmans for figure editing.