Bisulfite sequencing problems - T>C conversion (Mar/14/2005 )

Hello, I was wondering if somebody can help me with the following.

I have sequenced, after bisulfite (CpGenome, Chemicon), first PCR: 45 cycles, second nested PCR 30-35 cycles (SigmaREDTaq) and subsequent cloning, a 635 bp product, containing around 60 CpGs. Sources are different genomic DNA samples.

Although my results are negative for hypermethylation , I'm stuck with the following. In some of my samples (not all) I find a T>C conversion after sequencing. There is a clear preference for the location of this conversion (a total of 4 locations, although the locations are shared in different samples, not all combinations are the same).

Is this an artifact? (Due to 2x PCR? it occurs also in a few samples which I cloned without nested PCR) (Due to bisulfite? it doesn't occur in all samples treated at the same time).

A suggested mutation in the genomic DNA does not make sense to me, since normally this cytosine would have been converted after bisulfite (I'm now anyway trying to sequence directly).

I would be grateful to all suggestions.

---

Hi Hobbes,

I don't think it was coincidence that the same region gave this "artifact" is this region very G-rich?

I suspect that you have incomplete bisulfite conversion and that this region was not denatured properly during the bisulfite conversion.

I have not used the chemicon kit so i am not familiar with the protocol. To denature the DNA I ensure the DNA is denatured by placing it at 100C for five minutes prior to and after bisulfite conversion.

Another reason could be that you started with a gDNA prep with incomplete proteinase K treatment, you maybe seeing a footprint of a protein still bound to that region you are seeing that is unconverted.

All the best

Nick

---

Hi Nick,

Many thanks for the quick response. However, I'm not sure whether I understand it completely, sorry.

"I don't think it was coincidence that the same region gave this "artifact" is this region very G-rich?"

I think so too, it doesn't seem to occur at random, so there should be some explanation. The region is indeed G-rich.

"I suspect that you have incomplete bisulfite conversion and that this region was not denatured properly during the bisulfite conversion."

However, in case of incomplete bisulfite conversion for this region I would expect neighboring CpGs (within 2-10 bp) also to be non-converted or am I mistaken here?

Furthermore, it is a T>C conversion according to the database sequence (not C>T as in a bisulfite reaction, with a remaining C in case of non-conversion).

The option of a footprint of a protein sounds interesting, although how to explain the clear T>C conversion with that?

Thanks a lot! Hobbes

---

Hi Hobbes,

First, I think T->C conversion by bisulfite can be ruled out based on the experiments of Hayatsu et al (Biochemistry 1970, the first paper describing bisulfite reaction). They found bisulfite reaction was base specific in which only C ->U conversion and at some condition C->T conversion were possible, while other bases had no action with bisulfite. So the T->C conversion must be either PCR or sequencing error, although I have never noticed it myself. I assume there are no SNPs at those site, right?

---

Hi pcrman,

Thanks for your answer. So if we rule out bisulfite, PCR error seems most likely. There are no reported SNPs in this area. I'm now trying to get a sequence from this region (70-80% GC-rich......).

(sequencing error is unlikely as repeated sequencing of different cultures of the same clone gave the same result).

Hobbes

---

To me it sounds like you have a SNP in your region - One thing to note is that not all SNPs have been detected - just because it is not in a SNP database doesn't mean it isn't there.

Jon

---

ahh apologies for my misunderstanding.



So your T>C conversion you are seeing when you PCR and sequence the original DNA prior to bisulfite, does sound like a SNP to me.

Nick

---

SNP is probably most likely.

However, this still doesn't explain the following:

In the DNA prior to bisulfite, there is supposed to be (everything seems to depend on whether I can confirm that) SNP: T > C

After bisulfite, there should be a conversion: C > T


So why at this SNP location, after bisulfite, I still find a cytosine???

Adjacent CpGs are converted, so incomplete conversion is unlikely.

Hobbes

---

Very Curious Hobbes,

what system are you working with? Mammalian or Vegetable (Plant I mean)?

If you are working with plant (maybe bacteria also) and you can rule out the SNP you could very well be seeing unconventional DNA methylation at non-CpG sites, ie: CpNpG sites as has been observed in plants. A paper has been published where this has been observed in humans, although I have yet to see it.

The mystery continues...........

Nick

---

I'm working with human DNA.

That is the main question, is this something unconventional/new finding or a 'simple' artefact.

I'll keep you guys informed

Hobbes

---