Dealing with VERY Mg sensitive PCR reactions - (Aug/09/2007 )

I'm a research technician, and run a number of genotyping PCRs. A couple of them seem to be extremely Mg sensitive; in my initial Mg titration I saw absolutely nothing at 2 and 2.5 mM MgCl, but got lovely bold bands with 3 mM. These reactions continue to be very touchy, and it seems like they often only work when prepared with a fresh (thawed only once or twice) vial of MgCl. I'm going to try higher Mg concentrations but could there be something going on that's breaking down or inactivating the Mg over time? Never had that happen before, but I haven't really worked with such 'touchy' reactions either... Thanks!

Magnesium ions can not break down. But it can be sequestered, either by chelators like EDTA, or precipitated out by carbonate ions. Neither is likely in the story you have describe... unless for some strange reason you are using magnesium hydrogen carbonate to provide magnesium ions.

I believe the problem is a case of insufficient mixing of the magnesium solution... it needs longer and move vigorous vortexing. Freeze thaw cycles tend to concentrate the ions in the centre of the tube, the last place to freeze. Thus when the tube is finally defrosted, the rest of the solution is left with a lower salt concentration. Try vortexing more and see if that helps.

Heh, no, I'm not using magnesium hydrogen carbonate!

I'll definitely make use of your suggestion - it makes a lot of sense. The Mg tube contains over a mL of 50 mM MgCl, and since the magnesium is supplemental (my 10x buffer contains 1.5 mM MgCl already) I'm not using much at a time which would make any variation in concentration such as you describe show up all the more. I've been vortexing it for about 10 seconds, but longer certainly won't hurt.

If this works, I owe you one!

This could also be variations in the DNA template you are using. EDTA is common in these, which will chelate varying amounts of Mg++. I would guess that you might also want to try some other additives, such as Betaine. The amount of dNTPs in the mix is also important in that it is coupled directly to the required Mg++ concentration. Adjusting the annealing temperature or extension time might also help.

Bit of an update, for those who might be in a similar situation...

Vortexing my Mg better didn't solve the problem (though it did help), so I did further Mg titrations. I brought the Mg up to 3.5 - 4 mM, and voila - beautiful, bold bands. One of the things that was aggravating me earlier is that the reactions were working in small lots of, say, 3 or 4 rxns but not with the larger numbers I need to do my genotyping. What I think happened is that when I did the initial titrations I did it using small lots where pipettor error was more significant - so I was adding more Mg than I thought - and just passed the threshold for it to work. My new protocol gives a bit more margin for error.

Moral of the story? Don't forget about pipettor error, and if something's inconsistent that you might be right on the edge of where it works/doesn't work. I'm thrilled to have this settled finally!