Pathogen Quantitation in Complex Matrices: A Multi-Operator Comparison of DNA Extraction Methods with a Novel Assessment of PCR Inhibition

BACKGROUND Mycobacterium bovis is the aetiological agent of bovine tuberculosis (bTB), an important recrudescent zoonosis, significantly increasing in British herds in recent years. Wildlife reservoirs have been identified for this disease but the mode of transmission to cattle remains unclear. There is evidence that viable M. bovis cells can survive in soil and faeces for over a year. METHODOLOGY/PRINCIPAL FINDINGS We report a multi-operator blinded trial for a rigorous comparison of five DNA extraction methods from a variety of soil and faecal samples to assess recovery of M. bovis via real-time PCR detection. The methods included four commercial kits: the QIAamp Stool Mini kit with a pre-treatment step, the FastDNA® Spin kit, the UltraClean™ and PowerSoil™ soil kits and a published manual method based on phenol:chloroform purification, termed Griffiths. M. bovis BCG Pasteur spiked samples were extracted by four operators and evaluated using a specific real-time PCR assay. A novel inhibition control assay was used alongside spectrophotometric ratios to monitor the level of inhibitory compounds affecting PCR, DNA yield, and purity. There were statistically significant differences in M. bovis detection between methods of extraction and types of environmental samples; no significant differences were observed between operators. Processing times and costs were also evaluated. To improve M. bovis detection further, the two best performing methods, FastDNA® Spin kit and Griffiths, were optimised and the ABI TaqMan environmental PCR Master mix was adopted, leading to improved sensitivities. CONCLUSIONS M. bovis was successfully detected in all environmental samples; DNA extraction using FastDNA® Spin kit was the most sensitive method with highest recoveries from all soil types tested. For troublesome faecal samples, we have used and recommend an improved assay based on a reduced volume, resulting in detection limits of 4.25×10(5) cells g(-1) using Griffiths and 4.25×10(6) cells g(-1) using FastDNA® Spin kit.