Amplification products were sequenced directly using the same primers used for PCR, under the following conditions: 10.5-µL reaction volumes containing 0.5 µL BigDye terminator mix v3.1, 1.88 µL 5× sequencing buffer (Applied Biosystems), 1 µM primer and 0.5 µL PCR product. Sequencing thermal cycling parameters were 96°C for 2 min; 30 cycles of 96°C for 30 s, 55°C for 30 s, and 60°C for 4 min; and a 4°C hold. We cleaned cycle sequencing products from each reaction on Sephadex columns (Cat. no. S5897; Sigma-Aldrich, St. Louis, MO, USA) and ran the samples on an ABI 3730 sequencer (Applied Biosciences). Sequences generated under these conditions served as the baseline for quality comparisons. For the various trials below, these conditions were held constant while only the indicated parameters of interest were changed.Lowering PCR extension temperature
Reducing the extension temperature has previously been shown to be useful for amplification of A/T-rich regions of DNA (19). We hypothesized that if stutter products are caused by dissociation of the polymerase from the template strand, then the generation of slipped-strand products may be mitigated by decreasing extension temperatures in attempts to reduce DNA melting and increase binding affinity of Taq DNA polymerase. Since the affinity of Taq is not at its optimum at the normal extension temperature of 72°C (20) and DNA melting of A/T-rich regions may occur at 72°C (19), our first experiment was to determine whether sequence quality could be improved by lowering the extension temperature to 60°C. To accommodate the reduced extension speed of the polymerase at this temperature, we also increased the extension time to 3 min.Reducing the number of cycles in PCR
Some reports indicate that PCR-generated errors during amplification increase toward the end of the cycling phase of PCR (21,22). In order to determine whether the main cause of stutter product formation was a consequence of excessive cycle numbers, we performed a series of reactions using the same standard conditions, but with reduced total cycles: 22, 24, 26, 28, and 30 cycles instead of 35.Inclusion of co-solutes in PCR
With a priori knowledge of the target sequences of our samples, we were aware that most samples likely possess significant secondary structure due to the presence of two mononucleotide repeats of complementary bases. In attempts to minimize the amount of polymerase dissociation caused by template secondary structure, we tested the inclusion of three co-solutes— DMSO, betaine, and trehalose—that have been shown to improve PCR efficiency by lowering the DNA melting temperature (23,24) and reducing secondary structure formation (25). Therefore, we tested the inclusion of (i) a 3% (v/v) final concentration of DMSO, (ii) a 1 M final concentration of betaine, and (iii) a 5% (w/v) final concentration of trehalose in both the PCR and sequencing reactions.New generation polymerases
Following these experiments we progressed to the evaluation of three new-generation enzymes that are reported as having levels of processivity more than two times greater than Taq DNA polymerase, and lower error rates: Phusion (Finnzymes, Espoo, Finland), Herculase II Fusion (Agilent, Santa Clara, CA, USA), and KAPAHiFi (Kapabio-systems, Boston, MA, USA) (see technical data sheets at www.finnzymes.com/pdf/phusion_brochure_fph6_low.pdf, www.kapabiosystems.com/public/pdfs/kapa-hifi-pcr-kits/KAPA_HiFi_Brochure.pdf, and www.stratagene.com/lit_items/PCRBrochure6-21-07.pdf).
Phusion is a Pyrococcus-like DNA polymerase that has been attached to the nonspecific dsDNA binding protein Sso7d, resulting in increased catalytic activity and processivity (26). Herculase II Fusion is also a Pfu-based DNA polymerase with an added double-stranded DNA binding domain, and is provided with a reaction buffer that eliminates the inhibition of proofreading enzymes by PCR byproducts. KAPAHiFi is a high-fidelity engineered DNA polymerase that has been modified for increased affinity to DNA resulting in high processivity without fusion based proteins.
We also evaluated a fourth polymerase, Topo Taq HF (Fidelity Systems, Gaithersburg, MD, USA) which is a hybrid enzyme that is linked to multiple nonspecific DNA binding HhH domains (27) and is blended with a topoisomerase, conferring strand displacement activity.
Reaction and cycling protocols generally followed manufacturers' recommendations. For Phusion, we used 20-µL reaction volumes containing 0.2 U polymerase with 1× Phusion HF Buffer (containing 1.5 mM MgCl2), 0.2 mM dNTPs, 0.2 µM each primer, and 20 ng genomic DNA. For KAPAHiFi, we used 20-µL reaction volumes containing 0.4 U polymerase with 1× KAPAHiFi Fidelity Buffer (containing 2.0 mM MgCl2), 0.2 mM dNTPs, 0.2 µM each primer, and 20 ng genomic DNA. For Herculase II Fusion, we used 20-µL reaction volumes containing 0.4 U polymerase with 1× Herculase II reaction buffer (containing 2.0 mM MgCl2), 0.25 mM dNTPs, 0.2 µM each primer, and 20 ng genomic DNA. For Topo Taq HF, we used 20-µL reaction volumes containing 1 U polymerase with 1× amplification buffer (containing 3.0 mM MgCl2), 0.5 mM dNTPs, 0.3 µM each primer, and 20 ng genomic DNA.