User:Norman Wang/DNA Isolation Via Agarose Gel Filter+Membrane Extraction (detailed)
Via David Thompson 
Here is a rough protocol. I'm including any notes/precautions/etc that I can think of. Its a very simple procedure, so despite all my extra notes, cutting the gel and putting in the filter should take you a minute.
I have done this with gels ranging from .7% to 1.2%. It might be tougher to slide the filter/membrane into higher percentage gels.
- load your sample(s), run the gel until you can resolve your band of interest.
- while it is running, prepare the filter paper and tubing. Cut both to the same size. sometimes i'll make my filter paper a mm or two narrower than the tubing. Make sure they are a few mm wider than your band of interest. i think mine are usually ~1cm tall, though this varies depending on how thick your gel is. You want enough sticking out of the gel to grab solidly with tweezers. If you have multibands that you want to avoid, cut another piece of tubing to block them.
- stop the gel, take the gel tray out of the box. visualize the bands with your blue light box or handheld longwave UV lamp and make a straight cut a few mm below your band (towards the + electrode). Cut all the way down through the gel to the tray. make sure your cut is wider than your filter paper/membrane so you can easily insert/adjust it. I will usually cut with a razor blade. I have also cut with the end of a mini- spatula when i had many close together lanes that i wanted to purify on the same gel, though it is harder to make straight cuts straight this way.
- put the filter paper and the membrane together and match up their edges. grab them with tweezers. i prefer bent tip tweezers so i can hold my hand at a comfortable angle and still grip a wider surface, which is important when trying to stuff them into the gel.
- make sure your gel is wet, maybe dip it into the box/buffer quickly and drain off the excess. this helps things to slide into the cut.
- with the filter paper and membrane together, paper towards - and membrane towards +, push them into the cut with the tweezers. Note that once they touch the gel and are wet, its hard to pull away and start over- the membrane will start curling, the corners can bend, etc. so try to get the bottom edge in on the first try. sometimes i'll angle to push one corner in first, sometimes i'll wiggle/slide from side to side as I push in etc. Do what you have to. Once it is at least partly in the gel, you can grab either end of the filter/membrane and push each end down alternately. If your filter/membrane is very wide, then sometimes pushing from the middle will not work or can can cause things to bunch up in the cut. You want to push the filter/membrane all the way to the bottom, so it touches the tray all along its width.
- If you have larger bands on the gel that you want to avoid, you can take your other cut membrane and insert it into a cut above the band of interest to block migration.
- put the gel/tray back in the box, and run it for a few more minutes to push the band into the filter paper. Make a note of where dye fronts are, and keep track of how far the gel has run. usually 5 minutes is enough for me, but this varies with band size, where you made your cut, etc. You can visualize the gel to see if you band has entered the filter yet. Usually i can literally see the glowing band of dna on my filter paper if i look at the gel from an angle. I have never run things too long, somehow pushing the dna out of the filter and around the membrane. So if you are worried about getting all of your band, you can easily run longer.
- Stop your gel, remove the tray from the box, drain off excess buffer, pull the filter/membrane out of the gel and put both into a collection tube. Do not squeeze too hard since any liquid you push out could have your DNA in it. Also, having excess buffer over your gel could make recovery worse. My method uses the ultrafree-mc tubes, so i stuff the filter+membrane into these and put into a ultracentrifuge for a minute on highest speed. If you are worried about getting all the stuff out, you can run it longer. when it is done the filter and membrane will be dry, and your DNA will be in the liquid below. If you do it this way, don't stuff it in with the membrane on the bottom, because you could trap your dna (i think, haven't tested it, but i do worry about it and make sure that i do not block the way with the membrane).
- from here i take the eluate and add 5x volume of PBI buffer (the yellow stuff from the qiaquick PCR clean up kit) and run it through a normal qiaquick procedure. Alternatively, you could probably EtOH precipitate it, though I have not tried.
Yield varies for this method. Sometimes the recovery sucks, sometimes its awesome. However, I ALWAYS get some dna, and its always clean unlike qia gel purified samples. Best recovery i have gotten is 5/6 of my input. Variation probably depends on how careful i was with putting my filter/membrane all the way to the bottom, and running the gel long enough to push large bands (like a vector) into the membrane. We don't use SYBR, so when I'm doing EtBr i only use longwave UV to minimize any mutations. The bands are correspondingly harder to see, and afterwards if i go back and look at it with full-strength short-wave UV, i'll usually see that there is some DNA left in the cut. If you're using SYBR, then maybe you'll be able to more reliably tell when your stuff is the filter.
- Thompson-2008 Personal communication with David Thompson of Liu Lab @ Harvard