1. Introduction

Silicon (Si) uptake and accumulation is a functional trait with multiple implications for plant biology and ecology [1,2]. Silicon’s manifold functions in plant biology include protection from a myriad of abiotic and biotic stresses and confers many benefits to plants that are capable of taking up and accumulating large amounts of it, ranging from practically zero in some taxa to 5% dry weight (and in extreme cases even more) in grasses [1,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17], and probably even to those plants that take up considerably smaller amounts [5,18]. Moreover, its uptake from the soil and eventual reincorporation into the soil through plant litter and herbivore feces also affects soil properties and Si cycling [19,20,21,22,23]. Thus, Si uptake, accumulation and cycling by plants is a key phenomenon in many ecosystems [12,21,24,25,26], with direct and indirect implications for ecosystem properties and processes [1]. Here, we review the existing knowledge of Si in the soil–plant continuum, its roles in plant biology and ecology, in ecosystem processes, and the possible implications for various ecosystem services.

2. Historical Overview

In 1787, Antoine Lavoisier predicted the existence of the element Si, to which Sir Humphry Davy proposed the name “silicon” in 1808. It was eventually isolated and formally discovered in 1823 by Jöns Jacob Berzelius. The discovery of the occurrence of Si within plants quickly followed, owing to the works of some prominent scholars, including microscopy pioneer Christian Gottfried Ehrenberg (who suggested the word “phytolitharia” to describe plant mineral components) [27] and Charles Darwin, who provided him with some samples [28,29]. Silicon effects on plant performance are known for more than 170 years, starting with Struve in 1835 [30], and shortly after Davy’s pioneering publication from 1846, who suggested that Si is present in the epidermis of grasses, where it strengthens the plants and makes them more resistant to attacks by insects and “parasitical plants” [31]. A surge of research soon followed. Sachs [32] showed in 1862 that Si-accumulating plants were a less preferred food than their conspecific plants that were grown hydroponically in Si-poor media. He further found that Si accumulation started in plant hairs and further advanced into the epidermis and near leaf vascular tissues. He also suggested that not all Si deposits in plant leaves are hard, but some may remain in a colloidal state. Miliarakis [33] found in 1884 that basal (younger) leaf parts of Si-accumulating plants had lower Si concentrations than older leaf tips and could not deter feeding. He also found that horsetail (Equisetum) and sedge (Carex and Scirpus) old leaf sheaths had a high Si concentration and protected the younger and less silicified plant tissues from herbivory. Furthermore, in 1884, Kreuzhage and Wolff [34] suggested the importance of Si for oat plants. Kerner von Marilaun [35] suggested in 1887 that the sharp leaf margin of sedges may be due to Si deposits. In 1888, Stahl [36] summarized other studies and concluded that silica deposits in horsetails impeded grazing by snails. He also mentioned that plant tissue silicification reduced the food quality of tropical grasses in Africa for domestic animals.

A second surge of studies started in the 1920s. Between 1922 and 1925, Lemmermann and colleagues [37,38,39] found an increase in the yield of rye grown under phosphorus deficiency upon Si fertilization. Sommer [40] concluded in the year 1926 that rice plants without Si fertilization suffered from early increased leaf wilting, guttation, and retarded panicle formation. In parallel, a surge of studies on Si effects on rice took place in Japan, starting in the 1910s and continuing into the 1940s, summarized thoroughly by Ma and Takahashi [41]. In the 1910s, Onodera [42] found that blast-infected rice leaves had a lower Si concentration than healthy leaves. Miyake et al. [43] also found in 1922 that Si concentration was higher in blast-resistant plants than nonresistant (surprisingly, this specific Japanese research was published in German). Other studies showed an increased blast-resistance of rice plants following Si application [44,45]. In the 1930s Ishibashi [46,47,48] showed reduced growth and yield of rice plants under Si deficiency and reduced blast after Si fertilization. Raleigh [49] showed in 1939 that Si was strongly improving the growth of beet plants. A year later, Wagner [50] showed that Si protects plants against powdery mildew by silicifying host plants’ cell walls, hence reducing penetration.

The study of silicon in plants continued shortly following the end of World War II, and especially since the 1960s. Engel [51] found in 1953 that the Si accumulated in rye culm was ~1/3 easily (hot water) extractable, suggesting no strong binding of this fraction to the plant cell walls. Holzapfel and Engel [52] found that Si accumulation in wheat increases over a study period of 30 days. In 1962, Yoshida [53] discovered the cuticle–Si double layer and suggested that this layer may be responsible for plant resistance against fungal diseases. There were several publications by Okuda and Takahashi, who found that Si improves plant resistance to metals [54] and rice growth and nutrition of [55]. Many of these studies were reviewed by Lewin and Reimann at the end of the 1960s [56]. During the same period, Jones and his colleagues (mainly Handreck) published some seminal papers on the occurrence, uptake, localization and functions of Si in oats [57,58,59,60] and clover [61]. At around the same time, a group that developed around Parry and his successor Sangster (both of which have passed away only in the past decade in their 90 s) further looked into the physiology of Si uptake and deposition [62,63,64,65,66,67,68,69,70,71]. Their early studies set the basis for a greater effort in Si plant and soil research, and a third surge that started in the 1980s and continues to these days, initiated by many important scientists and continuing to these days by the many who follow their footsteps.

3. Silicon Uptake by Plants

3.1. Active Uptake by Intrinsic Transporters

Plant taxa vary in the amounts of Si they take up and accumulate, a variability that is manifested through variations in Si contents, uptake mechanisms, forms and deposition locations. Traditionally, plant taxa have been divided into three major types based on their Si uptake capabilities, defined by the amounts of Si taken up by the plant (often measured as Si content in the xylem) relative to the amounts of available Si in the soil solution. If the amount of Si in the plant is considerably larger than that in the soil solution, the plant takes up Si actively; if the amount of Si in the plant is considerably lower than in the soil solution, the plant excludes Si; if the two contents are comparable, the plant takes up Si passively [104]. As straightforward as this division is, it is over-simplified and lacks mechanistic rigor.

For once, the reference to available Si content in the soil solution can be misleading since the great variability of soil Si pools implies that a species may take up Si both actively and passively, depending on the soil type and parent material. A dynamic approach is more appropriate since it indicates the plant’s response to varying soil Si availabilities and can furthermore point to the underlying internal (physiological) drivers of such responses [7,16]. For example, some species seem to increase their active Si uptake when Si availability in the soil is lower, suggesting an active response to fulfill plant internal Si demands when passive uptake is insufficient [105,106]. In some cases, this is achieved by increasing the expression of Si transporter genes and the density of these transporters under low Si availability conditions [16], indicating a truly active uptake that does not only rely on active uptake mechanisms but also on physiological responses of these mechanisms. Furthermore, Si uptake also depends on transpiration rates, with some species demonstrating passive (transpiration-driven) Si uptake in addition to active (transporter-governed) Si uptake [105,106,107]. The modes and drivers of Si uptake and accumulation and its variability among species are, therefore, not as simple as an active/passive/exclusion division implies.

Several transporters and genes that are involved in Si uptake and accumulation have been studied so far. Although the study of Si transporters focuses on rice and other grasses (as is commonly the case in plant Si research [5,18]), the first plant gene to regulate Si accumulation was discovered in the gourd Cucurbita (Cucurbitaceae), regulating Si and phytolith formation in the fruit rind [108]. Shortly after, a surge of discoveries of the physiology and genetics of Si uptake in grasses has arisen, revolving around the four Lsi transporters, all belonging to the NIP aquaporin family. The first transporter to be discovered was the influx transporter Lsi1, located in the distal plasma membrane of root exodermis and endodermis cells [109]. An efflux transporter on the proximal plasma membrane of the same cells, Lsi2, transports Si from the exodermis to the cortex and further loads it from the endodermis onto the xylem [110]. A third transporter, Lsi6, exists in the shoots and is responsible for xylem offloading [111]. In grass shoot nodes, Lsi6 and Lsi3 (previously thought to be Lsi2 due to structural similarities) are involved in distributing Si among branches [112,113,114,115]. Together, these transporters constitute an elaborate cooperative system of Si uptake and distribution in grasses, with some variations in the details of where exactly each transporter is localized within each species [16,112,116] (Figure 1).

Figure 1. A simplified model of Si uptake from the soil to the shoot through the transpiration stream, including main transporters and responses to external factors.

Lsi1 and Lsi6 transporters were also identified in the soybean Glycine max [117]. In the Cucurbitaceae, a Si-accumulating dicotyledonous family, Lsi1 was also identified in all root cells of Cucurbita [118]. Wang et al. [119] identified two putative Si transporters in cucumber (Cucumis) of the same family. Together with the early-identified gene responsible for Si accumulation in Cucurbita rinds [108], these studies suggest that Si transport systems in grasses and dicotyledons share some similarities. The recent identification of a gene regulating Si uptake by the mangrove Rhizophora apiculate without identifying the transporter itself [120]. Multiple genes regulating Si uptake and accumulation were also found in the horsetail Equisetum arvense [121]. Finally, it appears that Lsi-like genes that govern Si uptake are common in many groups of land plants, suggesting that the origins of these mechanisms are as ancient as the origins of land plants [122]. These findings further suggest that the physiology and genetics of Si transporters in non-grass species are only beginning to reveal themselves.

Expression of the Lsi1 gene in rice is downregulated by Si supply, dehydration stress and abscisic acid (more strongly in Si-depleted plants), suggesting regulation of active Si uptake in response to changes in the transpiration stream and plant internal water balance [123]. Further studies have demonstrated how the expression of Lsi1, Lsi2 and Lsi6 genes is regulated by plant hormones [124] and internal Si and metal concentrations [125,126].

3.2. External Factors Affecting Silicon Uptake

In addition to intrinsic transporters, external factors also affect Si uptake and accumulation in plants. These include both passive uptake mechanisms driven by the transpiration stream (the soil–plant–air continuum; Figure 1) and active mechanisms induced or enhanced by biotic stressors. Since Si is taken up from the soil as monosilicic acid within the soil solution, passive Si uptake depends on the transpiration stream. Several studies have shown that plant Si content in grasses increases with soil water content and availability, most probably for the simple reason that the more water a plant absorbs, the more Si is taken up with it [127,128,129,130,131,132,133]. On the other hand, transpiration, acting as the motive force of water uptake, has also been shown to increase Si content in grasses, to the degree that Si content has been suggested to serve as an indicator to plant transpiration stress [113,133,134,135,136,137]. Hence, along large rainfall gradients, Si content tends to demonstrate a U-shaped curve (minimum Si content at approximately 200–300 mm mean annual rainfall), implying an interplay between water availability and transpiration motive force [134,136]. Nevertheless, high plant Si contents in extremely arid conditions may also occur because grasses under drought stress take up Si more actively for the benefit it confers in resisting drought or in herbivore deterrence (see Section 5 below) [134,136,137]. Nevertheless, the failure to observe this positive correlation in other studies of grasses [138,139] suggests that other variables may confound the simple positive effect of water availability [134,136] (For Si effects on soil water availability co-appearing with Si availability in soils, see Section 3.2 above). In the Asteraceae family, an intermediate Si accumulator with mostly passive Si uptake (as far as we know, no attempts were made so far to identify Si transporters in this family), there appears to be no clear, consistent pattern, suggesting that Si uptake is not simply driven by the transpiration stream [134,136]. That the expression of the Lsi1 gene in rice is down-regulated by dehydration stress and abscisic acid [123] is a further indication for the complex effects of the transpiration stream on Si uptake.

Several studies have shown that ambient CO₂ concentrations also affect plant Si uptake, content and form, but with contradictory results. Ambient CO₂ concentrations had no effect on root and shoot Si contents in sugarcane plants [140]. In rice, increased ambient CO₂ concentrations reduced husk Si deposition by as much as 60% [141]. Increased ambient CO₂ concentrations alter the composition of phytolith assemblages in Phragmites and reduce mean phytolith size [142], suggesting an effect on Si allocation and distribution. Despite these studies being limited and equivocal regarding the regulatory role of CO₂ on plant Si uptake and accumulation, they harness potential significance for our understanding of global Si–carbon relationships, namely the possibility of Si being a partial substitute for carbon in plants (see Section 5 below) and Si’s role in regulating the carbon cycle (see Section 6 below).

Among the biotic stressors known to affect plant Si content, herbivory is the one that was studied the most [143]. Exposure to invertebrate [144,145,146,147] and vertebrate [144,148] herbivores induces Si uptake and accumulation in grasses. Comparable induction by artificial clipping [149,150,151] further supports that this induction is directly associated with biomass removal or damage. While such induction was sometimes not observed in controlled experiments [144,151,152,153,154,155], this is likely because these experiments did not incorporate sufficiently long exposure times to initiate a response [134,144]. In natural landscapes, higher grass Si contents are associated with larger densities of herbivorous rodents [154,156,157,158,159], but not of larger herbivores [137,148,160] or following clipping [151], which is most likely explained by the involvement of other environmental variables in natural ecosystems having stronger effects on Si uptake and accumulation [134]. Nevertheless, recent evidence for cyclic dynamics of vole densities and grass Si contents [157,159,161] provides further support to the induction of grass Si uptake by herbivory. Among the Asteraceae, the only non-grass family in which the possible effect of herbivory on Si has been widely studied, such an effect was rare and weak [137].