How to deal with PCR contamination in molecular microbial ecology.

Microbial ecology studies often use broad-range PCR primers to obtain community profiles. Contaminant microbial DNA present in PCR reagents may therefore be amplified together with template DNA, resulting in unrepeatable data which may be difficult to interpret, especially when template DNA is present at low levels. One possible decontamination method consists in pre-treating PCR mixes with restriction enzymes before heat-inactivating those enzymes prior to the start of the PCR. However, this method has given contrasting results, including a reduction in PCR sensitivity. In this study, we tested the efficiency of two different enzymes (DNase 1 and Sau3AI) as well as the effect of dithiothreitol (DTT), a strong reducing agent, in the decontamination procedure. Our results indicate that enzymatic treatment does reduce contamination levels. However, DNase 1 caused substantial reductions in the bacterial richness found in communities, which we interpret as a result of its incomplete inactivation by heat treatment. DTT did help maintain bacterial richness in mixes treated with DNase 1. No such issues arose when using Sau3AI, which therefore seems a more appropriate enzyme. In our study, four operational taxonomic units (OTU) decreased in frequency and relative abundance after treatment with Sau3AI and hence are likely to represent contaminant bacterial DNA. We found higher within-sample similarity in community structure after treatment with Sau3AI, probably better reflecting the initial bacterial communities. We argue that the presence of contaminant bacterial DNA may have consequences in the interpretation of ecological data, especially when using low levels of template DNA from highly diverse communities. We advocate the use of such decontaminating approaches as a standard procedure in microbial ecology.