Microbial Community Shifts in Response to Acid Mine Drainage Pollution Within a Natural Wetland Ecosystem

Abstract

Natural wetlands are known to play an important role in pollutant remediation, such as remediating acid mine drainage (AMD) from abandoned mine sites. However, many aspects of the microbiological mechanisms underlying AMD remediation within wetlands are poorly understood, including the role and composition of associated microbial communities. We have utilized an AMD-polluted river-wetland system to perform rRNA sequence analysis of microbial communities that play a role in biogeochemical activities that are linked to water quality improvement. Next-generation sequencing of bacterial 16S rRNA gene amplicons from river and wetland sediment samples identified variation in bacterial community structure and diversity on the basis of dissolved and particulate metal concentrations, sediment metal concentrations and other water chemistry parameters (pH and conductivity), and wetland plant presence. Metabolic reconstruction analysis allowed prediction of relative abundance of microbial metabolic pathways and revealed differences between samples that cluster on the basis of the severity of AMD pollution. Global metabolic activity was predicted to be significantly higher in unpolluted and wetland sediments in contrast to polluted river sediments, indicating a metabolic stress response to AMD pollution. This is one of the first studies to explore microbial community structure dynamics within a natural wetland exposed to AMD and our findings indicate that wetland ecosystems play critical roles in maintaining diversity and metabolic structure of sediment microbial communities subject to high levels of acidity and metal pollution. Moreover, these microbial communities are predicted to be important for the remediation action of the wetland.

Keywords: 16S rRNA gene amplicon sequencing; acid mine drainage; bacterial community; metabolic prediction; metal pollution; microbial ecology; wetlands.

FIGURE 1

Sample sites within the Parys Mountain river catchment in Anglesey, Wales, United Kingdom. Sites on the southern Afon Goch river are marked S1–S3, with the wetland areas shaded in green. Sites on the northern Afon Goch river are marked N1–N4. Representative photographs of the samples sites along each river taken during times of sampling are shown. NA is the location of the adit that discharges AMD to the river.

FIGURE 2

Water chemistry [(A) pH and (B) conductivity] and total metal concentration in water [(C) dissolved and (D) particulate] and (E) sediment samples. Data are pooled from triplicate analyses taken on four to seven sampling occasions (June 2010, July 2010, November 2011, August 2013, October 2013, March 2014, and October 2014). Boxes show the minimum and maximum values and the line within the boxes shows the median values. Boxes that do not share lowercase letters are significantly different (p < 0.05) as determined by one-way ANOVA. Data for individual metals (Al, As, Cd, Cu, Fe, Mn, Pb, and Zn) are shown in Supplementary Figure S1.

FIGURE 3

Discrimination of sites on the basis of physicochemical parameters. Hierarchical cluster analysis (A) and PCA ordination plot (B) illustrating the discrimination between sample sites according to environmental properties and their correlation with each environmental factor analyzed. In (A), sites with non-significant clustering (p < 0.05) as determined by SIMPROF are indicated with red lines.

FIGURE 4

Relative abundance of bacterial taxa following OTU taxonomic assignment. All assigned taxa at phylum level where possible (A) and Proteobacteria shown to the highest resolution possible down to species level where possible (B) for each sample site. Selected taxa of high abundance in multiple samples are labeled, with ranges of relative taxa abundance given in parentheses.

FIGURE 5

Discrimination of sites on the basis of bacterial community structure. (A) Hierarchical clustering of sample sites based on species similarity. Sites with non-significant clustering (p < 0.05) as determined by SIMPROF are indicated with red lines. (B) Two-dimensional NMDS plot of sites based on species similarity showing a separation between unpolluted sites (UW, S3, N1, and N2), polluted sites (S1, S2, N3, and N4), and the highly polluted mine adit site (NA) on the basis of NMDS1. Assigned taxa assemblage is shown (gray circles) and the top five taxa with the most contribution (% contribution) to site assemblage based on SIMPER analysis are indicated and listed in red.

FIGURE 6

Metabolic prediction analysis and discrimination of sites on the basis of metabolic potential. (A) Hierarchical clustering of sample sites based on metabolic pathway similarity. Sites with non-significant clustering (p < 0.05) as determined by SIMPROF are indicated with red lines. Heat maps showing the predicted abundance of all prokaryotic metabolic pathways (B) and selected metabolic pathways related to element cycling (C).

FIGURE 7

Predicted changes in enzyme abundance in response to AMD pollution. Scatter plots of Enzyme Commission (EC) enzyme reaction classes showing the relationships between wetland (southern Afon Goch) and non-wetland (northern Afon Goch) log₂ fold-change values for polluted (site S1 or NA) versus unpolluted (site UW or N1) sites. Each dot corresponds to log₂ fold-change in abundance of a predicted enzyme reaction. The linear regression fit line is plotted.

FIGURE 8

Predicted changes in enzyme abundance in response to the presence of a wetland. (A) Scatter plots of EC enzyme reaction classes showing the relationships between wetland (southern Afon Goch) and non-wetland (northern Afon Goch) log₂ fold-change values for pollution source (site S1 or NA) versus downstream sites (site S3 or N4). Each dot corresponds to log₂ fold-change in abundance of a predicted enzyme reaction. The linear regression fit line is plotted. (B) Scatter plot for all enzyme classes. The red-shaded quadrant indicates enzymes that are significantly increased in abundance specifically at site S3 but not at site N4, relative to the pollution source sites. The enzymes of each class that are increased in abundance at site S3 are shown (right plot).