Arbuscular mycorrhizal fungi confer enhanced arsenate resistance on Holcus lanatus


Author for correspondence: A. Meharg Tel: +44 (0)1224 272264 Fax: +44 (0)1224 272703 Email:


  • • The role of arbuscular mycorrhizal fungi (AMF) in arsenate resistance in arbuscular mycorrhizal associations is investigated here for two Glomus spp. isolated from the arsenate-resistant grass Holcus lanatus.
  • • Glomus mosseae and Glomus caledonium were isolated from H. lanatus growing on an arsenic-contaminated mine-spoil soil. The arsenate resistance of spores was compared with nonmine isolates using a germination assay. Short-term arsenate influx into roots and long-term plant accumulation of arsenic by plants were also investigated in uninfected arsenate resistant and nonresistant plants and in plants infected with mine and nonmine AMF.
  • • Mine AMF isolates were arsenate resistant compared with nonmine isolates. Resistant and nonresistant G. mosseae both suppressed high-affinity arsenate/phosphate transport into the roots of both resistant and nonresistant H. lanatus. Resistant AMF colonization of resistant H. lanatus growing in contaminated mine spoil reduced arsenate uptake by the host.
  • • We conclude that AMF have evolved arsenate resistance, and conferred enhanced resistance on H. lanatus.


The evolution of arsenate (the dominant plant-available arsenic species in aerobic soils) resistance has been determined in a range of higher plants (Meharg, 1993, 1994; Fitter et al., 1998). The genetic basis of this adaptation involves a major gene in all species investigated to date (Watkins & Macnair, 1991; Macnair et al., 1992), with this gene encoding for suppression of high-affinity phosphate/arsenate transport in grasses (Meharg & Macnair, 1992a,b). Arsenate is a phosphate analogue and is transported across the plasma membrane via high-affinity phosphate transporters (Meharg & Macnair, 1990, 1992a). The most detailed studies on the mechanisms of arsenate resistance have been conducted on the grass Holcus lanatus (Meharg & Macnair, 1990, 1992a; Hartley-Whitaker et al., 2001b). Constitutive suppression of high-affinity arsenate/phosphate transport results in arsenate resistance in H. lanatus by decreasing the rate of arsenate uptake (Meharg & Macnair, 1992b). This enables the plants to reduce arsenate to arsenite and complex this with phytochelatins (Hartley-Whitaker et al., 2001b). In nonresistant plants, arsenate influx is at too great a rate to achieve detoxification by phytochelatin (PC)-complexation, leading to considerable oxidative stress, which is not observed in resistant plants (Hartley-Whitaker et al., 2001a).

Many plant species adapted to grow on metalliferous soils are colonized by mycorrhizal fungi in these environments (Meharg & Cairney, 1999). Experiments on arsenate resistance in grasses have, to date, not considered the role of arbuscular mycorrhizal fungi (AMF), which are known to colonize H. lanatus roots on highly arsenic-contaminated mine spoils (Meharg & Cairney, 1999). Meharg & Cairney (1999) outlined the role that mycorrhizal fungi may play in adaptations of plant hosts to metal(loid)-contaminated soils. They proposed that mycorrhizal fungi need to be resistant, like their hosts, to the contaminated environment and: (1) that they may just be fulfilling their normal role in the association, not conferring enhanced resistance to the host; or (2) that they also further enhance resistance. Arbuscular mycorrhizal fungi are known to benefit the P nutrition of many plant species by increasing phosphate capture from soil via the enhanced surface area of the fungal hyphae emanating from the mycorrhizas (Smith & Read, 1997). Because AMF have a role in acquiring P for mycorrhizal associations, this may provide a paradox in achieving arsenate resistance as the AMF may potentially enhance arsenate uptake.

In the only detailed physiological explanation of the role of mycorrhizal fungi in adaptation of its hosts to toxic metal(loids), Sharples et al. (2000a,b) found that the ericoid mycorrhizal fungus Hymenoscyphus ericae was responsible for the colonization of arsenic-contaminated mine spoils by its host Calluna vulgaris. Unlike grasses, C. vulgaris does not achieve arsenate resistance through suppression of arsenate uptake (the mechanism of arsenate resistance in uninfected C. vulgaris has not yet been elucidated). However, its fungal symbiont H. ericae is capable of arsenate exclusion and therefore enhances the resistance of C. vulgaris to arsenate. The fungus achieves this resistance by reducing cellular arsenate to arsenite, and then effluxing the arsenite (Sharples et al., 2000a,b), which is the mechanism used to achieve arsenate resistance in bacteria and yeasts (Rosen, 1999). This is an elegant solution as the fungus can obtain P for the association at the exclusion of arsenate.

In this study, we investigated the role of AMF in arsenate resistance in arbuscular mycorrhizal associations. First, the resistance of AMF spores from arsenic mine spoils to arsenate was investigated. Next, the role of arsenate-resistant and nonresistant strains of the arbuscular mycorrhizal fungus Glomus mosseae in the short-term uptake kinetics of arsenate by arsenate-resistant and nonresistant H. lanatus roots was ascertained. Finally, the long-term effects of AMF colonization on arsenate resistance of H. lanatus growing on arsenic-contaminated mine spoil soil was determined.

Materials and Methods

Mine soil characterization and morphological biodiversity of AMF

The field site was the Devon Great Consol mine, South-west of England (map ref. SX 427 736). This mine was the worlds largest producer of arsenic at the end of the nineteenth century and its soils are highly arsenic-contaminated (Meharg et al., 1994a). Adaptation of flora of this site to arsenic has been well characterized (Watkins & Macnair, 1991; Meharg & Macnair, 1992c; Hartley-Whitaker et al., 2001a,b).

Root samples of H. lanatus were taken at five selected locations at the mines. The root system of each plant was excavated and the roots sampled were placed in plastic bags and brought back to the laboratory. Field root colonization level (percentage of root length colonized) was approx. 40%, ascertained by clearing and staining (described later).

Trapping of AMF and culture purification

A genotype of H. lanatus, resistant to arsenate, was used for the propagation of AMF. This genotype was from the Devon Great Consols plant population. Mine root segments were used as inoculum. The culture growth substrate comprised a coarse sand and polluted soil mixture (1 : 1 v : v) as suggested by Morton et al. (1993). Two cycles of successive cultures were performed to reveal the species richness of AMF in the soils (Stutz & Morton, 1996). All trap cultures were maintained for 4 months in a plant growth room with a 16-h photoperiod using supplementary lighting; the temperature was 24°C during the day and 12°C at night. Plants were watered every 2 d with distilled water and weekly with modified, phosphate free, Rorison’s nutrient solution (25 mm KNO3; 10 mm Ca(NO3)2; 10 mm MgSO4; 250 µm H3BO3; 4 µm ZnSO4; 1.5 µm CuSO4; 80 nm (NH4)No7O24; 50 µm MnCl2; 0.5 mm NaCl; 10 µm ethylenediamine-tetraacetic acid (FeEDTA)). After 4 months of growth, watering was terminated and each trap culture was allowed to dry in situ for up to 2 wk. Plant tops were removed, and pot contents were placed intact into polyethylene bags and stored at 4°C for 1 month. Roots and substrate were mixed thoroughly and a new trap culture was initiated. The same host (arsenate-resistant H. lanatus seedlings) for another 4-month growth period.

Separation of mycorrhizal spores from soil detritus was conducted by sieving and decanting (Gerdermann & Nicholson, 1963). Extracted spores were selected under a stereoscopic microscope (Nikon SMZ-2T) and permanent slides of spores found were prepared using polyvinyl alcohol-lacto-glycerol (PVLG) (Brundrett et al., 1994). Taxonomic identification was based on spore morphology under a dissecting microscope and on diagnostic slides examined under a compound microscope at ×100–×400 magnification.

Pure cultures of the mine fungal isolates identified were obtained by placing a single spore of an isolate onto a young root of Plantago lanceolata growing in sterile, inert, attapulgite clay (Agsorb 8/16, Oil-Dri Ltd, Wisbech, UK.). To culture the mine AMF isolates under polluted conditions, a subculture was prepared containing the attapulgite clay mixed 100 : 1 with γ-irradiated (10 kGy) mine polluted soil. Nonmine fungi were propagated on P. lanceolata in attapulgite clay medium. All fungal cultures were maintained under the conditions outlined above. Weekly watering was done with the phosphate-free nutrient solution.

AMF arsenate resistance

An open-filter method was used to assess the germination rates of AMF spores in arsenate-dosed soils (Brundrett & Saito, 1995). Two strains of AMF isolated from the mines, G. mosseae BEG132 and Glomus caledonium BEG133 were used along with G. mosseae BEG25 and Gigaspora rosea BEG111 as reference strains from uncontaminated soils (cultured in attapulgite clay on P. lanceolata). Fifteen apparently viable (cytoplasm evident under transmitted light) spores were placed on a 24-mm diameter 0.45-µm filter. Three replicate filters were prepared for each treatment and placed in Petri dishes containing 50 g of attapulgite clay maintained at 80% of its maximum water-holding capacity. The attapulgite clay was dosed with arsenate (0, 50 and 220 mg kg−1 dry wt) using solutions prepared from NaH2AsO4. The Petri-dishes were then sealed with Parafilm and incubated for 6 wk, following which spore germination was assayed under a dissecting microscope at a magnification of ×160. Germination was counted as positive when a germ-tube greater than the diameter of the spore was visible (Dodd & Jeffries, 1989).

Kinetic studies

The resistant and nonresistant genotypes of H. lanatus were grown from tillers in 9-cm pots filled with attapulgite clay and fertilized weekly with the phosphate-free nutrient solution. Inoculum of AMF was placed into the hole created for planting before the tiller was inserted. The inoculum comprised 50 g of P. lanceolata roots colonized with either a nonmine isolate of G. mosseae BEG25 or a mine-isolated strain of G. mosseae BEG132. Uninoculated treatments received 50 g of nonmycorrhizal P. lanceolata roots.

After 4 months’ growth, roots were gently harvested by removing adhering attapulgite clay and washing them in ultrapure water. Immediately following washing, roots were excised at the node and a subsample of the roots was used to characterize arsenate uptake kinetics as follows. Roots were incubated for 20 min in an aerated solution containing 10 mm 2-(N-morpholino) ethanesulfonic acid (MES) and 0.5 mm Ca(NO3)2 adjusted to a pH of 5 (pH adjusted using NaOH). Roots were then transferred into containers with the same medium amended with arsenate (concentration ranging from 0 to 1 mm) for a further 20 min Arsenate was added as Na2HAsO4. Following arsenate exposure, roots were incubated in an ice-cold, arsenate-free medium, containing the MES and calcium nitrate, but amended with 1 m Na2HPO4; the phosphate desorbs arsenate in the root ‘free space’. The arsenic content of the roots was determined by hydride generation-atomic absorption spectrometry (HG-AAS) following nitric acid digestion. The methodologies outlined for incubations and arsenic analysis are those reported for previous kinetic experiments on H. lanatus (Meharg & Macnair, 1992b).

Another subsample of the roots was taken and the percentage root length colonized by AMF estimated. Roots were carefully washed with distilled water, cut into 1 cm segments and processed for root colonization. Subsamples of 1 g of fresh roots were randomly taken from these samples dispersed in water. Roots were cleared and stained with acid glycerol Trypan Blue according to Koske & Gemma (1989). One-hundred stained root segments were mounted on slides in 50% aqueous glycerol (v/v) and examined under a compound microscope (Nikon Alphaphot 2 YS2) at ×100 magnification. The frequency of AMF structures was estimated by rating the presence of arbuscules, vesicles and hyphae in the stained root segments. The frequency of fungal structures was expressed as a percentage in the colonized roots. There were three replicates for each root subsample (total of 300 root segments per plant root sample).

Growth of H. lanatus in arsenic-contaminated soil in the presence and absence of AMF

Clover trap cultures, inoculated with H. lanatus rhizosphere soil from an area of the mine with the highest AMF diversity (determined from a previous survey) were prepared. Clover roots that had 90% colonization (species composition of AMF not quantified), were washed, and were cut into 1 cm segments and mixed. Of these colonized roots, 5 g was placed in the planting hole before transplanting the unrooted tillers. One 5-cm long unrooted tiller of resistant and nonresistant clonal genotypes of H. lanatus with two leaves were transplanted into each pot. Nonmycorrhizal treatments were amended with the same amount of colonized roots, but these roots were autoclaved for 30 min at 120°C before inoculation. A mixture of coarse sand (985 g) and polluted soil (15 g) was used as the test substrate. The sand was acid-washed and autoclaved. The polluted soil was γ-irradiated (10 kGy) to destroy fungi. Soil and acid-washed sand were well mixed and each pot was filled with 750 g of this mixture, which had a pH of 5.6 and the following elemental concentration (g−1 soil): arsenic (As) 353 mg, copper (Cu) 27 mg, iron (Fe) 4100 mg, manganese (Mn) 19 mg, phosphorus (P) 29 mg, as determined by inductively coupled plasma-optical emission spectrometry (ICP-OES) following nitric acid digestion. Pots with the transplanted tillers were placed randomly onto trays and maintained in a plant growth room for 4 months. Plants received a 16-h photoperiod with supplementary lighting (400 W Na vapour lamps); temperature was maintained at 24°C during the day and 12°C at night. Plants pots were gravimetrically adjusted with distilled watered every 2 d to maintain 80% water-holding capacity (WHC). The experiment was a complete factorial of two treatments of inoculation × two genotypes × four replicates in a completely randomized design.

After 16 wk of growth, the number of tillers produced per plant dry weights of shoots and roots were obtained. Arsenic and phosphorus content were determined in shoots and roots after acid digestion with concentrated nitric acid and analysed by ICP-OES. A subsample (0.5 g) was taken from each root system to evaluate the percentage of root length colonized by native AMF, as detailed earlier.


AMF arsenate resistance

A diverse AMF population, as determined by morphological examination of spores, was present on the arsenic mine spoil soils with representatives from the Glomaceae (G. mosseae, G. caledonium, G. claroideum, G. constrictum, G. intraradices, G. fasciculatum and two unidentified Glomus spp.), Acaulosporaceae (Acaulospora delicata, A. undulata and Entrophospora infrequens). Strains of G. mosseae BEG132 and G. caledonium BEG133 were cultured for further investigation.

Germination tests using G. mosseae BEG132, and G. caledonium BEG133 isolated from arsenic mine spoil soil and nonmine isolates of G. mosseae BEG25 and Gigaspora rosea BEG111 in arsenate-dosed soils showed that germination of G. caledonium BEG133 was stimulated in the presence of arsenate, while the germination of the mine isolated G. mosseae BEG132 was unaffected (Fig. 1). By contrast, germination of G. mosseae BEG25 and G. rosea BEG111 from uncontaminated environments was strongly inhibited by the presence of 50 mg kg−1 of arsenate. The mine isolates appear to be considerably more arsenate resistant than the nonmine isolates.

Figure 1.

The percentage germination of arbuscular mycorrhizal fungi (AMF) spores in soil dosed with different levels of arsenate. Circles, Glomus mosseae; squares, Glomus caledonium; triangles, Gigaspora rosea. Closed symbols, arsenic mine isolates; open symbols, reference collection isolates. Data points represent mean ± SE, n = 3.

Kinetic studies

The percentage of total root length of resistant and nonresistant H. lanatus colonized by G. mosseae BEG132 (from mine site) was 3 ± 1.4% and 9 ± 5.8%, respectively. Colonization of resistant and nonresistant H. lanatus by G. mosseae BEG25 (nonmine site) was12 ± 4% and 27 ± 6.1%, respectively. Low infection levels in this experiment may have been a consequence of the growth conditions. The resistant plants had much lower AMF colonization levels, regardless of the site of origin of the fungus, and G. mosseae BEG132 was less effective at colonizing roots compared with the nonmine fungus G. mosseae BEG25.

The mycorrhization of plants had a considerable effect on the short-term uptake kinetics of arsenate (Fig. 2). Noncolonized roots behaved as expected, with nonresistant H. lanatus roots taking up much more arsenate than the resistant H. lanatus roots. This has been observed consistently in other studies (Meharg & Macnair, 1992b). Colonization of nonresistant roots by nonmine AMF considerably suppressed arsenate influx, yet this fungus (BEG25) had little effect on resistant plants. Colonization of nonmine plants by AMF isolated from mine sites slightly decreased arsenate influx further, except at the highest arsenate concentration (1 mm) where influx was much lower, compared with the nonmine AMF-colonized roots. The influx of arsenate was decreased following colonization by mine AMF in resistant H. lanatus plants. Despite low colonization levels recorded for mine AMF in H. lanatus roots, of both resistant and nonresistant plants, arsenate influx was decreased to 50% and 30% of the nonmycorrhizal plants at 1 mm arsenate. Analysis of variance (anova) showed that plant resistance, AMF treatment and concentration were all highly significant (P < 0.001) (Table 1). The interaction between plant resistance and AMF origin was also highly significant (P < 0.001).

Figure 2.

Influx of arsenate by nonmine (a) and mine (b) Holcus lanatus either nonmycorrhizal (circles and solid lines), colonized by nonmine Glomus mosseae BEG25 (triangles and dash line) or mine Glomus mosseae BEG132 (squares and dot–dashed line). Data points represent mean ± SE, n = 3.

Table 1.  Analysis of variance table for the arsenate influx into uninfected, nonmine Glomus mosseae- and mine G. mosseae-infected experiment, (data which is presented in Fig. 2)
SourceDegrees of freedomSequential sum of squaresF-valueP
  1. Analysis was performed on ranked data (the untransformed data were not normally distributed) using the statistics package Minitab ver. 11 (State College, PA, USA). ‘Conc.’, arsenate concentrations at which arsenate influx was measured; ‘plant’, compares resistant and nonresistant plants; arbuscular mycorrhizal fungi (AMF), describes if the plants were uninfected, infected with nonmine G. mosseae, or infected with mine G. mosseae.

Conc.  4 4882445.4< 0.001
Plant  1 2612792.8< 0.001
AMF (arbuscular mycorrhizal fungi)  2 1789431.2< 0.001
Conc. × plant  4  1690 1.7  0.164
Conc. × AMF  8  5117 2.4  0.024
Plant × AMF  2  4612 8.3  0.001
Conc. × plant × AMF  8  2626 1.2  0.298
Error 86 23172  

The main-effects plots for this anova showed that resistant plants took up less arsenate than nonresistant plants, regardless of AMF infection treatment (Fig. 3). For AMF infection, regardless of resistance status of the plants, uninfected plants took up more arsenate than nonmine G. mosseae-infected plants, with mine G. mosseae-infected plants taking up still less. The AMF–plant interaction plot showed that for nonresistant plants, infection with nonmine G. mosseae reduced uptake, with infection by mine G. mosseae reducing uptake still further (Fig. 3). By contrast, infection with nonmine G. mosseae did not alter arsenate uptake in resistant plants, but infection with mine G. mosseae did further suppress arsenate uptake.

Figure 3.

Analysis of variance (anova) main effects and interaction plots for the data presented in Fig. 2 for the arsenate influx experiment, with the anova presented in Table 2. Concentration, arsenate concentrations at which arsenate influx was measured; plant, compares resistant and nonresistant plants; arbuscular mycorrhizal fungi (AMF), describes whether the plants were uninfected, infected with nonmine Glomus mossae or mine G. mosseae.

Analysis of the apparent kinetic parameters showed that in five out of the six treatments that concentration-dependent influx of arsenate was better described by Michaelis–Menten kinetics as opposed to a linear model, nonmycorrhizal mine plants being the exception, according to the r2 of the model fits (data not shown).

It must be remembered that removing the plant roots from the growth medium to conduct the short-term influx experiments will damage and partly remove some of the external AMF hyphae. If hyphae were not damaged or removed, uptake of arsenate may have potentially increased in the AMF infected treatments.

Long-term soil experiments

The nonresistant, nonmycorrhizal colonized plants failed to grow. Inoculation of resistant plants with a mixed mine AMF community and growth for 16 wk on γ-irradiated mine soil/acid washed sand significantly enhanced shoot growth and tiller production (both at the P = 0.01 level of significance using an unpaired t-test) – approximately twofold that with uninoculated plants (Table 2). There was no significant difference in root biomass. Inoculation of plants with AMF decreased (by threefold) arsenic accumulation in the shoot (P = 0.01), and to a lesser, though still significant extent (P = 0.05) in the root. Uninoculated plants accumulated significantly more P in their shoots, but the converse was true of the roots, where inoculated plants had a higher P status. In both roots and shoots, however, the P : As ratio was approximately threefold higher in inoculated plants than in uninoculated plants (P = 0.01). There was no colonization by AMF in the uninoculated treatment, and 28% root length colonization in the inoculated treatment.

Table 2.  Effects of inoculation of arsenate-resistant Holcus lanatus when inoculated with its arbuscular mycorrhizal fungi community from its site of origin and grown on arsenate contaminated soil
TreatmentTillers per plantD. wt (g)As (mg kg−1 dry wt)P (mg kg−1 dry wt)P : As ratio
  • *

    Significant at the P = 0.05 level;

  • **

    significant at the P = 0.001 level using an unpaired t-test.

Uninoculated 50.78 26.11146 43
Inoculated101.51  6.7 824123
LSD 4.6**0.37** 14**  38** 38**
Uninoculated 0.13836 275  0.3
Inoculated 0.15567 447  0.8
LSD 0.11194* 205**  0.3*


We have demonstrated here for the first time that AMF can evolve arsenate resistance. Previous studies have ascertained that AMF can evolve cadmium and zinc resistances (Gildon & Tinker, 1981, 1983; Weissenhorn et al., 1994; Weissenhorn & Leyval, 1995). The evolution of arsenate resistance in AMF is interesting since arsenate is a phosphate analogue, and the proposed role that AMF have in obtaining phosphate from soils for their plant hosts (Smith & Read, 1997).

Evidence is presented that the AMF strains tested here, regardless of their arsenate resistance, reduced arsenate influx in both arsenate resistant and nonresistant H. lanatus plants. This strong suppression occurred even though plants had low levels of AMF colonization. This is in contrast to a comparable experiment on mycorrhizal and nonmycorrhizal tomato roots, where short-term phosphate influx was enhanced in AMF-colonized roots (Cress et al., 1979). It has been often noted that AMF colonization can improve P nutrition in pot experiments, whereas little difference is observed in colonized native plants under temperate field conditions (Fitter, 1991). Francis & Read (1995) outlined that AMF can have a neutral effect with respect to grass fitness. There is now an increasing body of evidence (Koide, 2000) highlighting an effect of AMF on plant reproduction, growing in natural systems, in P-limiting soils. Natural populations of AMF had no significant effect on the P uptake of field-grown Vulpia ciliata (Carey et al., 1982) but did increase seed production. Furthermore, seeds produced by mycorrhizal Avena fatua were lighter than those of nonmycorrhizal plants but contained more P (Lu & Koide, 1991; Koide & Lu, 1992).

The results presented here show that AMF colonization of H. lanatus suppressed the high-affinity phosphate transport system. This suppression may explain why AMF may not benefit the P nutrition of certain wild plants (Fitter, 1991). AMF can enhance the P-absorbing area of the root surface but in our experiments reported here, it appeared to downregulate phosphate transport in the host at least. It must be stressed that the AMF colonization levels in the experiments presented here were low. Under field conditions higher rates of colonization may improve the phosphate scavenging potential of the association, through increased hyphal biomass in soil, assuming that the AMF phosphate transporters are not downregulated in a similar manner to the plants. In our experiments, we could not distinguish between the fungal and plant contribution to arsenate uptake.

Because these kinetic studies were conducted on excised roots where external AMF hyphae may have been damaged or partly removed, the results of this experiment should be interpreted with this in mind, as more healthy external hyphae may have enhanced arsenate acquisition. Even though plant phosphate transport is downregulated, phosphate transport in AMF hyphae emanating from roots might not be downregulated. Nonresistant plants maintained higher levels of AMF than resistant plants for both arsenate-resistant and nonresistant AMF strains, implying that AMF may have a greater role in phosphate acquisition for nonresistant plants compared with resistant plants.

The mechanism by which AMF colonization downregulates phosphate/arsenate uptake in our experiments is not known. Molecular evidence shows that Mt4 cDNA is a phosphate-starvation inducible gene, which is downregulated in response to phosphate exposure and to AMF colonization (Burleigh & Harrison, 1999). The authors of that study hypothesized that Mt4 encoded for a signalling molecule involved in a phosphate starvation response. Furthermore, Liu et al. (1998) and Burleigh (2001) showed cDNA of MtPT2, which encodes for the high-affinity phosphate transporter, is downregulated in roots in response to AMF colonization.

Evidence for the mechanisms by which AMF downregulates arsenate/phosphate transport is found with the higher P concentrations in colonized roots compared with nonmycorrhizal roots when plants were grown in (contaminated) soil for 16 wk. The higher P levels in the root may have a role in regulating the kinetics of phosphate/arsenate uptake. Shoot phosphate levels, however, were lower and it is thought that shoot P status has an overall role in regulating P accumulation and metabolism in plants (Burleigh & Harrison, 1999). Transport of P to the shoots was decreased in plants colonized by AMF. The AMF may be the direct causal agent of decreased phosphate/arsenate transport through release of signalling molecules, rather than just simply altering the P status of the host. This hypothesis remains untested. The reduction in shoot P in colonized plants in the soil experiment presented here was also observed for field-grown arsenate resistant and nonresistant plants colonized by AMF when compared with nonmycorrhizal plants (Wright et al., 2000).

Arsenate resistance in H. lanatus is intriguing since, unlike other metal(loid) resistances, the gene for resistance and downregulation of phosphate/arsenate resistance is polymorphic in all populations investigated from uncontaminated (i.e. low As levels) soils (Meharg & Macnair, 1992c; Meharg et al., 1993, 1994b). There is evidence that this polymorphism is also present in other species, including both monocotyledons and dicotyledons (Fitter et al., 1998). It has long been considered that the rate-limiting step in phosphate uptake by plant roots is not the rate of transport across the plasma membrane, but diffusion of phosphate to the root surface since phosphate mobility in soil is low (Nye, 1977). This may partly explain why plants with suppressed phosphate uptake systems can compete effectively with those that have inducible uptake systems. Our present results indicate that AMF colonization of H. lanatus may make the nonresistant phenotype more like the resistant phenotype by making the differences in phosphate uptake kinetics less pronounced. Wright et al. (2000) argue that polymorphism is present in H. lanatus populations as a result of resistant and nonresistant plants having differing reproductive strategies, rather than differences in phosphate uptake per se. Resistant plants put more effort into reproductive growth while nonresistant plants put more effort into vegetative growth. However, it is thought that the differences in reproductive strategy in resistant and nonresistant phenotypes of H. lanatus is fundamentally related to P nutrition, with nonresistant plants being more P efficient than resistant plants (Wright et al., 2000).

It appears that AMF and their plant hosts have coevolved in arsenate-contaminated environments. While H. lanatus does not require its fungal symbiont to achieve arsenate resistance, its ability to grow on contaminated soils is greatly enhanced by being mycorrhizal. This is not because of enhanced P nutrition in mycorrhizal plants, as they had lower shoot P level. They did, however, have higher root P concentrations. It appears that AMF benefit the host on mine soils through a considerable reduction in arsenic uptake, particularly in the portion of arsenic translocated to the shoot. AMF colonization further reduces arsenate uptake compared with nonmycorrhizal plants, resulting in enhanced resistance. Furthermore, AMF have a wider role in the health of plants than just arsenate/phosphate uptake, including plant water relations and pathogen resistance of plants (Newsham et al., 1995). There remains the possibility that other benefits result from the mycorrhizal status, which may have helped to improve the health of colonized H. lanatus growing on mine spoil.

In the only comparable experiments on arsenate resistances in mycorrhizal associations, Sharples et al. (2000a,b, 2001) found that the ericoid mycorrhizal fungus H. ericae had a direct role in the arsenate resistance of its host C. vulgaris. This was through the ability of the fungus to efflux arsenite, while at the same time assimilating phosphate. Hymenoscyphus ericae has itself, similar to the AMF fungi under study here, evolved arsenate resistance. This is through reduced uptake of arsenate. The mechanistic basis of how arsenate-resistant AMF achieves arsenate resistance, and how it achieves reduced arsenate uptake by its resistant host is unclear. The fungus may have suppressed phosphate/arsenate uptake across its plasma membrane, which is the mechanism of arsenate resistance employed by most higher plants investigated to date (Meharg, 1993, 1994), or through enhanced efflux of arsenate, which is the mechanism employed by arsenate-resistant H. ericae (Sharples et al., 2000a,b). Unlike many ericoid mycorrhizal fungi, AMF cannot be cultured in the absence of the plant, and physiological studies on transport are restricted to germ tubes using radio-isotopes (Thompson et al., 1990). From the results presented here, it appears that one of the mechanisms employed by resistant AMF to reduce arsenate toxicity to its host is through further suppressing arsenate uptake by the association.


Some AMF have coevolved with their hosts resistance to arsenate, and colonization of H. lanatus by AMF on arsenate-contaminated soils further enhances the resistance of the host. This is achieved, at least partly, by the ability of arsenate-resistant AMF to further downregulate arsenate uptake into the host. The results from this study showed that nonresistant AMF also downregulated phosphate/arsenate transport in both arsenate-resistant and nonresistant hosts. This has major implications for the phosphorus nutrition of plants, and may explain why AMF generally do not enhance or suppress P acquisition by some wild plants.


Anne Dudley is thanked for her assistance in arsenic analyses. CONACyT, Mexico supported CG-C with a studentship.