To the best of our knowledge, no studies have incorporated the use of an IPC and amplification efficiency measurements in order to contrast their relative contribution in monitoring PCR inhibition. In this study, we present a comprehensive approach to detecting and quantifying PCR inhibition that incorporates both features, and recommend its use in routine sample processing (26), particularly when working with low-copy templates. Beyond simply identifying problematic samples, we demonstrate how quantifying inhibition effects can be used to determine an optimal combination of PCR facilitators (and dilution in some cases) for every DNA extract that maximizes quantitative accuracy and/or template recovery.Materials and methods Samples and DNA extraction
The tests described herein were performed on 47 DNA extracts from a diverse set of ancient samples (Table 1), including permafrost soil, mammoth bone and hair, and packrat paleofeces. DNA was extracted from the soil (~250 mg) according to Willerslev et al. (27), but using 25 mM TCEP [Tris (2-carboxyethyl) phosphine hydrochloride] in place of β-mercaptoethanol and with a 10-min vortexing step in Lysing Matrix E tubes (MP Biomedicals, Solon, OH, USA) for sample disruption. The extracts were purified either using columns as indicated or using a modification of the protocol of Boom et al. (28). In this case, the extracts were mixed with 4 mL binding buffer [5 M guanidinium thiocyanate (GuSCN), 50 mM Tris-Cl pH 8, 22.5 mM NaCl, 20 mM EDTA, 1.25% Triton-X 100] that had been previously incubated with 50 µL size-fractionated silicon dioxide (30 min with rotation at room temperature). Additional GuSCN was added to maintain a 5 M concentration upon addition of the extract and the pH was adjusted to 4.5–5.5 with glacial acetic acid to maximize binding efficiency (29). Following a minimum 1-h (room temperature) incubation, the silica-bound DNA was washed twice with 1 mL buffer (5 M GuSCN, 50 mM Tris-Cl pH 8, 22.5 mM NaCl) and once with 1 mL 80% ethanol (in 1× TE pH 7.5). The pellet was dried at 56°C for 5 min and the DNA eluted in 100 µL Buffer EB (Qiagen, Hilden, Germany) as per the procedure performed by Willerslev et al. (27). A subset of the soil samples were processed using the UltraClean Soil DNA Isolation Kit (MO BIO Laboratories, Carlsbad, CA, USA) according to the manufacturer's alternative protocol for maximum yields. DNA was extracted from the bone samples (~100 mg) according to Poinar et al. (30). As part of a separate study, the same protocol (without a demineralization step) was used for the feces samples. Although commonly used silica-based methods may offer high DNA purity, we have noted considerable DNA loss in exchange (unpublished results) and wished to evaluate whether an alternative protocol might yield enough DNA to outweigh the effects of increased PCR inhibition. The hair samples (54 mg and 2 mg) were processed as per Gilbert et al. (31).
To assess the levels of PCR inhibition in the samples, we monitored the effect of each purified extract on the amplification of an IPC during qPCR. As amplification of any endogenous templates would distort the quantification results, we selected an assay targeting the human β-2-microglobulin gene (B2M), which should not be present in these extracts, and used an IPC template derived from the B2M cDNA sequence (GenBank accession no. NM_004048), such that amplification of any contaminating human genomic DNA could be distinguished. The primers used in this assay (5′-3′: TGACTTTGTCACAGCCCAAGATA and AATCCAAATGCGGCATCTTC) flank two intronic regions (~2 kb total), such that amplification of the cDNA-derived template yields an 85-bp product that is easily differentiated from amplification of longer B2M genomic DNA (32). As the assay is primarily used for quantification of human cDNA, the template (5′-3′: GAACCATGTGACTTTGTCACAGC-CCAAGATAGTTAAGTGGGATC-GAGACATGTAAGCAGCATCATG-GCGGTTTGAAGATGCCGCATTT-GGATTGGATGA) was synthesized to include an internal base modification (underlined; primer binding sites are italicized) so that it may be distinguished from human cDNA contamination by sequencing as well, although this feature is not important for these experiments. We used an available synthetic ssDNA version of this template for convenience. As the same ssDNA template was used in all reactions, we perceive no disadvantages in its use compared to dsDNA template, nor must any corrections be made to the inhibition measurements.