Supposedly etablished MSP not working - E-Cadherin promotor (Jul/06/2010 )

Hi guys,

I don't usually study DNA Methylation, this is "just" a side project of mine. It sounded real easy at the start. I want to analyze E-Cadherin promotor methylation differences between cells (in culture) infected with different strains of my bacterium (the bacterial strains are my main project). Since this promotor has been analyzed quite thoroughly, I thought it wouldn't be hard. I found all necessary information in some papers: primer sequences, annealing temperatures, PCR conditions... these are two of my main sources:
Herman et al: Methylation-specific PCR: A novel PCR assay for methylation status of CpG islands, PNAS 93:9821-9826 (1996)
Kaise et al: CpG island hypermethylation of tumor-suppressor genes in H. pylori-infected non-neoplastic gastric mucosa is linked with gastric cancer risk, Helicobacter 13:35-41 (2008)

As a positive control for the MSP, I wanted to use completely methylated genomic DNA of the same cell line (methylated by M.SssI). But I get no PCR products. So I wanted to check if the methylation works, and digested the same DNA with ClaI or HpaII. But on a gel, the digestion pattern (which was really just one fat band) looked no different from not-methylated digested DNA. Am I misunderstanding something? This should work, right?

So as long as I cannot get my positive control to work, my other results aren't worth anything. But already I know that the MSP itself isn't working well, either, because I get unspecific bands (although the primers don't really bind anywhere else, according to PrimerBlast) and nothing is really reproducible.

I have not changed the PCR conditions very much because I figured that several papers can't be that wrong? I have, however, switched the annealing temperatures just for fun, and got even less product, but the bands appeared to be bigger (too big).

For the bisulfite conversion, I first tried a homebrew method but switched to a kit because I got no products whatsoever (I assume that pretty much all the DNA was degraded). I am now using Epigentek's Bisulflash kit, which doesn't seem to be popular on this forum, but it was cheap and I have no money. I have not yet checked if the conversion works because I just found out how to. I will try a digestion, but I don't know if it will be very informative for genomic DNA.

So I would be happy to get any input on this...
The biggest problem in doing MSP seems to be primer design, but since my primers have worked for other people for 15 years shouldn't they still be OK?

Thank you!

PS: By the way, this forum is awesome. I have learned a lot already today!

Just fyi, most commercial bisulfite kits don't need restriction digestion for your genomic DNA

epigenius on Jul 6 2010, 08:08 AM said:

Thank you, I know. I don't do a digestion for bisulfite modification. What I meant was, I will try to do a digestion after the modification to find out whether the modification is working.

>>As a positive control for the MSP, I wanted to use completely methylated genomic DNA of the same cell line (methylated by M.SssI). But I get no PCR products. So I wanted to check if the methylation works, and digested the same DNA with ClaI or HpaII. But on a gel, the digestion pattern (which was really just one fat band) looked no different from not-methylated digested DNA. Am I misunderstanding something? This should work, right?

It is unclear how you did the experiment--first PCR then digestion? What primers are used for the PCR--MSP primers?


>>So as long as I cannot get my positive control to work, my other results aren't worth anything. But already I know that the MSP itself isn't working well, either, because I get unspecific bands (although the primers don't really bind anywhere else, according to PrimerBlast) and nothing is really reproducible.

I don't trust MSP either. You should try bisulfite genomic sequencing if your purpose is to analyze Ecaderin in great detail rather than screen a lot of samples.


>>I have not changed the PCR conditions very much because I figured that several papers can't be that wrong? I have, however, switched the annealing temperatures just for fun, and got even less product, but the bands appeared to be bigger (too big).

Although you followed all the conditions used in the papers for PCR, there are still many other varibles that affect PCR results such as DNA template which I belive was not made the same way reported in the papers. In addition, the conditions given in the papers might not necessarily the actual conditions used.

>>For the bisulfite conversion, I first tried a homebrew method but switched to a kit because I got no products whatsoever (I assume that pretty much all the DNA was degraded). I am now using Epigentek's Bisulflash kit, which doesn't seem to be popular on this forum, but it was cheap and I have no money. I have not yet checked if the conversion works because I just found out how to. I will try a digestion, but I don't know if it will be very informative for genomic DNA.

I agree that digestion won't be very informative.

Hello pcrman, thank you for the reply.

For the control, I wanted to make my own completely methylated DNA to use for the MSP, so that I would see what an MSP on that DNA should look like. So I took genomic DNA from my cell line and methylated it with M.SssI. Afterwards I digested it and wanted to compare the restriction pattern to non-methylated DNA digested with the same methylation-sensitive enzyme. I wasn't expecting an actual clear band pattern or anything, I just thought that it should look different, but it doesn't. There's just one big band that's much bigger than all the marker bands, and then nothing.

I'll look into bisulfite genomic sequencing, but I might not have time to try it. Thank you for the suggestion though.

What you said about the PCR conditions makes sense. This is actually the first time that I really tried to use a protocol out of a paper, and I was expecting it to work, but lately I have gotten doubts that it's really that easy The MSP primers I copied out of the papers. The template is prepared in a different way, but from the same cell line. I was hoping that that would be enough. I guess it's not.

>>For the control, I wanted to make my own completely methylated DNA to use for the MSP, so that I would see what an MSP on that DNA should look like. So I took genomic DNA from my cell line and methylated it with M.SssI. Afterwards I digested it and wanted to compare the restriction pattern to non-methylated DNA digested with the same methylation-sensitive enzyme. I wasn't expecting an actual clear band pattern or anything, I just thought that it should look different, but it doesn't. There's just one big band that's much bigger than all the marker bands, and then nothing.

I see. So you digested genomic DNA and expected to see different pattern from methylated and unmethylated DNA. The big band you saw is probably the undigested genomic DNA. You should have seen a smear of DNA after digestion. I think there must be something wrong with your digestion reaction.

pcrman on Jul 10 2010, 09:24 PM said:

Yes, exactly. That's what I expected to see. I have tried the digestion twice with two different enzymes (ClaI and HpaII), but the picture is always the same. So I don't think it's the digestion, but rather the methylation that's not working, although I followed NEB's instructions for the methylation reaction and I don't understand what could possibly go wrong. The enzyme is new, it has never been left out or anything.

How much DNA did you use in the digestion reaction? Are you using too much DNA so that the marjority of it could not be cut?

pcrman on Jul 12 2010, 06:38 PM said:

I used up to 1 µg of DNA with 4-5 U of the restriction enzyme. For the methylation, I used at least 4 U/µg, as recommended by NEB. And I let both reactions go on for 2-3 h. I think that should be enough? I didn't want to digest longer for fear of star activity.

I came across a different problem that is puzzling me. When preparing my positive control (completely methylated DNA), I have to purify the methylated genomic DNA from the enzymatic methylation reaction. I realized that my cleanup kit is not designed for DNA fragments bigger than 40 kb, and we don't have anything more fitting. How do you do this? Is there special kits, or do you do a Phenol-Chloroform-extraction or something?