Your Cart(0)

DNA Purification from Agarose Gels


By John Schloendorn



Over the past couple of years, I had the good fortune to learn a whole lot about DNA, while assembling Gene And Cell Technologies’ molecular armamentarium from my own genome and a variety of other sources.  Today I’d like to highlight one particular aspect of the process, the extraction of DNA fragments from agarose gels for subsequent cloning.  

Cloning experiments of course require excellent quality DNA fragments, absent both DNA and non-DNA contaminants.  The kinds of molecules that need to go include damaged plasmids, uncut plasmids, PCR template carryover, false PCR products, primer dimers, unused primers, agarose, gel dyes, buffer salts and solvents.

Generally the most effective way to get rid of both DNA and non-DNA contaminants is to resolve them from our fragment by agarose gel electrophoresis, and then to physically cut out the band representing our desired fragment.  This is very effective in removing wrongly-sized DNA contaminants that virtually no other method can get.  But in doing so, we introduce a major non-DNA contaminant, namely the agarose gel itself.  This necessitates an effective way to extract our fragment from the surrounding gel slice.

Some of you may be thinking:  But isn’t gel-purification a solved problem?  Don’t we have convenient spin-column kits for that?  Well.  Definitely maybe. 


Chemical Melting / Spin Column Kits

Commercial gel purification kits generally employ a set of so-called proprietary buffers that melt an excised gel fragment under elevated temperature, followed by a spin-column purification scheme using silica-adsorption or ion-exchange chromatography.  Those of us willing to dig into the decades-old literature on this topic are of course able to assemble their own kits at a small fraction of the cost.

Unfortunately, both the commercial kits and DIY chemistries have strict size-limitations, work only in a narrow concentration range, suffer seemingly random failures, and simply never achieve the advertised yield and purity.  If you don’t believe me, try measuring your DNA concentration before and after gel purification…  I do not want to blame specific manufacutrers for this.  It seems that all manufacturers suffer these problems equally.  This broad lack of reliability, throughout the industry, is one reason we at Gene And Cell Technologies decided against offering such a kit product ourselves in our own store.  

Despite their limitations, I think there would still be a place for the commercial kits in a complex cloning strategy, if the limitations were clearly acknowledged and quantitatively described.  But unfortunately no manufacturer systematically reports the bounds of the rather narrow space of (fragment size) x (DNA amount) x (GC-content) x (…) in which their kit works well enough for subsequent cloning.  When your fragment is long, difficult, low concentration, or in some other way unconventional, then you are stuck with discovering by trial and error whether the kit is for you.  Since my sequences tend to be at least one of the above, this creates so much uncertaincy that I generally stopped working with the kits altogether in my own experiments.

But don't get me wrong -- a lot of the time, the purity of a gel-prep with a commercial kit is great.  The yield is usually far below specifications (>70-80%).  In reality, if I get 10%, I’d be happy.  The discrepancy may be perceived as annoying, but it doesn’t necessarily prevent cloning, as long as the purity is there.



(Figure 1) - The problem:  UV spectrum of DNA prepared by "gel purification kit", undisclosed manufacturer.  Large contaminants absorbing in the 230 nm range obscure any signal from the DNA.  In my hands, replicas of the same band will either all work semi-reliably with something like ~25% random failure rate, or for another fragment, it will simply not work at all.  This seems to happen equally with all the major manufactures.  I tried.



In theory, electroelution is a cleaner way to extract your DNA, because there is nothing involved in it that isn’t already involved in running your gel.  If you can resolve your fragment on the gel at all, then it should be possible to run it all the way through the gel and catch it in a suitable container when it comes out at the edge.  Therefore, the main question is, what is a suitable container, and how do we arrange the geometry of it all so that we can efficiently catch our band?

Dialysis enclosure method

One way to do gel electroelution is to excise a the band in a narrow slice, and place it inside a dialysis membrane.  Any of the shelf “snake skin” with 10kd or another suitable cutoff will work fine.  Pick a small-diameter tubing and choose narrow clips.  There are commercial electroelution chambers that accomplish this with maximum convenience for ~$3 / prep.

Place the slice enclosed in the membrane inside your regular gel-tray, completely submerged in buffer.  If there are air-bubbles trapped in the membrane, it may be necessary to drop some suitable weight on it, in order to submerge it completely.  For a regular-sized slice (200 mg – 400 mg), DNA is eluted for ~60 min under normal conditions, e.g. 100V.  You can follow the elution with a handheld lamp matching your gel dye (see the chapter on UV and Vis exposure below).

When the DNA has completely left the gel slice, most of it will be stuck at the opposing wall of the dialysis tubing.  Reverse the voltage for 2 minutes, to knock it of the wall.  Then, retrieve the DNA from the dialysis tubing with a pipette.  Mix vigorously to wash the residual DNA off the gel slice and off the tubing walls, before retrieving it.

Trough method

An even more straightforward way to capture your band is to cut a trough into the gel right in front of your band, and run the band into the trough.  Then, it can be retrieved with a regular pipet tip.

The only complication of this method is that some kind of blockade is necessary to keep the DNA from zipping through the buffer-filled trough as soon as it enters it.  Some people report using a single piece of dialysis membrane or filter paper inside the trough, but I have not tried that myself.  I generally use PEG 8000 to trap my DNA in the trough. 

Cut a trough, slightly larger than the band itself, or slightly U-shaped.  This is necessary, because the electrical current will prefer to avoid the PEG 8000 - filled areas, causing your DNA to try to weasel its way around the trough rather than getting trapped in it.  It takes some experience to get the geometry of this right.  It’s best to practice on some disposable bands, before using this skill on any critical bands.

After cutting the trough, remove any residual running buffer from it with a pipette.  Then, fill the trough with 20% PEG 8000 dissolved in running buffer (PEG in dH2O will make your DNA try even harder to route its way around the trough).  Continue running the gel for approximately another 10 minutes at 100V (less for small bands, more for large bands), to push your DNA into the trough.  The DNA should be clearly visible as a fluorescent blob in the middle of the trough at this point.  Continue to observe under appropriate illumination, as you retrieve the DNA from the trough by pipette.



(Figure 2) Removing DNA from a PEG-filled trough under blue/green illumination.  It's easy to follow the DNA under blue/green illumination all the way into the pipet tip.  Note how the DNA tries to run around the trough.  A larger or slightly U-shaped trough can prevent this problem. 


Next, the DNA fragment needs to be separated from the added PEG.  This can be done by ethanol precipitation (1/10th volume NaAc pH 4.8;  2.5x volume ethanol;  incubate at -20oC for >1h), or through a commercial “reaction clean-up” spin-column kit.  In my hands, the reaction clean-up kits are much more reliable and have much better recovery than the gel clean-up kits.  Similarly, a combination kit, which can do both reaction mixtures and gel slices, will work much better when used on this PEG solution than when it is used on a solid gel slice.



(Figure 3) Much better!  Even at low concentration, the trough elution method reliably eliminates nearly all UV-absorbing contaminants.


Harmful Effects of UV and Vis Exposure

It is often said that exposure of DNA to UV radiation during gel electrophoresis may damage the DNA and compromise subsequent cloning experiments.  Unfortunately, cloning experiments can often fail for all kinds of reasons, and this combination of high unpredictable failure rates and invisible dangers creates a fertile breding ground for superstition.  The systematic study of the effects of UV exposure on transformation efficiency remains rare.  Here is one notable exception from China.  They use two wavelengths of uv, and find that at 302 nm, no more than seconds of UV exposure are acceptable, and at 365 nm, minutes are acceptable.  Here is another report confirming the 302 nm result. 

These studies have a couple of limitations:  They use UV exposure in a cuvette, do not report the dose, and do not compare it against blue/green dyes.  We can’t really apply these lessons quantiatively to a gel purification / ligation scenario, where the geometry and circumstances different.  Blue/green dyes might by some theories be even less mutagenic than long UV, but in practice, I’d like to see that data before I believe it.  That said, I have cloned many bands prepared under blue light, and had many successful ligations, so I know there is at least some way to use these dyes safely. 

From these anecdotal results, it seems safe to use either long UV (365) or blue light for the time reasonably required to extract a single band (seconds to a few minutes).  More systematic studies are needed to compare the damage caused by the different methods quantitatively, and in the context of sequence features that might plausibly make your DNA particularly vulnerable (length, high T content, secondary structures, other?).



I hope this document helps you take care of one important aspect of your cloning workflow – the retrieval of DNA fragments from gel slices.  Please feel free to ask any questions or share your own experiences using the comment box below. 







Perhaps a simple hack from South Africa?

- Cut the gel band out
- Make a hole in the bottom of a 0.5mL PCR tube with a hot, small gauge needle.
- Place a small wad of compressed fishtank wool in the bottom of the tube (can use tweezers)
- Place the gel band in the PCR tube (I used to chop it up), cut the lid off, but leave some connector on.
- Place into a larger 1.5mL tube, and centrifuge at high speed for a minute. You’ll have to arrange the lids as you can’t close them with the PCR tube in.
- Make up the volume with buffer
- P:C clean, precipitate with ethanol.
- All done

Posted by Liam Thompson on August 18, 2016

Hi Raf,
Not sure about glycerol. We haven’t tried it. My theory would be that the long chains of the PEG would help to trap the DNA, and you would not get that with a small molecule like glycerol.

But as the saying goes, it theory practice is just like theory, but in practice it’s not. So if you get to try it, please let us know what happens!

Posted by John Schloendorn on July 29, 2015

would glycerol be ok instead of PEG8000?

Posted by Raf on July 22, 2015

great article, I will be trying it soon.

Posted by Li on July 22, 2015

I use TBE. In my hands, I cannot get the DNA to stay in the well for long enough to retrieve it, in running buffer alone. (or at least I’ve never succeded in timing it correctly, trying only once or twice). So I’ve always added the 20% PEG. I never tried cloning it with the PEG in there. I’ve always used ethanol precipitation to clean it up.

Posted by John Schloendorn on July 10, 2014


Great blog. As someone who is currently trying to purify DNA from agarose gels for subcloning, I know the headaches that can result from using Gel Purification kits. I had a quick question; for the Dialysis Tube electroelution method, after getting the DNA out of the gel and have it in your running buffer (we use TAE) is it ready to be used in downstream applications (i.e. subcloning)? Or would I need to do some form of DNA precipitation?

Thanks in advance,

Posted by Sean on July 03, 2014

Heh, thanks Nicole. Yes I remember the “am I crazy” feeling from grad school vividly. I figured someone had to say it ;-)

Hopefully this can help all the crazy-feeling grad students out there see the light at the end of the tunnel. The number of things that can go wrong is large, but it’s finite. After enough of it happened to you, things will get noticeably more reliable.

Posted by John Schloendorn on June 13, 2014

John Schloendorn,

This blog trumps all the molecular biology forums I have ever visited! Thank you! I wish this had been around when I was a first year grad student.

I really appreciate that you confront all of the common – yet never talked about- strange things that happen during cloning and with the use of certain types of DNA kits. I used to feel a little crazy, but now I know all of these things are normal, even if not disclosed in protocols or manuals.


Posted by Nicole on June 13, 2014

Leave a comment

Comments have to be approved before showing up

We Accept...

We Accept Visa We Accept Discover We Accept American Express We Accept Mastercard We Accept PayPal We Accept Google Checkout