A comparison of five methods for extraction of bacterial DNA from human faecal samples

The purity of DNA extracted from faecal samples is a key issue in the sensitivity and usefulness of biological analyses such as PCR for infectious pathogens and non-pathogens. We have compared the relative efficacy of extraction of bacterial DNA (both Gram negative and positive origin) from faeces using four commercial kits (FastDNA(R) kit, Bio 101; Nucleospin(R) C + T kit, Macherey-Nagal; Quantum Prep(R) Aquapure Genomic DNA isolation kit. Bio-Rad; QIAamp(R) DNA stool mini kit, Qiagen) and a non-commercial guanidium isothiocyanate/silica matrix method. Human faecal samples were spiked with additional known concentrations of Lactobacillus acidophilus or Bacteroides uniformis, the DNA was then extracted by each of the five methods, and tested in genus-specific PCRs, The Nucleospin(R) method was the most sensitive procedure for the extraction of DNA from a pure bacterial culture of Gram-positive L. acidophilus (10(4) bacteria/PCR), and QIAamp(R) and the guanidium method were most sensitive for cultures of Gram-negative B. uniformis (10(3) bacteria/PCR). However, for faecal samples, the QIAamp(R) kit was the most effective extraction method and led to the detection of bacterial DNA over the greatest range of spike concentrations for both R. uniformis and L. acidophilus in primary PCR reactions. A difference in extraction efficacy was observed between faecal samples from different individuals. The use of appropriate DNA extraction kits or methods is critical for successful and valid PCR studies on clinical, experimental or environmental samples and we recommend that DNA extraction techniques are carefully selected with particular regard to the specimen type. (C) 2002 Elsevier Science B.V. All rights reserved.