PCR contamiantion sanger sequencing. Bacterial DNA contamination of commercial PCR enzymes: considerations for microbiome protocols and analysis.

Background: Bacterial contamination of laboratory and sequencing reagents is an issue that continues to plague many studies of the microbiome. This is of particular concern when investigating tissues that are expected to have low bacterial burdens. However, this issue of contamination remains underappreciated. Existing studies have typically examined this issue broadly, grouping together contamination that could be due to extraction kits, laboratory consumables, and molecular biology reagents. Here we specifically identify differential bacterial contamination across a broad sampling of commercial polymerase chain reaction (PCR) enzymes and their associated reaction components. Results: Six of the nine tested enzymes were shown to be contaminated with bacterial DNA when comparing reactions with E. coli template DNA to reactions containing no template DNA. These enzymes were found to contain a variety of contaminating sequences as determined using Sanger sequencing followed by querying GenBank via Blast search. The predicted identities of the bacterial contaminants were also found to associate with the enzyme used in the reaction, as determined by phylogenetic analysis using the greatest likelihood algorithm with 1000 bootstrap replicates. Conclusions: We found that contamination was common among commercially available PCR enzymes and their reaction components, with six of nine tested enzymes having bacterial DNA contamination. This contamination was also highly diverse and included several bacterial families that have been previously reported as common sources of contamination in microbiome studies. Importantly, when comparing the contamination detected in the reactions, there was strong association between reactions that were generated by the same enzyme, indicating that this contamination was a function of the enzyme used, and was unlikely to be introduced by our own workflow. Together these data indicate that selection of PCR enzymes for microbiome work is important not only because of intrinsic characteristics of the enzymes, but also because they can generate detectable and specific contamination profiles, and that this contamination should be controlled for by inclusion of robust negative controls.