Recovering DNA from agarose gels

Paul N. Hengen, Ph.D. (July 14, 1999)

Introduction
Methods and reagents is a unique monthly column that highlights current discussions in the newsgroup bionet.molbio.methds-reagnts, available on the internet. A commonly occurring theme on the net is the recovery of DNA, and this month's column discusses the pros and cons of various methods used to extract DNA fragments directly from agarose gels. For details on how to partake in the newsgroup, see the accompanying box.

DNA electrophoresed through agarose gels is frequently used as a primary or re-amplification template for the polymerase chain reaction (PCR) as well as for hybridizations, sequencing, ligations, and many other molecular techniques. To recover the DNA, the band of interest is excised with a sterile scalpel and the DNA extracted by various means. Common practices include organic extraction with phenol and chloroform, use of electroelution devices, `freeze and squeeze' methods, `crush and soak' methods and a few different spin techniques.

The freeze and squeeze technique involves freezing the gel piece in liquid nitrogen within a micropipet tip [1] or centrifuge tube [2], and spinning out the liquid by centrifugation, while the crush and soak method involves adding buffer to the agarose slice before squashing it with a glass rod. The slurry is placed at 37 degrees C for several hours before the sample is centrifuged through siliconized glass wool or non-toxic polyallomer fibers.

Electroelution

An electrophoresis unit [3] or a mini-gel system can be used for DNA extraction by placing the gel fragments inside a dialysis bag along with electrophoresis buffer so that, when electroeluted, the trapped DNA can be recovered by precipitation. This can also be accomplished by using a centrifuge tube mini-unit [4-7], or microtiter dish with capillary tubes that span two of the titer-dish wells.

Another way is to cut a small trough ahead of the migrating DNA band and to electrophoretically elute the DNA onto diethylaminoethyl (DEAE)-cellulose paper [8], dialysis tubing [9], affinity membrane [10] or into a dead space in the gel containing 0.3 M sodium acetate pH 6.0, 10% sucrose [11]. The problem with this method is that the gel must be visualized with UV and constantly monitored to ensure collection of the sample.

Resin binding

A very popular method is to bind the DNA to silica particles by using commercially available binding resins, diatomaceous earth or glass fibers. This method was previously described in Methods and reagents [see TIBS 19,182-183]. Briefly, fragments from the gel are placed in a tube together with a chaotropic substance and a binding matrix, usually composed of a silica-based resin. The DNA is quickly bound to the resin and the complex is washed several times with a 70% ethanol. The DNA-matrix is centrifuged out of suspension, dried and the DNA is eluted in a small volume of sterile water.

Using this technique, DNA fragments up to 15 kb in length can be subcloned. However, the efficiency of recovery drops dramatically when the size of the fragment is larger than 6 kb. This presumably occurs because linear fragments bind to more than one particle at once, causing them to be physically sheared when pipeted.

Spin techniques

Another useful method is purification by placing the gel slice within a microfuge tube containing a membrane with a small pore size, such as a Costar Spin-X microfuge tube fitted with a 0.22 μm filter [12]. Alternatively, gel slices composed of low melting point (LMP) agarose can be placed in a microfuge at 70 degrees C until the gel has melted. The molten agarose is then quickly frozen and thawed, and the tube centrifuged for 5min. This technique has the advantage that very large fragments of DNA can be recovered with minimal damage, especially if care is taken to expose it to long-wave UV radiation (366 nm) for only short periods of time [13].

Netters have reported problems with this method - the precipitated material forms a pellet that does not easily dissolve in water, making it difficult to work with. This can be rectified by not precipitating, but using the supernatant directly instead.

Spin techniques may also cause co-elution of contaminants that could inhibit DNA ligase and presumably other enzymes, or perhaps interfere with transformation. In addition, the yield of DNA is usually not more than 60-70%. Some netters are enthusiastic about this technique, however, claiming that it works very consistently if the procedure is optimized [14].

GELase

Contaminating LMP agarose can also be removed by heating the gel slice to 65 degrees C for 10min, lowering the temperature to 40 degrees C, and then adding GELase[TM] (available from Epicentre Technologies), a combination of enzymes including beta-agarase, which degrades the agarose into multimeric subunits. [15,16] This technique generally gives high yields of DNA, but can only be used with the high-grade LMP agarose because it has fewer sulfate groups, which inhibit GELase activity. Unfortunately, this method is time consuming since the digestion takes several hours or overnight. In addition, the DNA sample may still require phenol extraction and precipitation.

There may also be some further problems with the method, since one netter reported that DNA gel-purified with beta-agarase cannot be amplified by PCR, while another reported that DNA fragments purified in this way could not be enzymatically labeled by nick translation for subsequent use as a probe in a hybridization experiment. The problem was only corrected by switching to a recovery protocol that didn't use agarase.

Syringe squeeze

All the above protocols are time consuming. A faster and easier way is the syringe-squeeze method of Li and Ownby [17], which can give 90-100% recovery in less than 30 sec. However, the down side is that DNA may not be sufficiently purified from contaminating agarose or buffer components for further manipulations. For example, one netter complained that DNA extracted in this way could not be ligated efficiently without an extra step of phenol/chloroform extraction.

Summing up

Netters have found that the best choice of extraction procedure is highly dependent on what the DNA is to be used for afterward. Typically, methods involving extraction with organic solvents, electroelution, or binding of the DNA to silica particles or ion-exchange resins give quite pure DNA, but yields are relatively low. The apparently poor yield of DNA recovered by the spin techniques has been found to be clean enough for ligating directly from the eluate without organic extraction or any other further manipulation. On the other hand, high-yield techniques tend to be problematic in enzyme reactions. Clearly, there is a distinct trade-off between recovery and the purity of the DNA sample.

References

[1] Koenen, M. (1989) Trends Genet. 5,137

[2] Heery, D. M., Gannon, F., and Powell, R. (1989) Trends Genet. 6,173

[3] Pollman, M. J., and Zuccarelli, A. J. (1989) Anal. Biochem. 181,12-17

[4] Karuppiah, N., and Kaufman, P. B. (1992) BioTechniques 13,368

[5] Pascali, V. L., et al. (1991) Electrophoresis 12,317-320

[6] Peloquin, J. J., and Platzer, E. G. (1991) BioTechniques 10,159-160

[7] Sandhu, G. S., and Kline, B. C. (1989) BioTechniques 7,822-823

[8] Dretzen, G., et al. (1981) Anal. Biochem. 112,295-298

[9] Girvitz, S. C., et al. (1980) Anal. Biochem. 106,492-496

[10] Zhu, J., et al. (1985) Bio/Technology 3,1014-1016

[11] Hansen, H., Lemke, H., and Bodner, U. (1993) BioTechniques 14,28-30

[12] Schwarz, H., and Whitton, J. L. (1992) BioTechniques 13,205-206

[13] Hartman, P. S. (1991) BioTechniques 11,747-748

[14] He, M., et al. (1992) Genetic Analysis Techniques and Applications 9,31-33

[15] Gold, P. (1992) BioTechniques 13,132-134

[16] Serwer, P., et al. (1992) Biochemistry 31,8397-8405

[17] Li, Q., and Ownby, C. L. (1993) BioTechniques 15,976-978