Managing microbial communities for finding new enzymes and sequentially reconstruct genomes from complex metagenomes

INVITEE

International audience

English. Global understanding on environmental microbial communities is currently limited by the bottleneck of genome reconstruction. Soil is a typical example where individual cells are currently mostly uncultured and metagenomic datasets unassembled. In this study, the microbial community composition of a natural grassland soil was managed under several controlled selective pressures to experiment a "multi-evenness" stratagem for sequentially attempt to reconstruct genomes from a complex metagenome. While lowly represented in the natural community, several newly dominant genomes were successfully reconstructed under various "harsh" tested conditions. These genomes belong to several genera including Leifsonia, Rhodanobacter, Bacillus, Ktedonobacter, Xanthomonas, Streptomyces and Burkholderia. So far, from 10 to 78% of generated metagenomic datasets were reconstructed, so providing access to more than 88 000 genes of known or unknown functions and to their genetic environment. Functions of potential industrial interest including chitinases were also discovered by metagenomic approaches after chitin enrichment. An integrative approach coupling rrs phylochip and high throughput shotgun sequencing of the metagenomic DNA has been applied to investigate the shift in bacterial community structure and functions after incubation with chitin. Results indicate that chitin enrichment leads to an increase of Actinobacteria, γ-proteobacteria and β-proteobacteria suggesting an increase of the observed bacterial diversity and the specific selection of chitin degrading bacteria. This "divide and conquer" strategy could be applied to other environments and using auxiliary sequencing approaches like single cell to detect, connect and mine taxa and functions of interest while creating an extensive set of reference genomes.