High Performance DNA Purification using a Novel Ion Exchange Matrix

Abstract

Ion exchange chromatography has emerged as a reliable alternative to classic CsCl-ethidium bromide gradients for isolating nucleic acids of the highest purity. A plasmid purification method based on a unique anion exchange membrane (IEXM) was developed for the production of superior quality plasmids. This method was simpler and more efficient than conventional bead-based methods. Plasmids were extracted from bacterial cells through alkaline lysis. The crude lysate was clarified by a sequential filtration device that not only removed cell debris but micellar aggregates as well. The clarified lysate was mixed with an extraction solution and loaded into a spin column containing IEXM. Binding, washing, and elution conditions were optimized to achieve efficient isolation of plasmids from the impurities. IEXM had an exceedingly high dynamic binding capacity, excellent selectivity, and a near 100% recovery for plasmids. The binding capacity for pUC19 was 2.93 mg/cm³ of IEXM, which is several times greater than the values for conventional ion exchange beads. The superior selectivity of the method was reflected in the extremely low levels of endotoxin, and thus it is well-suited for critical applications in eukaryotic systems.

Nucleic acid purification is a fundamental methodology in the field of biotechnology and all disciplines of life sciences. Plasmids are basic vectors for recombinant technologies and are commonly used in molecular and cell biology laboratories. There are a variety of methods for purification of plasmids at different scales and with different levels of quality. Among them, alkaline lysis of bacterial cells followed by capture on silica in the presence of chaotropic reagents is commonly used for general molecular biology purposes. However, this method lacks the purity required for use in eukaryotic systems. A large and increasing number of advanced applications including transfection and microinjection of animal cells and gene therapies are demanding plasmids in large quantity and extremely high purity. Ion exchange chromatography has replaced CsCl-ethidium bromide gradient centrifugation as the method of choice to produce ultrapure plasmids at a scale from laboratory to manufacturing.1–4

Several ion exchange-based plasmid purification kits have been developed and are widely used in research laboratories. These products are disposable, cost-effective, and minimize cross-contamination, and sophisticated chromatographic instrumentation is not required. Most of them are gravity-flow, liquid chromatographic columns containing porous silica beads modified with diethylaminoethanol. The gravity columns require minimal use of equipment but have limited diffusional efficiency and large volume retention. Consequently, only binding sites on the outer surface are effective for DNA capture. Thus, a large bed volume is needed, and considerably large volumes of buffers are required for sufficient buffer exchange in each chromatographic step.5

In contrast to conventional bead-based chromatography, membranes have the advantages of high flow rate, convenient format, low cost, and potential for high throughput applications.6–8 Ion exchange membranes have been used for separation of nucleic acids, endotoxins and virus.5–9 For DNA used in eukaryotic cell manipulations, in addition to eliminating impurities such as RNA, complete removal of bacterial endotoxins is particularly critical.10,11 Endotoxins are components of the bacterial cell membrane. Introducing DNA containing endotoxins into eukaryotic cells may trigger undesired cellular responses which may harm the health of the cells and interfere with designed experiments.12–15 Here we report the development of a DNA purification method using a unique anion exchange membrane (IEXM). By using IEXM, the purification procedure was considerably simplified and could be used to produce ultrapure plasmids with extremely low levels of endotoxin.

MATERIALS AND METHODS

Chemicals, Plasmids, Bacterial Culture, and Alkaline Lysis

All chemicals were purchased from Sigma-Aldrich (St. Louis, MO) or Fisher Scientific (Pittsburgh, PA). Plasmids pUC19 and pBR322 (New England Biolabs, Beverly, MA) were transformed into ElectroEB10B competent cells (EdgeBio, Gaithersburg, MD) with a MicroPulser Electroporator (BioRad, Hercules, CA). Following selection on agar containing LB media and 50 μg/mL ampicillin, a single colony was inoculated into LB medium containing 50 μg/mL ampicillin and grown overnight at 37°C. Cells were harvested by centrifugation at 5000 × g for 15 min in RC-5 Superspeed Refrigerated Centrifuge (Sorvall, New-town, CT) and subjected to alkaline lysis. The bacterial pellet from each 150-mL culture was resuspended in 10 mL of resuspension solution S1 (50 mM Tris-HCl, 10 mM EDTA, 100 μg/mL RNase A, pH 8.0), then treated with 10 mL of lysis solution S2 (0.2 N NaOH, 1% SDS) for no more than 5 min, and then neutralized with 10 mL of ice cold neutralization solution S3 (3 M potassium acetate, pH 5.5) by gentle mixing. The crude lysate was clarified with a filtration assembly consisting of a 60-mL syringe with a 20 μm nominal pore size polyethylene frit (Porex, Atlanta, GA) and a 1-μm membrane syringe filter.

DNA Purification

Solutions and devices used in DNA purification were from EdgeBio. Approximately 1/10 volume of ice-cold Extraction Solution S4 was added to the cleared lysate and mixed by inverting 10 times. The treated lysate was transferred to a spin column containing IEXM membrane and centrifuged into a collection tube at 2000 rpm (650 × g) in an Allegra 6R Centrifuge (Beckman Coulter, Fullerton, CA) using a GH-3.8 swinging bucket at room temperature. Using the same centrifugation parameters, the column was washed once with 16 mL of Wash Solution C5 and eluted into a new collection tube with 5 mL of Elution Solution C6. The eluate was transferred into a nonpyrogenic disposable 14-mL BD Falcon round-bottom centrifuge tube. DNA was precipitated by mixing with 0.7 volumes of isopropanol and centrifugation at 15,000 × g for 15 min at 4°C. The DNA pellet was rinsed once with 5 mL of 70% ethanol and centrifuged at 15,000 × g for 10 min at 4°C. After decanting the ethanol and air-drying the pellet, DNA was dissolved in endotoxin-free Tris-EDTA buffer or water.

DNA Quantitation and Analyses

DNA concentration was determined with a NanoDrop Spectrophotometer (NanoDrop Technologies, Wilmington, DE) or by using Quant-iT PicoGreen dsDNA reagent (Invitrogen, Carlsbad, CA). DNA samples were analyzed by electrophoresis in agarose gel containing ethidium bromide and visualized under UV transillumination. Purified DNA was digested with restriction enzymes (New England Biolabs) by following the manufacturer’s instructions. DNA sequencing was performed on a PTC-100 thermal cycler (MJ Research, Watertown, MA) with a thermal cycling program of 95°C for 15 sec, 45°C for 5 sec, and 60°C for 2 min for 35 cycles. The total cycling reaction was 10 μL with 3 μL of pUC19 template and 0.25 μL of BigDye Terminator v3.1 reagent (Applied Biosystems, Foster City, CA). The extension products were purified by gel filtration spin plate (EdgeBio, Gaithersburg, MD), and analyzed on an ABI 3730xl DNA Analyzer. Sequencing data was analyzed using Phred software CodonCode InterPhace v2.3.1 (CodonCode, Dedham, MA). Endotoxin presence in DNA samples was evaluated by Limulus amebocyte lysate assay with Pyrogent Single Test (Lonza, Walkersville, MD) and by kinetic chromogenic quantitative endotoxin assay independently performed by Clongen Laboratories, LLC (Germantown, MD).

RESULTS AND DISCUSSION

Purification Procedure

Isolation of plasmid DNA by ion exchange chromatography consists of several major steps: alkaline lysis, clarifying lysate, loading cleared lysate on ion exchange matrix for DNA capture, washing to remove impurities, elution, isopropanol precipitation and ethanol rinse to concentrate DNA and desalt, and finally dissolving purified DNA in a chosen volume of a solvent suitable for downstream applications. An ideal process is characterized by a user-friendly design, high-capacity components, minimal hands-on time and reagent consumption, high recovery, and excellent purity. The described method meets most or all of those objectives.

Nevertheless, even “optimal” methods have their vagaries. In this method, complete mixing by gentle inversion and slow rotation of the tube after lysis and during neutralization was found to be critical for keeping the precipitate intact. Too much small debris will block the filtration. Incubation of the neutralized lysate in ice resulted in better aggregation of precipitates and removal of contaminants as compared with incubation at room temperature.

The distinctive features of this method are described as follows. The filter device is comprised of a 60-mL syringe with a 20 μm polytheylene frit for removal of large debris, and a 1-μm glass fiber filter to further remove small debris and micellar aggregates. The sequential filtration design minimizes particulate load in the clarified lysate, and provides a simple and rapid process with minimal clogging and low retention volume. The IEXM is a diethylamine-modified porous membrane configured in a spin column format. The porosity was optimized to provide high interior surface area accessible for plasmid binding and high flow rate for easy sample processing. The spin column is stored dry and ready to use without a pre-wetting or equilibration step. The spin columns are configured to fit in standard microcentrifuge tubes for mini preparations or in 50-mL conical centrifuge tubes for maxi preparations. Because of its exceedingly high dynamic binding capacity and high flow rate, buffer volume and operation time for each step are considerably smaller than for most conventional ion exchange resin-based methods. For example, instead of washing twice with 30 mL wash buffer, washing once with 16 mL wash buffer is sufficient for an IEXM maxi column. Bound DNA is eluted from the washed column in 5 mL, compared with the usual 15-mL elution.14 Elution of DNA in small volume and high concentration facilitates DNA precipitation by isopropanol. The precipitation can be done conveniently in a 14-mL disposable centrifuge tube or in several microcentrifuge tubes.

Binding efficiency and capacity for Plasmid

Plasmids pUC19 (high copy) and pBR322 (low copy) were isolated from the cleared lysates. Each fraction eluted from the IEXM spin column was subjected to isopropanol precipitation to remove salts, and resolved on agarose gel (Figure 1A). No plasmid was found in either unbound fraction (flow-through from IEXM spin column) or wash which contains elevated salt concentration. Plasmid was released to near completion from the column in a single elution with high-salt elution buffer. Negligible amounts of plasmid were recovered when the column was further washed with elution buffers containing up to 1.5 M NaCl (Figure 1B). The high recovery ratio was confirmed by applying a known amount of plasmid on IEXM spin column. Twenty-five micrograms of purified pUC19 in 500 μL of 10 mM Tris-HCl (pH 8.0) was loaded on a minicolumn adapted to a microcentrifugation tube and effluent was collected. After washing and elution, 25.3 ± 0.2 μg pUC19 was recovered, thus confirming that the capture and elution steps were providing close to 100% recovery. After elution, the column was further washed with elution buffer containing 1.5 M NaCl. No DNA was found in effluent, column wash, or 1.5 M NaCl elution.

FIGURE 1.

Isolation of plasmids with anion exchange membranes. A: Purification of high-copy plasmid pUC19 (lanes 1–4) and low-copy plasmid pBR322 (lanes 5–8) using IEXM. Fractions were precipitated by isopropanol to remove excess salts and redissolved in Tris-EDTA of the original sample volumes, and 10 μl was loaded on a 1% agarose gel. M, 1-kb DnA ladder. Lanes 1, 5: cleared lysates; lanes 2, 6: flow-through fractions; lanes 3, 7: column washes; lanes 4, 8: elutions. B: Recovery of pUC19 in the process of purification with IEXM. Fractions were precipitated by isopropanol to remove excess salts and redissolved in 30 μl of Tris-EDTA. Ten microliters was loaded on a 1% agarose gel. M, 1-kb DNA ladder. Lane 1: flow-through in binding step; lanes 2–5: low-salt washes with increasing concentrations of salt; lane 6: elution; lanes 7–10: high-salt washes with increasing concentrations of salt. C: RNA contaminants were copurified during pUC19 isolation with quaternary ammonium membrane. Ten microliters of each fraction was loaded on a 1% agarose gel. Lane 1: cleared lysate; lane 2: flow-through; lane 3: column wash; lanes 4–6: repetitive elutions.

The matrix has a much lower affinity for RNA binding than for plasmid. Most RNA was found in the flow-through; the trivial amount of RNA retained on the column was eliminated in the washing step. No RNA contamination was visible on an agarose gel electrophoresis even when 5 μg of purified plasmid was loaded (data shown in Figure 3). The result suggests high binding efficiency, i.e., high recovery and high selectivity of IEXM for plasmid isolation. In contrast, when a quaternary ammonium-modified anion exchange membrane (strongly basic) was used for plasmid purification, large amounts of degraded RNA copurified with the plasmid (Figure 1C). It was reported that quaternary ammonium membrane is suitable for plasmid purification.5,10 However, our observation indicates the advantageous performance of diethylamine-substituted membrane in producing high-purity plasmid.

FIGURE 3.

Analyses of the yield and purity for pUC19 purified with IEXM or anion exchange beads. Four replicate preparations were performed. A: Yields of pUC19 were determined by PicoGreen dsDNA assay (gray bars) and NanoDrop spectrophotometer (white bars). Results are mean ± SD for four replicates. B: Agarose gel analysis of purified pUC19 plasmids. M, 1 kb DNA ladders.

To test column capacity, cleared lysate from 1 L LB culture of Escherichea coli EB10B harboring pUC19 was repeatedly loaded into one IEXM maxi column with a bed volume of 0.58 cm³. After washing, plasmid was eluted and recovered by isopropanol precipitation, and 1.7 mg of pUC19 was purified. No plasmid was found in the effluents during column loading. The presence of plasmid in column wash indicates the saturation of the column (Figure 2). Therefore the estimated dynamic binding capacity for pUC19 is 2.93 mg pUC19 per cm³ of matrix. This is about ten times greater than the typical capacity for conventional porous beads.

FIGURE 2.

Binding capacity of IEXM for pUC19 plasmid. Cleared lysate from 1 L LB culture was loaded repeatedly on one IEXM maxi column which had a bed volume of 0.58 ml. M, 1-kb DNA ladder. Lane 1: cleared lysate; lanes 2–12: flow-through fractions for repetitive loadings; lane 13: column wash.

Yield and Purity of DNA

Yield and purity of plasmid purifications in this method were compared with methods based on anion exchange beads from two major commercial entities. In these studies, plasmid pUC19 was purified from 100 mL LB culture. Four replicate preparations were performed. Both PicoGreen dsDNA quantitation assay and NanoDrop spectrophotometer were used to determine plasmid concentrations. In PicoGreen assays, RNA and other UV-absorbing contaminants do not significantly affect quantitation of plasmid DNA; contaminants having absorbance at 260 nm (e.g., RNA, protein) will interfere with DNA quantitation by UV spectroscopy. Therefore, the PicoGreen assay provides values which are close to actual yields, and the ratio of the values from the two methods can be used to assess the purity of the plasmid. The yields of pUC19 purified by IEXM and the two anion exchange bead methods are comparable. The yield by IEXM is otherwise slightly higher than that of both bead methods (Table 1, Figure 3A). The small difference in yield is consistent with the fact that IEXM’s bed volume (0.58 cm³) is substantially smaller than the volume of anion exchange beads in the gravity columns (approximately 5 mL). Quantitations by PicoGreen assay and NanoDrop spectrophotometer are almost identical for plasmids purified with IEXM and bead 1. For all four pUC19 plasmids purified using bead 2, quantitation by NanoDrop was higher than that with PicoGreen. The elevated yield determined by absorbance method indicates possible contaminations in plasmid preparation with bead 2. To further investigate the purity, 5 μg of purified pUC19 plasmids were loaded on agarose gel and run for a short time (approximately 15 min) to examine residual RNA contamination (Figure 3B). The presence of degraded RNA was observed in all four pUC19 plasmids purified with bead 2, but not in plasmids purified with IEXM or bead 1.

TABLE 1.

Yield of pUC19 Purified With Anion Exchange Membrane IEXM and Beads

	Yield (mean ± SD), μg
	NanoDrop	PicoGreen	Ratio NanoDrop/ PicoGreen	Ratio A260/A280
IEXM	194 ± 2.3	188 ± 7.6	1.03	1.89
Bead 1	178 ± 0.4	177 ± 4.1	1.01	1.88
Bead 2	207 ± 10.1	179 ± 17.9	1.16	1.90

The yield was quantified by both NanoDrop spectrophotometer and PicoGreen dsDNA assay. Ratio of the two values and the ratio of UV absorbances at 260 nm and 280 nm were calculated. The highest values for yield quantified by both methods are shown in bold.

The quality of plasmid purified with IEXM was examined by routine restriction digestion analyses and the DNA sequencing. No inhibition was observed in both applications (data not shown for restriction digestion; Figure 4 for sequencing). Purified pUC19 was diluted serially and sequenced using diluted BigDye v3.1 Terminator at 1/16 of regular concentration. Interestingly, base reading lengths (Phred20 scores) remained high for extensively diluted pUC19. The Phred20 scores are comparable to that for pUC19 with the concentration normally used in the sequencing reaction (Figure 4A). Additionally, the close-to-linear signal to template concentration response suggests high quality of the plasmid (Figure 4B).

FIGURE 4.

DNA sequencing using pUC19 purified with IEXM. Purified pUC19 was serially diluted and sequenced. Results shown are average from duplicate sequencing reactions. A: Phred20 scores. B: Intensity of G-signal. The line shows linear regression of G-signal to log of pUC19 concentration.

Endotoxin levels

Endotoxins are lipopolysaccharides (LPS) generated by Gram-negative bacteria such as E. coli. Endotoxin is toxic to animal cells, and may have inhibitory effects on mammalian cell transfection.12,13 High levels of LPS contamination may interfere with downstream enzymatic reactions. Endotoxin is often copurified with plasmids in almost every common DNA purification method except ion exchange chromatography. Due to the abundance of LPS in bacterial lysate and the similar negative charge between LPS and DNA molecules, preventing LPS from copurification with DNA is a major technical challenge for developing high-performance DNA purifications based on an ion exchange platform. Elimination of endotoxin from DNA preparations not only ensures success in downstream applications, but also provides a sensitive parameter for demonstrating high purity.

The presence of endotoxin in pUC19 plasmids purified with IEXM was first evaluated with Limulus amebocyte lysate by Pyrogent Single Test from Lonza which determines level of endotoxin instead of quantitative measurement of endotoxin concentration. The endotoxin level was found to be < 0.01 endotoxin unit (EU) per microgram of pUC19. Quantitative chromogenic endotoxin assay was then performed to determine endotoxin concentration. The endotoxin concentration was 0.00246 ± 0.00125 EU/ μg DNA, well below 0.1 EU/μg DNA which is considered to be “endotoxin-free.” The result was compared with the published results for other methods13,14 (Figure 5). It is noted that a control set of undiluted sample spiked with 0.5 EU/mL of standard were included in chromogenic endotoxin assays to monitor inhibition if any to rule out the possibility of a false negative. The spiked values were almost equal to (with an average CV of 4.5%) the combined concentration of endotoxin from the standard and the plasmid samples.

FIGURE 5.

Endotoxin levels in plasmids purified with different methods. 1: Qiagen Plasmid Maxi kit. 2: Two rounds of CsCl-ethidium bromide gradient centrifugation. 3: Qiagen EndoFree Plasmid Maxi kit. IEXM: pUC19 maxi preparation from 100 ml of LB culture using IEXM maxi spin column. (Data for 1–3 are cited from published results. Data for IEXM were from quantitative chromogenic endotoxin assay.)

Besides plasmid purification, we tried using IEXM in isolating bacterial genomic DNA and BAC constructs with promising recovery efficiency (data not shown). Due to the microporous structure of IEXM, the recovery is expected to be better than other membranes with much smaller pore sizes, which may trap DNA with high molecular weights.

In conclusion, the IEXM-based method described here was successfully used in production of ultrapure plasmids with high yield and efficiency, and the developed method can be potentially applied in purification of large DNA molecules such as BACs, cosmids, fosmids, and genomic DNAs. The method is faster and simpler than other commercially available methods for plasmid purification and provides superior quality.