(515p) A Systems Biology Approach To Acid Mine Drainage Remediation

Acid mine drainage (AMD) is a major environmental and public health issue throughout the world due to its deleterious effects on humans and the environment, with treatment costs estimated in the tens of billions of dollars worldwide. The oxidation of the sulfide minerals exposed to air and water during mining activities creates increased levels of sulfate, acidity, and Fe(II) in the drainage. The increased acidity also mobilizes toxic trace metals, resulting in contamination of both the runoff and downstream waters. One attractive means of passively treating AMD is through sulfate-reducing bioreactors (SRBRs). These bioreactors contain a solid organic substrate, typically wood chips or compost, which provides a slow-release carbon substrate to support a complex anaerobic microbial community that drives remediation. A systems biology approach offers great potential to improve the design and performance of microbial-based SRBRs remediating AMD. In this approach we treat the ecological community as a ?meta-organism' and study its components and their interactions to better understand how the network behaves as a whole. In order to begin modeling this ecological community-scale network, several questions need to be answered: Which microorganisms are present in the community? What are the important pathways? What are the rate-limiting steps? What information is available about pathway fluxes? Lab- and field-scale studies were used to investigate these questions and to shape a conceptual model of the SRBRs.

To date, the microbial community of SRBRs has been represented as a carbon flow-based network consisting of four types of microorganisms. Polysaccharide-degrading species hydrolyze the organic substrate into short chain biopolymers (e.g., cellobiose, glucose, and mannose) that are then converted to alcohols and low molecular weight organic acids by the fermentative species. These small organic molecules then act as electron donors for the SRB, which produce sulfide and bicarbonate that remediate the AMD by forming metal sulfides and buffering the acidity, respectively. Methanogens compete with SRB for electron donors (e.g., hydrogen and acetate).

Short-term batch tests supplemented with cellobiose, glucose, and organic acids were performed to verify the carbon flow and to explore the general kinetic properties of the network. Material from two columns simulating SRBRs treating AMD was used as substrate for the batch anaerobic microcosms. The results also indicate that methanogenic activity was low and that competition between SRB and methanogens did not limit SRB activity. Cellulolytic bacteria were not substrate-limited at any time. However, fermenters experienced substrate limitation immediately and SRB after four weeks, indicating a dependence of both groups on the upstream populations to provide substrates. These results further validated the importance of the microbial community and pointed towards cellulose degradation as the rate-limiting step.

The importance of the microbial inocula on the remediation of AMD was further explored with a column study. Columns inoculated from two sources, dairy manure (DM) and acclimated microcosm material from the anaerobic microcosms studies (SRC), and uninoculated columns (U) were fed a simulated mine drainage and compared on the basis of pH neutralization and removal of metals and sulfate. Removal of metals and sulfate in the SRC columns were higher than either the DM or the U columns, which both performed the same. The microbial community of the SRC columns was the first to reach a pseudo-steady state based on denaturing gradient gel electrophoresis (DGGE) analysis targeting the V3 region of the 16S rRNA gene. A higher proportion of the DGGE band DNA sequences were related to microorganisms that carry out cellulose degradation in the SRC columns compared to the DM and U columns. Quantitative polymerase chain reaction (Q-PCR) revealed a consistently higher proportion of SRB of the genus Desulfobacterium in the SRC columns. This initial analysis demonstrated the importance of inocula on performance, and highlighted the need for additional inocula screening.

Further batch tests were conducted to screen several inocula and identify those with key capabilities for AMD remediation. Five inocula were screened: (1) dairy manure (DM), (2) anaerobic digester sludge (ADS), substrate from two field scale sulfate-reducing bioreactors (3) Luttrell (LUTR) and (4) Peerless Jenny King (PJK), and (5) material from the SRC columns (ACC). The microbial communities were characterized over 14 weeks using DGGE targeting the V3 region of the 16S rRNA gene and Q-PCR targeting the SRB genera Desulfovibrio and Desulfobacterium. Sulfate and zinc removal and pH neutralization were used as indicators of performance. The best performing cultures (inoculated with the LUTR, PJK, and DM materials) had in common the presence of all the microorganisms necessary for the desired functioning of SRBRs (polysaccharide degraders, fermenters, and sulfate reducers) as well as a relatively high proportion of one or both of the SRB quantified by Q-PCR. The results further demonstrated that the type of inoculum influences performance and provided insight into key aspects of successful inoculum composition.

The knowledge gained from the laboratory tests was then applied to study the communities of the two functioning field-scale SRBRs that served as inocula in the batch screening study. Analysis of the two field communities again revealed the three main groups of bacteria (cellulose degraders, fermenters, and SRB) were detected at both sites. In addition to the three main groups, a high diversity of species covering a broad phylogeny were also detected at both sites, including denitrifiers, rhizosphere microorganisms, acetogenic bacteria, metal-reducing species, and AMD-generating bacteria. Phylogenetic analysis of cloned 16S rRNA genes revealed no significant differences between the overall microbial communities, so the analysis scope was narrowed to the SRB community by targeting the adenosine 5'-phosphosulfate reductase (apsA) gene. Significant differences between the two bioreactors based on the apsA gene analysis. LUTR was dominated by uncultured SRB most closely related to Desulfovibrio spp., while PJK was dominated by Thiobacillus spp. The fractions of two genera of SRB, Desulfovibrio and Desulfobacterium, were also higher at LUTR compared to PJK as determined Q-PCR. The differences between the SRB communities were attributed to oxygen exposure at the PJK site due to varying flow conditions.

The field analyses revealed a much more complex microbial network than expected based on the laboratory studies. Analysis of the community targeting the apsA gene revealed information about the community complimentary to the 16S DNA analysis. With this in mind, a more function oriented approach was implemented that targeted genetic markers related to other functions of interest in the community, such as cellulose degradation and fermentation. Additionally, an approach that targeted only active species in the community was applied.

Anaerobic cellulose degraders were analyzed by targeting the glycoside hydrolase families 5, 9, and 48, and significant spatial variations within the PJK SRBR were found. Fermenters and methanogens are currently being analyzed by targeting hydrogenase and methyl coenzyme M reductase genes, respectively.

The active microbial population present at PJK was characterized using active community profiling (ACP). ACP involves monitoring the composition and activity of a mixed microbial culture via comparative measurements of 16S rRNA and rDNA using capillary electrophoresis single-strand conformational polymorphism (CE-SSCP), as the cellular ratio of total RNA to DNA is known to be proportional to growth rate in a variety of organisms. A more complex and diverse microbial community was observed in the rDNA profiles than in the rRNA profiles, indicating that many of the species present in the system are not active. Species representative of the three main groups (cellulose degraders, fermenters, and SRB) were more dominant in the rRNA profiles, verifying their active role in the remediation process. Several species with unknown function were also prevalent in the rRNA profiles. These species may be actively involved in AMD remediation, or may be particularly suited to the environment of the bioreactor.

The results of these laboratory- and field-scale studies provided experimental information used to update and improve the current network model. The community profiling identified many microorganisms that do not fall under the three main categories currently associated with SRBRs. Based on these additional network components, we propose an electron donor/acceptor network as the new basis for a conceptual model. The network offers an improved experimentally-based scaffold to evaluate a multitude of aspects of the microbial community associated with AMD remediation. The identification and evaluation of potential electron-transfer bottlenecks, a framework for the design of site-specific microbial inocula, and the ability to easily integrate additional ?omics'-based information are a few of the potential applications of this systems biology-developed model.