PE bisulfite seq: could the EB solution inhibit the PCR rxn?

[Simone78]

Hello everybody,
I am performing Paired End BISULFITE sequencing, with a modified Illumina protocol. I do the Covaris shearing, end repair, adenylation, ligation and size selection (E-gel), followed by bisulfite treatment using the EZ DNA methylation kit.
I then performed the elution in 30 ul EB solution (according to the protocol) but I recently ran into some problems in the next step (PCR).
I use 1-3 ul of bisulfite-treated DNA for PCR and, even if it looks I have enough DNA for seq, the library complexity is not enough to guarantee good results.
I then thought that, if I use more DNA for the PCR I will get more but, on the contrary, I get even less!
It seems there is an inverse correlation between input DNA and amount of PCR product. Could it be due to an inhibiting effect of the EB solution? Should I rather do the elution in TE or water?

Finally, it is well known that sodium bisulfite damages the DNA but I also noticed that it seems there is no relation between amount of DNA after bisulfite treatment and after PCR, as if not all the DNA is "amplifiable". This makes it very hard to predict the amount of input DNA and number of cycles needed for PCR. Does anybody have seen the same thing?
Thank you very much!

[C.R.]

Hi, which protocol do you actually you refer to, since there are several publications out for shotgun bisulfite sequencing. One example: http://www.ncbi.nlm.nih.gov/pubmed/21289626
There the converted library is enriched by 4–8 cycles of PCR.

[Simone78]

yes, that is very similar to what we do. But the closest to our protocol is probably this one:

http://www.ncbi.nlm.nih.gov/pubmed/21072239

[GW_OK]

If you are using the same EB that I'm using it is 10mM Tris, pH 8, which shouldn't have any affect on PCR. I wouldn't use TE for any PCR reaction, as the "E" in "TE" is EDTA, which is Not Good for the Mg2+ the PCR reaction needs.

Could your primers somehow be the limiting factor, with too much template preventing good amplification?

[Simone78]

yes, I sadly realized that, even if I do the elution in water (after bisulfite treatment) I get the same results as with EB.
I don't think the primers are the limiting factor, since I have so little DNA left after size selection and sodium bisulfite. My primers are 40 uM, then I mix equal volumes of F and R and I take 1.25 ul of this solution (PCR in 50 ul). There are still tons of primer (dimers?) after Qiagen column, as it's seen with the Bioanalyzer.
I think it all depends on the harsh treatment with sodium bisulfite, which destroys most of the DNA. Unfortunately, I didn't notice any clear relationship between what I get after size selection and what is left after bisulfite conversion, which makes it difficult to predict how much DNA I need in order to get a good library.

[csquared]

I apologize in advance for asking what may be an obvious question:

You are using methylated adaptors correct? If not, the bisulfite treatment will cause the amplification primers to no longer be complimentary.

EB will not have any effect on PCR, but it looks like your water results confirm that the EB is not the issue.

[Simone78]

yes, we are using Early Access Methylation Adapter Oligos (Illumina) and PE PCR primers. Today I did 12 cycles and I got 5-6 ng DNA/ul. I perform the purification with Ampure XP beads, (in order to get rid of the primers, which are retained by the Qiagen columns) so what I have is "usable" DNA. Even though I see a small peak with the High Sensitivity chip after 8 cycles, it looks much better after 12 cycles. Unfortunately this doesn't really help me...I'm simply getting more of the same molecules...
I read that we need around 8 x 10E+08 different molecules / lane. If I back-calculate what I have after PCR (an increase of 256 times of the initial DNA when I perform 8 cycles and of 4096 times when I perform 12 cycles), then it looks like my library has 1/10 of the complexity that is needed for a good run!