SURE gel electrophoresis: A method for improved detection and purification of dilute nucleic acid samples

Abstract

Agarose gel electrophoresis is performed routinely by molecular biologists as both an analytical and a preparative method for characterization of nucleic acids. Gel analysis of highly dilute DNA solutions is challenging because of the limited sensitivity of detection available with conventional methods. In this study a new approach is described for concentrating samples directly within gels called SURE (successive reloading) electrophoresis. The approach involves loading of dilute samples multiple times into a single well, with each loading followed by a brief pulse of electrical current before the next sample is loaded. The procedure generates single bands created by molecular stacking that exhibit strongly enhanced signal intensities and minimal band broadening. Using optimized voltages and time intervals as many as 20 successive loadings could be performed and up to 800 μL could be loaded into a single well. Gel extraction and fluorescent quantitation demonstrated that approximately 97 % of the DNA from each loading was incorporated into the resultant band. Highly dilute DNA samples (<0.0007 ng per microliter) could be readily detected after six loadings. The method produced good results with either TAE or TBE as electrophoresis buffers, using loading dyes with or without SDS, and in both minigels and large gels.

1. Introduction

DNA and RNA molecules that have been extracted from cells or synthesized in the laboratory are commonly analyzed by electrophoresis in agarose gels to assess yield and purity [1–3]. Agarose gel electrophoresis is also frequently used to purify nucleic acids by extracting bands out of gels followed by separation from the agarose gel matrix [4–6]. Samples used for this type of analysis are mixed with a loading dye prior to loading into a well. The dye mixture contains a a high density agent such as ficoll, sucrose or glycerol to promote sedimentation of the sample to the bottom of the well plus one or more tracking dyes such as bromophenol blue or xylene cyanol [7,8]. Some loading dyes also contain sodium dodecylsulfate (SDS), which can terminate enzyme reactions and may also reduce band thickness.

Purification of DNAs or RNAs from cells or use of methods such as polymerase chain reaction (PCR) and other laboratory synthesis methods can generate products that are highly dilute. Nucleic acid samples whose concentrations are low often present a problem: they may be so dilute that loading the maximum volume possible into an agarose gel well results in (a) insufficient sample to be visualized by conventional detection methods such as staining with ethidium bromide or with more sensitive stains like SYBR Gold [9] or (b) insufficient sample to extract from the gel and produce practical amounts for use in downstream experiments. In such cases it is sometimes possible to concentrate the sample by methods such as precipitation with ethanol or isopropanol [10–12] or one can repeat the nucleic acid preparation method used initially but at a larger scale. Each of these responses has drawbacks, including potential sample loss due to inefficiency of precipitation methods and the loss of time associated with repetitive sample preparation.

In the current study we describe a method for loading samples repeatedly into the wells of agarose gels, overcoming detection problems associated with highly dilute solutions. Using appropriate electrophoresis parameters, the method permits loading of volumes that are many times the capacity of a single well. The new approach can improve both analytical and preparative experiments, i.e., permitting the visualization of previously undetectable DNA bands from highly dilute samples and enhancing yields of DNAs extracted from agarose gels.

2. Materials and methods

2.1. Materials

Most experiments were performed using a 2704 bp DNA fragment produced by PCR amplification of the Saccharomyces cerevisiae EXO1 gene with primers 5-ExoC (CTGAGGTTGACTACTACGAGCTATACGAATATC) and 3-ExoD (GGCAAACATACTTGTGGCTTAATTTGACACATC). The template was chromosomal DNA prepared from yeast strain BY4742 [13] using a chemical extraction method [14]. pBluescript plasmid DNA [15] was prepared using a Qiaprep plasmid miniprep kit from Qiagen. Electrophoresis buffers were prepared from 50X TAE or 10X TBE stock solutions (ThermoFisher Scientific). Glycerol was purchased from EMD Millipore and sucrose was obtained from Mallinckrodt. Phusion PCR reaction buffer (5X) and purple loading dyes with or without SDS were purchased from New England Biolabs. Each commercial 6X loading dye contained 15 % Ficoll-400, 60 mM EDTA, and 20 mM Tris (pH 8) plus two tracking dyes at 0.001 % and 0.02 %. When present in the 6X loading dye, the concentration of SDS was 0.48 %. Similarly, glycerol and sucrose were present at 20 % and 30 %, respectively, when used in loading dyes. New loading dye prepared in the laboratory contained 15 % Ficoll-400 (Research Products International), 20 mM Tris base (pH 8) and 0.02 % bromophenol blue with or without EDTA. TE DNA storage solutions contained 10 mM Tris (pH 8.0) and 0.5 mM EDTA. Ethidium bromide solution (10 mg/mL) was purchased from IBI Scientific and SYBR Gold (10,000X) was from ThermoFisher. All gels were prepared using agarose LE from Gold Biotechnology.

2.2. Agarose slab gel electrophoresis methods

Most experiments were performed using Horizon 11 × 14 cm gel rigs (Labrepco) along with 10-well combs. Minigels were run using a BioRad mini-sub cell GT 8 system with 8-well combs. Applied voltages were varied using an EPS601 power supply (GE Healthcare). Unless specified otherwise, all gels contained 0.8 % agarose prepared using 80 mL 1X TAE (40 mM Tris, 20 mM acetic acid and 1 mM EDTA), producing gels that were ~5 mm thick. Samples were run until the tracking dye had migrated halfway down the gel to a marked line in the rig and were then stained with ethidium bromide (~0.5 μg per milliliter) for 10 min except where indicated otherwise. SYBR Gold was diluted 1:10,000 into water before use. Gel images were obtained using a Bio-Rad Chemi-Doc MP instrument after staining.

2.3. SURE gel electrophoresis protocol

In a typical SURE electrophoresis experiment, 25 μL of a dilute DNA sample was loaded slowly into each well, the electrical leads were connected, and the power supply was turned on for 20–40 s. Optimum voltage was approximately 84 V, or 6 V/cm for the 14 cm gels employed for most experiments. The electric field was then turned off, the leads were disconnected, and another 25 μL was loaded carefully into the same well. This loading and running process was then repeated multiple times. “Blowing out” the wells by pipetting electrophoresis buffer up and down into the wells before loading each new aliquot of sample, as is done in some polyacrylamide gel-based methods, did not improve results using SURE. In addition, we note that diluting samples using either water or TE (10 mM Tris [pH 8.0], 0.5 mM EDTA) prior to loading multiple times produced similar results. After all loadings were complete, gels were run until the tracking dye reached a line marked on the gel rig and then stained for 10 min using ethidium bromide. Six loadings of 25 μL into wells that can hold approximately 35 μL were used for most experiments. Loading volumes of 15, 20 or 25 μL consistently produced good results. Up to 800 μL could be loaded successfully in large gels (150 mL instead of 80 mL total volume) and as many as 20 loadings could be performed with minimal impact on the resultant quality of the bands. Negative images of photographs were created in most experiments to improve detection of faint bands.

3. Results

3.1. Dilute DNA samples can be loaded multiple times into the wells of agarose slab gels and produce single bands with minimal band broadening

The amount of a DNA sample that can be loaded onto a conventional agarose slab gel is limited because of the small sizes of the wells. We investigated the idea of successively reloading a well with the same DNA sample multiple times, running the DNA into the gel briefly after each loading, to increase the total number of DNA molecules in the resulting band. Our initial experiments employed a 2.7 kb PCR fragment that was run on a 0.8 % agarose gel using 1X TAE as electrophoresis buffer (Fig. 1A). The DNA was diluted, mixed with 6X purple loading dye (without SDS; New England Biolabs), and loaded onto an 11 × 14 cm gel. An undiluted sample, corresponding to ~30 ng of total DNA, was loaded into the first lane (labeled “C”). The original DNA solution was diluted 1:6 and then loaded six consecutive times into wells of the next lane, applying a constant voltage of 112 V (8 V/cm) for 20 s after each loading. The capacity of the wells was approximately 35 μL and a total of 150 μL was loaded, adding 25 μL at a time. The process was repeated in the next two lanes, where the diluted sample was loaded six consecutive times, running at 112 V for either 30 or 40 s before each subsequent loading. Note that the different heights of the bands are because of the staggered loadings, e.g., each new well was not loaded until after all successive reloadings of the previous well had been completed. When all loadings were finished the gel was run at 130 V until the tracking dye reaching a line marked on the gel rig and it was stained and photographed. The results indicated that the DNAs from all six loadings could be combined into a single relatively thin band and that the timing of the pulses was important; pulses of 112 V for 20 s after each loading produced the best results, while 40 s caused the band to broaden (Fig. 1A, left side). In the experiment shown on the right side of the gel in Fig. 1A, a constant pulse time of 30 s was used after each loading and the voltages were varied. Best results were observed at the lowest electric field strength of 84 V, where a single thin band was observed.

Fig. 1.

SURE electrophoresis of DNA fragments generated by PCR. (A) Dilute DNA samples were run into the gel briefly at the indicated voltages and times and the wells were subsequently reloaded repeatedly for a total of six loadings. After all loadings were complete, the tracking dye was run a defined distance down the gel before staining with ethidium bromide. C, a single loading of the original DNA sample before dilution. (B) Model for stacking of DNA molecules to form a single band from multiple sample loadings. (C) Unstacked bands produced by pulsing the voltage for too long after each loading (112 V for 1 min). (D) Model for production of unstacked bands under non-optimized conditions. Original DNAs used were 2.7 kb PCR fragments that had been cleaned and suspended in 10 mM Tris (pH 8.0).

These observations suggested that the DNA molecules in each sample, which are dispersed in three dimensions within the lower half of the well initially, are able to combine or “stack” upon each other and also upon the molecules in the subsequent loadings. It is likely that brief application of voltage causes the molecules to hit the gel surface and stack upon one another at the interface, followed by passage into the gel so that they are located just inside the gel matrix (Fig. 1B). When the next sample is loaded under optimal conditions, the new molecules move into the gel matrix and remain in close proximity to the nucleic acids from the previous load. The result is formation of a single apparent band composed of many more molecules than would be present after a single loading, leading to a stronger fluorescence signal after staining.

Proper stacking to form a single relatively thin band requires that the magnitude of the applied voltage not be too high and that the length of time it is applied after each loading not be too long. The image in Fig. 1C shows a broad band produced by pulsing the voltage for too long after each loading (112 V for 1 min). Under such non-optimal conditions the molecules in each loading migrated independently from each other and formed distinct small bands corresponding to the six loadings (Fig. 1C and D).

3.2. Band intensity can be influenced by the presence of EDTA

Commercial loading dyes such as the one used here (Purple Gel Loading Dye without SDS - New England Biolabs) contain Ficoll-400, a high density agent that causes the sample to fall to the bottom of the well, plus Tris, EDTA and one or more colored tracking dyes. Normally a 6X loading dye is diluted by 1:6 into the DNA sample. However, this might not be optimum for the SURE method because chemicals within the loading dye mixture could affect migration and stacking of subsequently loaded samples. To test this, SURE was performed using six 25 μL loadings run at 112 V for 30 s and using the same PCR fragment employed in Fig. 1. Band intensities were improved when the purple loading dye without SDS was diluted by more than 1:6, with darker bands at 1:8, 1:10 and 1:12 (Fig. 2A). Similarly improved results were seen when a commercial purple loading dye containing SDS was used (not shown). We note that the presence of SDS in the dye made reloading less efficient because samples did not fall as quickly to the bottoms of the wells.

Fig. 2.

Impact of components within commercial loading dyes and TAE electrophoresis buffer. (A) SURE electrophoresis band intensity increased when less loading dye was added to samples. (B) High levels of EDTA in the loading dye reduced band intensities. (C) Use of new loading dye without EDTA produces bands that are thinner and darker in TAE gels. (D) Use of TA electrophoresis buffer plus a loading dye without EDTA produces results similar to TAE.

The possibility that a component within the commercial dyes was interfering with band stacking was assessed by testing new loading dyes that contained Ficoll-400 and the tracking dye bromophenol blue with or without the other reagents, Tris and EDTA. Band intensities were found to be strongly affected by EDTA, but not Tris. The final concentration of EDTA in a sample containing 1X commercial dye is 10 mM; tests of loading dyes prepared with varying EDTA concentrations revealed that band intensity was reduced when this amount was used (Fig. 2B). Intensities were strongest and bands were thinnest when the final EDTA concentration was less than 5 mM.

These experiments indicated that strongest band intensities were produced by either (a) diluting commercial 6X dyes by more than the normal six-fold, e.g., 1:10 or 1:12 (Fig. 2A) or (b) using a simpler dye mixture that has a lower concentration of the metal chelator EDTA. We prepared a new dye solution with 0.02 % bromophenol blue plus the same concentrations of Ficoll-400 and Tris as the commercial dye, but without EDTA. A new TAE gel was run using the new 6X loading dye along with the same voltages and interval times that were employed in Fig. 1A. Use of the new dye without EDTA produced noticeably thinner and darker bands than before (compare Fig. 2C versus Fig. 1A). Regardless of the dye used, the strongest intensities and lowest band broadening effects occurred at the lowest voltages and interval times.

The TAE electrophoresis buffer within and surrounding the gel contained Tris, acetic acid and EDTA. The possibility that EDTA in this solution might also influence SURE DNA migration rates was tested by performing the same experiment as in Fig. 2C, but using TA electrophoresis buffer without EDTA (Fig. 2D). Band intensities were similar and strong using both buffers, with slightly thinner bands observed when TA was used, e.g., compare the bands produced using time intervals of 30 s and 40 s in Fig. 2C versus 2D.

3.3. Impact of molecular biology enzyme buffers and agarose gel strength

DNAs generated by PCR are synthesized in buffered solutions optimized for DNA polymerase enzyme activity. These buffers contain salts and other components that could potentially interfere with band stacking during SURE electrophoresis. To test this possibility, the 2.7 kb PCR fragment used previously was placed into different concentrations of the buffer used for its synthesis (Phusion DNA polymerase buffer). Using six loadings of 25 μL as before, a slight reduction of band intensity was observed at 0.75X buffer concentration when samples were run on a 1 % agarose gel, a common agarose percentage used in research labs (Fig. 3A). Lower buffer concentrations of 0.25X and 0.5X buffer had minimal impact on band signal strength or band broadening.

Fig. 3.

Impact of enzyme buffers and alternative electrophoresis buffers on SURE electrophoresis of PCR fragments. (A, B and C) A high concentration of Phusion DNA polymerase buffer can reduce band intensities; softer 0.5 % gels are more impacted than 1.0 % or 1.5 % gels. (D) Band quality using 1X TBE electrophoresis buffer and loading dye without EDTA is similar to that seen with 1X TAE. (E) Phusion DNA polymerase buffer produced less interference when 1X TBE was used.

Researchers frequently adjust the concentration of agarose in a gel in order to improve separation of molecules in different size ranges. Low concentrations such as 0.5 % produce softer gels with reduced gel matrix structural integrity, a factor that might impact the efficiency of band stacking at the interface. The intensities of bands were modestly reduced in both 0.5X and 0.75X Phusion buffer when a 0.5 % gel was tested (Fig. 3B), suggesting a small negative impact caused by the weaker gel matrix when compared to the 1 % gel in Fig. 3A. Use of a stronger gel that contained 1.5 % agarose produced similar results to those seen with 1 %, but bands were even thinner (Fig. 3C). Overall, these results indicate that the presence of the enzyme buffer can influence the amount of DNA in the final multiply-loaded band, but this effect can be minimized by modest dilution of the sample.

3.4. SURE electrophoresis works in both TAE and TBE

The most commonly used electrophoresis buffers in molecular biology labs are TAE and TBE [2,16]. Tests performed with 1X TBE indicated that band quality was similar to that seen with 1X TAE, e.g., compare Fig. 3D versus Fig. 2C. Interestingly, bands produced from samples containing Phusion PCR reaction buffer were consistently stronger in TBE gels than in TAE gels (compare 0.75X buffer in Fig. 3E vs. Fig. 3A).

3.5. Band quality remains high using different loading dye preparations and different forms of DNA

The high density agents that are used most often in commercially available loading dyes are Ficoll, glycerol and sucrose [7,8]. Assessment of new dyes prepared using glycerol or sucrose instead of Ficoll indicated that changing this component did not affect the results (Fig. 4A). Circular plasmids are routinely analyzed by agarose gel electrophoresis. They differ from linear DNAs in that they consist of multiple topoisomers that are identical in mass but have variable levels of supercoiling. The plasmid pBluescript II KS+ (2961 bp) was purified from E. coli cells, diluted and analyzed using six loadings of 25 μL as before. Intensities of both the supercoiled monomer and dimer bands were increased after six loadings, without substantial broadening (Fig. 4B).

Fig. 4.

Flexibility and sensitivity of the new method. (A) SURE electrophoresis of 2.7 kb PCR DNAs loaded six times using high density agents ficoll, glycerol or sucrose. (B) Enhanced signal produced by six loadings of a dilute solution of a circular plasmid DNA (pBluescript). (C) A solution of PCR fragments was diluted six-fold and the original sample was then loaded once on a gel and the diluted solution loaded six times. (D) Concentrated and diluted PCR DNA samples (three of each) from (C) were run on a gel and the DNA bands extracted and quantified by fluorometry. (E) Demonstration that DNAs that are undetectable after one loading can be readily visualized after multiple loadings. (F) Band quality remained high when loading as much as 800 μL of liquid into one well of a large gel. Volumes of 20, 40, 60 or 80 μL of different dilutions of a DNA ladder were loaded ten times, corresponding to different loaded volumes of 200, 400, 600 or 800 μL, while each well received the same total amount of DNA.

3.6. Sample loss due to inefficient DNA stacking is minimal under optimized electrophoresis conditions

A major potential advantage of SURE electrophoresis is that large amounts of a nucleic acid sample can be loaded into a single well and be subsequently purified by gel extraction. We tested the efficiency of recovery of PCR DNA fragments that had been diluted by 1:6 and loaded six times and compared it to the recovery of DNA from a single loading of the original undiluted sample. Examples of each type of loading are shown in Fig. 4C. Samples were run three times and the 2.7 kb bands were extracted and purified using a Monarch gel extraction kit. The concentration of DNA in each sample was subsequently determined using a Qubit fluorometer in conjunction with the Qubit HS (high sensitivity) DNA detection kit. Recovery of DNA loaded six times was 84.6 % as efficient as recovery of the concentrated sample that had been loaded only once (Fig. 4D). This result indicates that losses due to inefficient stacking are minimal when samples are loaded multiple times under optimized conditions.

3.7. SURE electrophoresis allows detection of DNAs in highly dilute solutions

Another potential application of the method is in detection of DNAs that are present at extremely low concentrations. To illustrate this capability, DNA ladder fragments that were 4, 3 and 2 kb in size were diluted to very low concentrations such that a single loading of 25 μL would produce bands containing only 18, 67 or 22 pg of DNA, respectively (Fig. 4E). Six loadings of 25 μL of this sample were then performed, followed by staining with SYBR Gold, which is a more sensitive stain than ethidium bromide. Bands produced after loading a single time were undetectable (Fig. 4E, left side). By contrast, the DNAs loaded using SURE had approximately six times more DNA in each band and were readily visualized. This result demonstrates that dilute samples containing as low as 0.7 pg/μL (or 0.0007 ng/μL; corresponding to the concentration of the 4 kb DNA fragment that was loaded six times in Fig. 4E) can be detected using successive reloading.

Most of the experiments performed in this study involved six loadings of 25 μL (150 μL total) into wells with a capacity of approximately 35 μL. The possibility that more loadings and higher total volumes might also work was investigated by creating a thicker (taller) gel than the ones used previously. The new gel contained 150 mL total volume instead of 80 mL and the height of the gel was 10 mm instead of 5 mm, producing deeper wells and allowing larger volumes to be loaded. A DNA ladder containing fragments ranging in size from 100 bp to 10,000 bp was diluted to different extents and then loaded onto a gel such that all lanes had the same total amount of DNA. Volumes of 20, 40, 60 or 80 μL of different dilutions were loaded ten times each into separate wells, corresponding to total loaded volumes of 200, 400, 600 or 800 μL. The volumes differed but the total mass of DNA added to each well was identical (83 ng total, with 10 ng in the 3 kb band) and the gel was stained with ethidium bromide (Fig. 4F). The intensities of the bands in all lanes were similar, indicating that use of larger loading volumes, even as much as 800 μL, was not detrimental.

4. Discussion

In this study we have characterized a new method for concentrating nucleic acids and enhancing band signal intensities in agarose gels. The experiments revealed that a DNA sample could be loaded into a well, run into the gel matrix at low voltage, and then a new aliquot of the sample loaded into the same well. Samples could be loaded multiple times and produce a single band containing the combined molecules from all of the loadings. The intensity and thickness of the band was dependent upon the magnitude of the applied voltage and the length of time that each sample was run before the next aliquot was loaded. Lower voltages and shorter run times produced the best results, likely due to more efficient stacking of the molecules adjacent to each other. The process is conceptually similar to the protein compaction that occurs during polyacrylamide gel electrophoresis of proteins using systems that employ both a stacking gel and a resolving gel. Protein molecules organized by isotachophoresis in a low percentage acrylamide stacking gel are run into a more solid (higher percentage) polyacrylamide resolving gel that causes them to stack upon each other at the interface [17–19]. This compaction reduces broadening effects that may be caused by large well sizes, large sample volumes, or by the random diffusion that occurs naturally during the run.

Band broadening was reduced and intensity increased when EDTA in the loading dye was reduced to less than 5 mM. We speculate that EDTA may chelate and remove some of the metal ions normally associated with the DNAs, changing their net charges to be more negative, and thereby interfere with aggregation (stacking) needed to form a thin band. The presence of a PCR enzyme buffer (Phusion DNA polymerase buffer) modestly reduced band intensities, but the effect was eliminated by diluting samples to 0.5X buffer or lower. TBE was less affected by the presence of buffer than TAE. The default high density agent in the loading dyes used here was Ficoll, but glycerol and sucrose were also shown to produce good results. The method worked well in both TAE and TBE electrophoresis buffers, with large gels and minigels, and using either linear DNA fragments or circular plasmid DNAs. Good results were also observed when gels contained ethidium bromide and therefore did not require a post-run staining step (not shown).

The yields of DNAs in the bands produced after multiple loadings was assessed by extracting them and measuring their concentrations using fluorometry. The tests revealed that 85 % of the initial amount of DNA present in the sample could be recovered after six loadings. If there is a consistent small loss after each loading, this number suggests that approximately 97 % of the DNA in each loaded sample is retained in the final band (0.97 × 0.97 × 0.97 × 0.97 × 0.97 × 0.97 = 83.3 % remaining). The result demonstrates that losses are small when optimized electrophoresis conditions are used.

The capabilities of the method were assessed in multiple ways. As many as 20 loadings could be performed using gels whose sizes are typical of those used in molecular biology labs (~5 mm thick). Employing larger gels with deeper wells, as much as 800 μL of sample could be loaded without reduction of band quality. Extremely dilute DNA solutions that did not produce a visible band on a gel, even after staining with SYBR Gold, became detectable using SURE. Tests demonstrated that solutions containing as low as 0.0007 ng/μL could be readily detected by simply loading the sample six times (Fig. 4E).

These results demonstrate that the sensitivity of detection of DNAs present at low concentrations can be strongly improved with the new approach. This signal amplification is advantageous for many types of experiments. For example, SURE allows the loading of large amounts of DNA into a single well for subsequent gel extraction, avoiding sample concentration handling steps such as alcohol precipitation or loading and elution from spin columns [20–22]. The method improves visualization of faint bands produced by PCR and other molecular biology enzyme reactions and may be useful for the analysis of extrachromosomal circular DNAs and other nucleic acids present at very low concentrations in prokaryotic or eukaryotic cells [23–25]. In addition, the new approach can potentially facilitate electrophoretic detection of nucleic acids present either in natural water and soil environments [26] or in human blood and urine samples [27–29].

Data availability

Data will be made available on request.