Optimizing the indirect extraction of prokaryotic DNA from soils

The objective of this work was to develop protocols to selectively extract prokaryotic DNA from soils, representative of the whole community, amenable to high-throughput whole genome shotgun sequencing. Prokaryotic cells were extracted from soils by blending, followed by gradient centrifugation. Detergent (sodium deoxycholate) was required for complete dispersion of soil aggregates and detachment of prokaryotic cells from a broad range of soil types. Repeated extractions of a given soil sample were critical to maximize cell yield. Furthermore, cells obtained through repeated extractions captured unique prokaryotic assemblages that would otherwise have been missed in single-pass extractions. DNA was isolated from extracted cells using one of the following treatments: i) lysozyme-SDS-proteinase K (enzymatic) digestion; ii) potassium ethyl xanthogenate digestion; or iii) enzymatic digestion of cells embedded in agarose plugs. In addition, these methods were compared to a commercial bead-beating extraction kit (MoBio UltraClean). Of the indirect DNA extraction methods, plug digestion generated the largest yields (up to 70% of yields obtained by direct DNA extraction) of high-molecular weight DNA (>400 kb). Thus, plug digestion is amenable to large-insert metagenomic library construction and analysis. Comparisons of banding patterns generated by RAPD-PCR and DGGE indicated that sequence composition and inferred community composition of a given extract varied greatly with DNA isolation method. While overall diversity did not change significantly with the cell lysis method, analysis of 16S rRNA gene clone libraries revealed that each extraction procedure produced unique distributions of prokaryotic phyla within the sample population.