Primer Extension Capture: Targeted Sequence Retrieval from Heavily Degraded DNA Sources

We present a method of targeted DNA sequence retrieval from DNA sources which are heavily degraded and contaminated with microbial DNA, as is typical of ancient bones. The method greatly reduces sample destruction and sequencing demands relative to direct PCR or shotgun sequencing approaches. We used this method to reconstruct the complete mitochondrial DNA (mtDNA) genomes of five Neandertals from across their geographic range. The mtDNA genetic diversity of the late Neandertals was approximately three times lower than that of contemporary modern humans. Together with analyses of mtDNA protein evolution, these data suggest that the long-term effective population size of Neandertals was smaller than that of modern humans and extant great apes.

1. The DNA template here is a 454 library prepared from a low copy number DNA source as described in Hofreiter, 2007a, 2007b) and (Maricic and Pääbo, 2009 Recommended: keep remaining 5 ul of the eluate from step 2.3 for qPCR quantification. 3. Transfer the whole mixture to a fresh 1.5ml tube (this helps to reduce carryover of non-target library fragments). Pellet the supernatant using the Magnetic particle collector (MPC) and discard the supernatant. 4. Wash 5 times with 500 μl 1xBW buffer (many washes help to remove as much background as possible). After the last washing step, spin down briefly and remove the last traces of supernatant. 5. Add 500μl 1x Hot Wash (HW) buffer. Shake for 2 min at 65°C (or 5C above the PEC primer Tm) on a thermal block. Remove the supernatant quickly after this, to minimize cooling down. (This step is to remove background fragments still associated with the PEC primers but not actually extended on during the extension step). Remove the last traces of supernatant. 6. Resuspend the bead pellet in 30 μl EB buffer. Transfer the whole mixture to a fresh 1.5ml tube (this helps to reduce carryover of non-target library fragments). Incubate at 95°C for 3 minutes in a thermal cycler for elution. Place in the MPC and remove the supernatant, taking care to leave the beads behind.
1. Prepare a PCR master mix for the required number of samples. Representative Results:

Reagent μl per reaction
It is strongly recommended when starting out with the PEC protocol to perform first a capture reaction where a small amount of a 'positive control target sequence' (e.g. a~100bp PCR product -1 picogram is easily enough) is mixed in with the normal recommended amount of amplified 454 library template (see step 2.1), and capture is performed with a single PEC primer designed to capture the positive control product. If this control reaction is performed, qPCR with both positive control-specific and 454 adaptor-specific primer pairs can be used to quantify the amount of 'control target' versus background in the purified primer extension reaction (1.3), primer extension product (2.3) and eluted bead capture product ( 3.6). In this way, both the effectiveness of background removal ("specificity") and the efficiency of target recovery ("sensitivity") in your experimental conditions can be directly measured.
A successful PEC protocol normally has the following properties - If there is less than 1% of the target remaining in the final product, the primer extension step may have been unsuccessful and should be investigated.
If there is much more than 0.01% of the background remaining in the final products, the washing steps may have been incorrectly performed and should be investigated.

Discussion
The PEC method is simple, quick, sensitive and specific. Therefore we the authors envisage multiple applications outside ancient DNA, such as capture of small RNA fragments from an RNA library, interrogation of structural variation in a pooled sample or capture of 16S (or other loci) diversity from a metagenomic sample. One point to mention is that the sensitivity of capture becomes lower as the number of PEC primers in a multiplex capture reaction increases. PEC is therefore not ideally suited for capture of very large (e.g. a megabase or more) capture regions, but is extremely well suited for capture of small target regions or even SNP positions from many individuals in a rapid fashion.