Protocol Online logo
Top : New Forum Archives (2009-): : DNA Methylation and Epigenetics

Bisulfite PCR Inconsistent, often smears - (Jul/14/2009 )

Hi all,

I'm an undergrad working to optimize a bisulfite assay that is compatible with the promoter of the gene our lab works with. I currently have 4 overlapping primer sets that I'm using to "walk" down the promoter, which happens to be one giant CG island, with a GC content of 60-85%. Thus far, I've been using two primer sets for each region. One specific to normal DNA, and one specific to bisulfite treated DNA, and both avoiding CG sites. Upon treatment and PCR, I can then compare the presence of bands between the two in order to verify that I have a fairly high conversion rate. I've tested out a handful of protocols, and even a kit, and have suffered incomplete conversion (bands produced by primer sets specific to converted AND unconverted DNA) in all 4 regions. Because of this problem, I consequently know that all 4 bisulfite-treated-specific primer sets at least work, if nothing else.

In my most recent attempt to convert DNA, I was successful in converting enough of my DNA so that the untreated-specific-primers would not amplify a product (thus suggesting a high conversion %) Furthermore, the first of my 4 primer sets specific to bisulfite treated DNA amplified a spectacular band. The other three primer sets, however, saw long smears from well to end, except for one, which had a faint background band of the right size in the middle of the smear, but not sharp enough for me to try to isolate and clone. It should also be noted that I use sheared salmon sperm DNA to act as a carrier when precipitating my bisulfite product.

Is there a way to remove this smear phenomenon? Furthermore and more importantly, does anyone have thoughts on getting my pesky A/T rich primers to work?

PCR has 40 cycles with 49 degree annealing temp.

Potential problems could be:

-Partial conversion is causing neither (treated/untreated specific sets) to work, but I find this unlikely.
-Salt concentrations are too high
-Too much DNA/carrier DNA? In the one primer set that was specific to treated DNA that finally generated a hot band, the well to end smear was much less pronounced. Furthermore, I tried making a new primer set for a region that barely ever generated a product, and the new set does not generate a band ever, nor a smear...
-I've also heard that impurities in the DNA can generate smears. I should mention that I used a very intense genomic DNA purification protocol that should help eliminate that problem.

Sorry for the long post/question, and I can provide PCR conditions and bisulfite protocol if it helps. Thanks to anyone with advice.

-Brett-

Brett on Jul 14 2009, 08:12 PM said:

Potential problems could be:

-Partial conversion is causing neither (treated/untreated specific sets) to work, but I find this unlikely.
-Salt concentrations are too high
-Too much DNA/carrier DNA? In the one primer set that was specific to treated DNA that finally generated a hot band, the well to end smear was much less pronounced. Furthermore, I tried making a new primer set for a region that barely ever generated a product, and the new set does not generate a band ever, nor a smear...
-I've also heard that impurities in the DNA can generate smears. I should mention that I used a very intense genomic DNA purification protocol that should help eliminate that problem.

Sorry for the long post/question, and I can provide PCR conditions and bisulfite protocol if it helps. Thanks to anyone with advice.



I agree, incomplete conversion seems to be unlikely. From my point of view an annealing temp of 49C indicates short primers. Due to the reduces complexity of the DNA after bisulfite treatment you should try to design longer and therefore more specific primers or to increase the annealing temp to avoid unspecific amplification.

What type of carrier do you use? Avoid any type of genomic DNA (yeast, salmon sperm, etc...) but switch to a less complex carrier, like for instance poly(A).

Hope that helps...

MoB

-MoB-

My primers are super AT rich because of the high GC% of my DNA, but they are between 25-30 base pairs long each. I do use sheared salmon sperm as a carrier, about 10 ug of it actually. I often see protocols describing the use of glycogen. Thoughts on that? I'll see if our lab has any poly A DNA as well. Thanks for the input though.

-Brett-

Brett on Jul 15 2009, 07:12 PM said:

My primers are super AT rich because of the high GC% of my DNA, but they are between 25-30 base pairs long each. I do use sheared salmon sperm as a carrier, about 10 ug of it actually. I often see protocols describing the use of glycogen. Thoughts on that? I'll see if our lab has any poly A DNA as well. Thanks for the input though.


I think that's the problem: 25-30 bases and an annealing of 49C indicates a low complexity of your primer and this favours unspecific primer-binding- and amplification.

Glycogen will do it as well. But avoid DNA-carrier, I recommend to you poly(A) = RNA. poly(dA) might work as a primer and will also result in unspecific amplification.

Anyhow in my opinion a carrier is only needed if the starting amount of your DNA is extremely low. Depending on your bisulfite-protocol everything above 100 ng DNA won't need a carrier.

Best,

MoB

-MoB-

My protocol starts with 10 ug of DNA. I assume (and then normalize) a 75% recovery post digest/phenol/chloroform cleanup, so I add 7.5 ug of DNA in each of my rxns. I've always kind of wondered about the carrier addition. Starting with that 7.5 ug, I end up resuspending in 100 uL of TE. I'll try it without the carrier in a day or two and post back how that worked out, perhaps a side by side reaction to see if one works better than the other. Thanks for the ideas.

As a side note, do most labs that do this assay have their own alterations? So few papers use the same methods, unless they use kit that is, and I wondered if its common for this assay to have to be customized to each DNA region's needs.

-Brett-