Arbuscular mycorrhizal colonization has little consequence for plant heavy metal uptake in contaminated field soils

Abstract

The factors affecting plant uptake of heavy metals from metalliferous soils are deeply important to the remediation of polluted areas. Arbuscular mycorrhizal fungi (AMF), soil-dwelling fungi that engage in an intimate exchange of nutrients with plant roots, are thought to be involved in plant metal uptake as well. Here, we used a novel field-based approach to investigate the effects of AMF on plant metal uptake from soils in Palmerton, Pennsylvania, USA contaminated with heavy metals from a nearby zinc smelter. Previous studies often focus on one or two plant species or metals, tend to use highly artificial growing conditions and metal applications, and rarely consider metals' effects on plants and AMF together. In contrast, we examined both direct and AMF-mediated effects of soil concentrations on plant concentrations of 8-13 metals in five wild plant species sampled across a field site with continuous variation in Zn, Pb, Cd, and Cu contamination. Plant and soil metal concentration profiles were closely matched despite high variability in soil metal concentrations even at small spatial scales. However, we observed few effects of soil metals on AMF colonization, and no effects of AMF colonization on plant metal uptake. Manipulating soil chemistry or plant community composition directly may control landscape-level plant metal uptake more effectively than altering AMF communities. Plant species identities may serve as highly local indicators of soil chemical characteristics.

Keywords: Lehigh Gap Nature Center; Palmerton Zinc Superfund Site; arbuscular mycorrhizal fungi; heavy metals; hyperaccumulation; plant-soil feedback; pollution; restoration; soil chemistry.

Figure 1

Conceptual diagram of the relationships examined in this study. Soil metal concentrations may relate to plant metal concentrations directly, or indirectly via mycorrhizal fungal colonization.

Figure 2

Interspecific differences in plant (A–C), soil total extractible (D–F), and soil exchangeable (G–I) concentrations of the primary pollutants Zn (A, D, G), Cd (B, E, H), and Pb (C, F, I). Note that the vertical axes are different for each panel. Species are abbreviated as follows: AA, Ageratina altissima (Asteraceae); AP, Agrostis perennans (Poaceae); DF, Deschampsia flexuosa (Poaceae); ES, Eupatorium serotinum (Asteraceae); MP, Minuartia patula (Caryophyllaceae). Families are color-coded as follows: Asteraceae, blue; Poaceae, red; Caryophyllaceae, gray.

Figure 3

Interspecific differences in plant root colonization by AMF (A) and in the integrative soil variables pH (B), CEC (C), and base saturation (D). Species and families are abbreviated and color-coded as in Fig. 2.

Figure 4

(A) Plant metal concentration profiles, (B) soil total extractible metal concentration profiles, and (C) soil exchangeable metal concentration profiles each clearly segregate plant species in CAP ordination space (P < 0.001 for each). Species and families are abbreviated and color-coded as in Fig. 2. (D–F) Contributions of individual metal concentrations, LOI, and AMF colonization (D only) to the ordination spaces in (A–C), respectively. Percentages on axis labels show the amount of constrained variation accounted for by individual CAP axes. Species DF does not appear in (C) because of insufficient sample size.

Figure 5

Total extractible soil Zn (A) and Cd (B) concentrations decrease significantly with distance to the nearer zinc smelter (Zn: P < 0.001, R² = 0.249; Cd: P < 0.01, R² = 0.076), but total extractible Pb and Cu, and exchangeable Zn, Cd, and Pb concentrations do not change with distance to the smelter. Exchangeable Cu was not included in this analysis because most measurements were below the detection limit of the ICP-OES. Points represent individual soil samples, and lines are best-fit lines.

Figure 6

Proportion of variation in (A) exchangeable metal concentrations, (B) exchangeable contaminant concentrations, (C) total extractible metal concentrations, and (D) total extractible contaminant concentrations accounted for by spatial proximity of soil samples (“Prox”), plant species identity (“Sp”), both, or neither. Variance components within a graph may not sum to 1 due to rounding error or production of small negative variance estimates in the overlap region of the Venn diagram. Negative values here are artifacts of subtraction and do not indicate major problems with the model (Oksanen et al. 2013).