Arsenic and old plants


Terrestrial life, like its marine progenitors, developed around vents in the earth's surface from which spewed a mineral-rich substrate high in a suite of metals and metalloids. To survive in this environment, organisms either had to maintain cytoplasmic homeostasis so that cellular processes could function or have biochemistry that functioned in the presence of what we consider toxicants. Typically, terrestrial and ocean vents release arsenic at much higher concentrations than found on average in the earth's crust, with life evolving in an arsenic rich environment. The most comprehensive picture of the early evolution of land plants comes from the Rhynie Cherts (c. 400 million yr old), NE Scotland, which was a subaerial hot-spring environment rich in arsenic (Rice et al., 1995). Could the elevated arsenic in the environment in which fern progenitors evolved be a selection pressure leading to the remarkable phenomenon of arsenic hyperaccumulation in ferns?

Who hyperaccumulates?

Ma et al. (2001 ) first reported arsenic hyperaccumulation in ferns in Pteris vittata . Pityrogramma calomelanos was rapidly added to this list ( Visoottiviseth et al., 2002 ), followed by three more species from the Pteris genus ( P. cretica , P. longifolia and P. umbrosa ), reported in this issue by Zhao et al . (pp. 27–31). However, not all ferns hyperaccumulate arsenic, as determined by Zhao et al. (2002 ), and others ( Kuehnelt et al., 2000 ; Visoottiviseth et al., 2002 ). With > 10 000 fern species and their allies to screen, it may be predicted that many more hyperaccumulating ferns will be discovered. Though the number of ferns tested for arsenic hyperaccumulation is small, some patterns are emerging. Fern phylogenies, while complex and evolving, place all five hyperaccumulating ferns in the Pteridaceae family ( Fig. 1 ), within the order Pteridales. The Pteridaceae alone contains > 400 species, with c . 10% of all ferns being in the Pteridales.

Figure 1.

Evolutionary tree for ferns derived from Pryer et al. (1995 ), with families containing arsenic hyperaccumulating ferns shown in red, and those that have had nonaccumulating ferns identified shown in green. Branch lengths are arbitrary.

Ferns are among the most primitive plants (Fig. 2) with molecular (rbcL sequences) and morphological characteristics placing the Pteridaceae quite late on in fern evolution (Fig. 1). The Pteridaceae are closely related to other families which, as far as we are aware, do not at present exhibit hyperaccumulation (Fig. 1), but with c. 1/1000 of all fern species screened, there is someway to go before the uniqueness of the Pteridaceae with respect to arsenic hyperacummulation can be assumed. Similarly, the hunt for a nonaccumulating Pteris species is on.

Figure 2.

Evolutionary tree for land flora from Pryer et al. (1995 ), with groups containing arsenic hyperaccumulating species shown in red, and those that have had nonaccumulating species identified shown in green. Branch lengths are arbitrary.

The most primitive fern screened to date, Dicranopteris linearis in the Gleicheniaceae, does not hyperaccumulate arsenic (Visoottiviseth et al., 2002). The most primitive land plants are in the Lycopodlophytes (Fig. 2). The only member of the Lycopodlophytes investigated to date, a Selaginella sp., also did not hyperaccumulate arsenic (Visoottiviseth et al., 2002). There are a number of studies on the Equisetacae, which evolved early on from the Lycopodlophytes, which show that although a range of species have high arsenic levels (> 100 mg kg−1 As), none hyperaccumulate arsenic (Wong et al., 1999; Kuehnelt et al., 2000; Hozhina et al., 2001). Caution should be maintained in stating that these members of primitive land flora do not hyperaccumulate arsenic, as all the mentioned studies were from field screening, and issues of the arsenic bioavailability and its heterogeneity need to be addressed.

The hyperaccumulating accessions studied by Ma et al. (2001) (P. vittata) and Visoottiviseth et al. (2002) (Pityrogramma calomelanos) were from contaminated sites. Zhao et al. (2002) also showed that P. vittata and the other three Pteris species from uncontaminated soils also exhibited hyperaccumulation. Hence, hyperaccumulation is not just a character selected in these species on arsenic contaminated soils, it appears to be constitutive. The constitutive nature of arsenic hyperaccumulation is not unique in this respect as the zinc and cadmium hyperaccumulation in Arabidopsis halleri and Thlaspi caerulescens is constitutive (Baker & Whiting, 2002). What is unique is that the arsenic hyperaccumulating ferns are extremely widespread, unlike A. halleri and T. caerulescens which have a limited geographical and ecological distribution.

Evolution of arsenic hyperaccumulation

Land plants that evolved in arsenic rich environments would have required mechanisms for coping with this element, with hyperaccumulation being one strategy. An excluder strategy has been well studied in grasses (Meharg & Hartley-Whitaker, 2002). Tolerance mechanisms may have been lost as plants spread out from hot-springs into nonarsenic contaminated environments, with members of the Pteridaceae for some reason retaining these primeval mechanisms, either as evolutionary baggage or because this trait conferred them with some advantage, arsenic related or not. Alternatively, members of the Pteridaceae evolved hyperaccumulation at a later stage in response to a particular selection pressure, or perhaps they were confined to arsenical habitats until late on in their evolutionary development.

Not all primitive plants may have evolved in arsenic rich environments. We have biased our view of where terrestrial life-forms evolved from the particular conditions that give us the best fossil record of early plant evolution, such as the silica rich extrusions from hot springs which led to the excellent plant preservation in the Rhynie Cherts (Rice et al., 1995).

Perhaps arsenic did not provide the selection pressure that led to arsenic hyperaccumulation. The physiological mechanisms responsible for arsenic hyperaccumulation may have evolved for other reasons, and just happen to confer arsenic hyperaccumulation to the ferns. This is indeed the case for arsenic resistance in angiosperms where resistance is a consequence of having suppressed high-affinity phosphate transport (Meharg et al., 1993; Fitter et al., 1998), as arsenate and phosphate are analogues. Arsenic hyperaccumulation may therefore be a consequence of unusual phosphorus metabolism in ferns, though this seems not to be the case (see below). It is imperative that we understand the physiology of hyperaccumulation to understand the evolutionary basis of this phenomenon.


Arsenate is the dominant plant available form of arsenic in aerobic environments. Plants will mobilise and take up large quantities of arsenate, if present, in their quest to obtain phosphate. Many ferns grow on peats subject to routine water logging, and the resulting reduced conditions in the production of arsenite from arsenate. It has recently been demonstrated for the first time that arsenite is actively taken up at a high rate by plant roots (rice) (Abedin et al., 2002), probably through aquaporins (Wysocki et al., 2001). Arsenite has a pKa of 9.2 and is predominantly in an undissociated form at neutral to acid pHs. In essence, arsenite or arsenate present in the ferns environment would be efficiently accumulated.

Zhao et al. (2002 ) report that there is nothing unusual in the total phosphorus status of arsenic hyperaccumulators. Recent transport studies on P. vittata show that there is nothing particularly abnormal about the kinetics of arsenate/phosphate transport, and that arsenite is taken up actively, though at a lower rate than arsenate ( Wang et al., 2002 ). This begs the question: how do arsenic hyperaccumulators mobilise arsenic in such vast quantities from the soil?

The physiological mysteries deepen once inside the plant. Consistently it has been shown that hyperaccumulators store up to 1% arsenic in their shoots, primarily as free arsenite (Ma et al., 2001; Francesconi et al., 2002; Wang et al., 2002). The presence of arsenite is not a surprise in itself as many plants rapidly reduce arsenate to arsenite (Meharg & Hartley-Whitacker, 2002), but the normal model for plant detoxification of arsenic is complexation of As(III) with phytochelatins (PCs) (Meharg & Hartley-Whitacker, 2002). The arsenite (As III) must be compartmentalised in the vacuole, unless the ferns have extraordinary protein biochemistry as As (III) has a high, and therefore disruptive, affinity for –SH groups. The arsenic could be shuttled through the cytoplasm, and indeed translocated from shoot to roots, via As-PC complexes, until deposited in the vacuole. Tonoplast function could be key to solving the mystery of hyperaccumulation. Alternatively, efficient detoxification (through complexation) and translocation to shoot vacuoles is the basis of this phenomenon. It could be that more than one adaptation is required.

In any case, arsenic acting as a phosphorus analogue does not appear to be at the basis of the tolerance these ferns have to arsenic, or their ability to hyperaccumulate arsenic. Potential candidates for arsenite analogues are notable by their absence. It appears that arsenic is the selection pressure that led to arsenic hyperaccumulation. This raises more problems in trying to work out the logic of arsenic hyperaccumulation, given that the phenomenon is found in species not endemic to arsenic contaminated sites.

The ecological advantages conferred by hyperaccumulation, of any metal that exhibits this phenomenon, is still a mystery (Baker & Whiting, 2002). There are no studies published to date examining if these ferns hyperaccumulate arsenic in soils with low arsenic.

Concluding remarks

These are exciting times for those interested in metal(loid) tolerances in plants. It is not every day that a phenomenon such as arsenic hyperaccumulation is revealed. Many questions are raised by arsenic hyperaccumulation of ferns. The growing list of hyperaccumulators may help us to piece together the evolutionary origin of this phenomenon. Phylogenetic trees point clearly to where we should look for more hyperaccumulating ferns.