Geochip-based functional gene analysis of anodophilic communities in microbial electrolysis cells under different operational modes.

A microbial electrolysis cell (MEC) is a bioelectrochemical system that can produce hydrogen from acetate at high hydrogen recoveries, but the composition and structure of the microbial communities in this system have not been extensively studied. We used a high throughput metagenomics technology (GeoChip) to examine the microbial community functional structure in MECs initially operated under different conditions. We found that startup conditions had little or no effect on reactor performance in terms of Coulombic efficiencies (CEs) and COD removals, somewhat greater effects on CO(2) and CH(4) production, and very large effects on hydrogen production. Hydrogen yields were generally higher for reactors that were always operated as MECs than those initially operated as MFCs. Hydrogen yields were nine times larger for MEC reactors with an applied voltage of 0.7 V (64%∼80% efficiencies) than 0.3 V (<7-8%), independent of startup conditions. GeoChip analysis revealed that the functional and phylogenetic diversity of MEC microbial communities after 4 months was quite high despite the use of only a single substrate (acetate). MECs with the largest hydrogen yields had the highest microbial diversity. Multivariate analyses showed that communities that developed in the MECs were well separated from those present under startup conditions, indicating reactor operation altered microbial community composition. Community shifts based on a Mantel test were significantly related to CEs and COD removals in these reactors, suggesting that there were significant changes in microbial community composition as a result of conditions that affected MEC performance. Common well-known exoelectrogenic bacteria (e.g., Geobacter, Shewanella, Desulfovibrio, and Anaeromyxobacter) were found in these systems, but their importance in determining reactor functional performance was not supported with a high confidence in our statistical analysis.