We developed a method for selective purification of DNA using the cationic detergent, cetyltrimethylammonium bromide (CTAB), accompanied with urea and controlled high-salt (NaCl) concentration. This method is effective for rapid separation of DNA fragments from artifacts such as PCR primer dimers or ligation adapters. The CTAB-associated purification completely removed the short PCR artifacts and primers, as well as enzymes and buffer, while recovering a sufficient quantity of amplicons for subsequent experiments such as preparation of libraries. This method could also be applied to the fractionation of nucleic acids generated by other types of reactions.

Many molecular biology techniques require purification of cDNAs, DNAs, or DNA-RNA hybrids, and/or fractionation of nucleic acids from various biological sources and reactions to improve downstream performance. Generally, PCR amplicons with short primers (~20 bp) and no artifacts can be readily purified with commercially available PCR purification kits that have a maximum cutoff of 100 bp. However, in multi-step techniques requiring PCR, these purification kits often cannot remove artifacts longer than 100 bp; this is especially problematic when DNA libraries must be prepared (1,2,3). Next-generation flow-cell sequencers offer DNA sequencing directly from PCR, which is advantageous because subcloning is avoided, but it also requires the subsequent purification of the amplified products to increase the sequencing quality, which is proportional to the purity of the template (4). In cap analysis of gene expression (CAGE) technologies with deep sequencing or deepCAGE (2)—in which PCR is used as a key step to prepare libraries for flow-cell sequencing—sometimes artifacts <150 bp are generated (1,2). In a search for ways to remove these artifacts, we systematically investigated the effect of NaCl concentration on DNA fractionation in the presence of cetyltrimethylammonium bromide (CTAB) and urea (5,6), using a glass fiber matrix in a spin column (illustra GFX Column; GE Healthcare, Little Chalfont, Buckinghamshire, UK). The GFX kit exploits a glass fiber matrix in conjunction with chaotropic salt to capture nucleic acids (7,8). Here, to better modulate the fractionation of differently-sized nucleic acids as a substitute for chaotropic salt, we have introduced a CTAB-urea solution to capture the nucleic acids. Selective purification was performed with variable NaCl concentrations. CTAB has been used for decades to extract and purify DNAs from biological sources (5,6, 9,10,11). It has recently been shown that CTAB along with urea can improve the purification of RNAs by removing polysaccharides and other contaminants (6). We have also found that RNAs shorter than tRNAs are not precipitated efficiently under certain conditions (6). With these problems in mind, we have explored various options to optimize this purification method.

To set up the methodology, we started with 500 ng 50-bp DNA ladder marker (Invitrogen, Carlsbad, CA, USA) or PCR reaction (5 µL). The DNAs were mixed with 1% CTAB-urea solution (4 M urea, 50 mM Tris-HCl pH 7.0, and 1 mM EDTA pH 8.0). CTAB forms positive micelles in solution with nucleic acids: it interacts with nucleic acids through the phosphates and its quaternary ammonia, and the long chain of the detergent localizes away from the DNA groove to reduce contact with the surrounding water. It does this by forming aggregates and thereby promotes condensation (10,12). We applied this property of CTAB to capture the DNAs in the glass fiber. The exact mechanism of this capture is currently uncharac-terized, but several possibilities are that (i) micelles are absorbed to the glass filter matrix, (ii) CTAB coats the fiber and makes it positively charged, or (iii) the CTAB-DNA complex is simply retained very efficiently to the surface. The interaction of CTAB and nucleic acids was controlled by adding 5 M NaCl solution to final concentrations of 0.4–0.6 M. Then, the mixture was shaken briskly and heated at 65°C for 10 min. Following incubation, samples were kept at room temperature for 10 min and applied to the GFX columns. Each column was centrifuged at 16,000× g for 1 min. Flow-through was collected and eluted DNA was recovered by ethanol precipitation for later control electrophoresis. To remove the CTAB, we washed the column with 600 µL wash buffer (60% ethanol, 0.3 M NaCl, 100 mM Tris-HCl pH 7.5, and 5 mM EDTA pH 8.0). The column was washed a second time with 600 µL 80% ethanol, which removes the CTAB while keeping the DNA precipitated onto the matrix. The column was replaced with a new 1.5-mL siliconized tube and DNA was eluted with 50 µL nuclease-free H₂O, then concentrated by ethanol precipitation. DNA obtained in the first flow-through and final elution was analyzed on 6% polyacrylamide gel or 1.5% agarose gel. Gels were stained either with ethidium bromide or Gel Star (Lonza, Rockland, ME, USA) and visualized with BioDocIt (UVP, Upland, CA, USA). We estimated how much DNA was recovered by comparison with a fragmented ladder and quantified by measuring the concentration with a NanoDrop 1000 spectro-photometer (Thermo Fisher Scientific Inc., Wilmington, DE, USA). The overall recovery was ~65%; recovery was ~30% for smaller fragments just above the 150-bp cutoff, and ~80% for larger fragments (data not shown).