Bisulfite Sequencing PCR help! - BSP is Failing like Gangbusters. Please help! (Dec/11/2009 )

Hi Everyone,

First of all this forum is a Godsend for me given that I've been banging my head on a wall trying to get my Bilsulfite PCRs to work for a month now with absolutely no luck. By that I mean I think I've gotten the bisulfite treatment part down but that my primer design for whatever reason is the cause of complete and utter failure. Please read and let me know what you think. I apologize that it's long but if I figure if I put it all out there it will help with troubleshooting. Thanks guys!

Here is my bisulfite treatment protocol:

- Either digest 4 ug of genomic DNA with a restriction enzyme that doesn't cut in my targeted CpG islands (I digest 4 ug because I lose 50% of this during column purification in preparation for the next step) OR run 2 ug DNA through a 26.5 gauge needle 5X to shear it.
- I UV spec DNA at this point just to make sure I have enough going into bisulfite treatment. I usually have 1.5 - 2 ug going forward. So far so good.
- Denature DNA in FRESH 0.3 M NaOH and incubate at 37C for 20 mins
- Add FRESH Bisulfite/Hydroquinone (final concs 2.4 M/0.5 mM respectively) and incubate at 50C for 16 hours wrapped in foil
- Column purify into 40 ul H2O (Promega Wizard)
- Add 4.5 ul of FRESH 3M NaOH to desulfonate at room temp for 5 mins.
- Ethanol precipitate with ammonium acetate (no glycogen) overnight. At this point I get good pellets so I figure I'm good so far. NOTE: I spin multiple times and pipette out the 100% EtOH. I do not wash with 70% EtOH before resuspension in 1X TE. Is this crucial?
- Resuspend in TE. Store at -20.

At this point I test for full conversion. I take my bisulfite treated DNA and UV spec it and it always has a positive reading (~30 ng/ul or almost 2 ug usually each time) with OD 260/280 readings from 1.2 - 1.5. Furthermore I've tried doing regular PCRs (amplicon size= 300 bp) using non-bisulfite PCR primers on bisulfite-treated DNA to see if I can amplify any residual unconverted DNA and in most cases it comes up negative (a few cases here and there are positive which means incomplete conversion). If my bisulfite-treated DNA passes these two checks then I assume it's good-to-go for bisulfite PCRs. Is this assumption correct? Should I be checking for something else?

Next is PCR time where usually a lot of cursing occurs. I've tried manually designing primers and using MethPrimer as a primer design tool. In all cases I have never gotten a single PCR band. I've tried about 10 different PCRs now and in a subset I've even tried hemi-nested and fully-nested PCRs and they've all failed to yield any bands. These amplicons are 300-500 bp long so I don't think I'm pushing the envelope too hard here. My standard thermocycler profile is as follows:

95C for 5 mins

40 cycles:
95C for 30 sec
45C for 30 sec
72C for 3 mins

72C for 7 mins
4C hold

I used to do 55 anneal but when I was getting no bands, I switched to 45C anneal. After this switch, I did nested and hemi-nested PCRs on a some samples but those didn't work either. In some cases I see some primer-dimers but most of the time there's absolutely nothing even when I run the whole PCR reaction out on a gel.

My PCR conditions (50 ul reactions) are as follows:
200 nM each primer
200 uM dNTPs each
3 mM MgCl in standard Roche 10X PCR buffer
2.5 U Platinum Taq (Invitrogen) -- hotstart and no proofreading (I think).
1X Betaine
~100 ng of bisulfite-treated DNA in TE per reaction (usually ~3 ul)

In regards to primer deisgn, I've been very careful not to include any CpGs within my primer sequences. Just to give you an idea of some of the primers I've been working with:
5'- TTTTGTTTTTTTTTAAGTTAATTGT
5'- CATCCCTCCACTTCTTCCCCATAAAAAAT

5'- TTTCCTAAACAATAAAAAAAAACAT
5'- GATTTTATGATAAATTGTTTTAGTT

5'- AGTAGTTTTTTTTTTAGGAGTGAAGGAGGTTA
5'- CTTCTCAAACTCCTCCTCTCCCCTTA

5'- TAAGGGGAGAGGAGGAGTTTGAGAAG
5'- AAAATACCTTCAACCAATCACCTCAATACCT

I know some of my primers have long stretches of Ts which aren't a good thing but these tend to be the 5' portion of the primer and not the 3' half. At this point, I am at a loss as to where to go from here despite having read this forum religiously over the past day and a half. The biggest questions in my head:

Is my bisulfite treatment of DNA working? I think it is but maybe I'm missing something.

Should I have washed my pellet with 70% EtOH to desalt? I don't think this is that crucial, but maybe it is?

Are my primers poorly designed? I'm thinking yes and this is probably the cause of my failure, but I have no way to confirm since my 'positive control' primers (i.e. those taken from published papers run under their published conditions) have also failed. Maybe my PCR conditions aren't optimal. Major ARGH.

The only things I can think of at this point are to try a new bisulfite treatment method (agarose beads) or maybe try Methyl Primer Express as a primer design tool, but I suspect it's something other than that. Any suggestions as to where to go from here would be appreciated. As Tiger Woods would say, "Quickly. Huge. Bye."

Thanks!
- Chris

i think your problem is in the primer design, methprimer just doesn't do a good job with bisulphite primer design.

do try methylprimer express it designs way better primers.

Hi Chris,

I understand your frustration. When I first did busulfite PCR more than 10 years ago I had the same problem. But now many reagents have improved, making bisulfite PCR no longer a formidable task. Here are my suggestions:

1. If cost is not a problem, use a kit for the bisulfite treatment. When I first did bisulfite treatment there was no kit available, but I had to use DNA wizard kit from promega to purify my DNA. I ended up with little DNA recovery. Then there came the first kit from Chemicon (yes it was called Chemicon, then bought by upstate, then by ...). The kit helped a little but not that much till a kit from Qiagen which really boosts my success rate. I have also tried a kit from Zymo which works great too. These kits comes with column based purification system which give very good recovery of DNA -- critical for bisulfite PCR.

2. PCR conditions: After the first 40 cycles, re-amplify the product using the same primers or nested primers. 45C is kind of too low. keep at 55C. If you are not using a very old PCR machine, only spend 5-10" at 95C for every cycle. the extension time is also way too long, using a max. 30". My cycle conditions look like this:

Assemble a 20 ul reaction (for the first round of PCR, you don't need to run a 50 ul reaction so that you can save your precious DNA)
94C 2 min
94C 10" - 55C 20" -- 72C 20" x 40
72C 5 min.

Then re-amplify the first product using the same condition for ~30 cycle in a 30 ul reaction to obtain enough product for sequencing or cloning.

3. Get the JumpStart RedTaq polymerase from Sigma. It helps a lot! No additional additives are required.

4. I don't think primer design is an issue in your case because you have also tried primers known to work.

I hope that helps. Good luck!

Hi Chris!

There are some points which might be a problem.

First of all, I don't think your problems are caused by the bisulfite-treatment. It's an older but well known protocol. What seems odd is your OD 260/280. For bisulfite-converted DNA I would expect a ratio between 1.9 and 2.1!

Another point: Do you dry your DNA pelet after ethanol-precipitation? I found that even smallest traces of EtOH might work as a strong PCR-inhibitor. Also try less DNA. 10 ng should be enough for a standard PCR. Too much DFNA could also inhibit your reaction.

Regarding your primer, there are sometimes high differences in melting temperatures. I never trust programs like MethPrimer, because I never (!) received good PCR results from this primers. In my opinion these programs are good to get an idea in which region the primers should be designed. But final designing is done manually. I use Primer Premier 5 (also available as NetPrimer online) for checking dimers, cross-dimers, hairpins and melting temp. Try to design your primer with nearly identical length and (if possible) GC-content. Avoid G-stretches and Gs at the 3'-end. I try to design primers with melting temps in the range of 52 degC and run the PCR with an annealing temp of 56-60 degC. I never tried the Invitrogen Taq (I use Qiagen HotStart Taq), but 5 minutes activations seems to be very short.

Hope that helps...

MoB

Thank you guys for your suggestions. I will do all three suggestions:

MethylNick: I'll use Primer Methyl Express and get some new primers. When I was manually designing primers I would just find any stretch of 20-30 bp immediately flanking the CpG island that didn't have any CGs in it and use that to make primers off of. As long as it didn't have a long stretch of Gs or Ts I would pretty much accept and order it. I admit this is not an ideal way and I hope Methyl Primer Express will fix it.

pcrman: I'll try out your new cycle conditions as you've described along with the JumpStart RedTaq polymerase. Although I'm surprised Platinum Taq doesn't work given that it is also hot start and lacks proofreading activity. I'll try ordering a kit as you suggested going with the Qiagen one since their kits seem to be very reliable in my experience (though hugely overpriced).

MoB: Yea, I was also concerned with my 260/280 ratios as well. However, I've had ratios in this range before for regular PCRs and they've worked fine, but yes I admit BSP is somewhat finicky and probably requires cleaner conditions. I'm hesitant to clean-up since I have so little left. As for drying my pellet, yes I did. I gave it several minutes to air dry before resuspension. I usually do wash with 70% EtOH however I was using a protocol that said it wasn't necessary so I just followed it but now I'm wondering...

I'm praying that it's just the PCR conditions perhaps along with the incorrect Taq. I read a previous post where the problem was just fixed by switching Taqs. I'll try pcrman's suggestions today so long as I can find the new Taq. Thank you so much for your help and if you can think of anything else please post!

Chris

Hey Guys,

Thanks to all who replied to my post. I finally got those damn bisulfite PCRs to work! I think my problems arose from a combination of poor PCR cycling conditions and poor primer design. The latter is especially so because all the primers pairs that work now were designed with MethPrimer and all the ones that have failed were manually designed. I knew bisulfite PCR primer design was less forgiving than regular PCR primer design but apparently I underestimated just how much. I'm also using JumpStart Taq (Sigma) instead of Invitrogen Platinum Taq though I imagine there shouldn't be much a difference between the two -- they're both essentially the same Hot Start Taq.

One more question I do have is what is the current gold standard to ensure that you have near 100% conversion? Right now I'm employing three checks:

1. A single regular "test PCR" using regular PCR primers on bisulfite treated DNA to see if I can amplify any residual unconverted DNA. If that is negative then it supports complete conversion. The main problem with this method is probing a single locus in the genome and extrapolating that finding to the entire genome can be dangerous.

2. Once I start getting my sequencing data (direct and clonal) I have internal control non-CpG cytosines within the sequence that I will check. If all of them are converted to "T"s then this should be pretty telling as to whether near-full conversion occurred though I suspect from the outset the data is not going to be this clean. Assuming it is, this check still suffers from the fact that it is locus-specific. Other loci with higher GC content may not achieve conversion to the degree this one does assuming this one works.

3. I'm also using published primer sequences against human MLH1 and they should only amplify bisulfite-treated DNA but not untreated DNA.

This may sound like overkill but I don't want to be nailed by colleagues (and future reviewers) as to the integrity of my bisulfite PCR sequencing data. Comments about my checks would be appreciated but more importantly I'd like to know what controls you all employ to ensure (near) complete conversion of your DNA.

Thanks again!

Chris

Hello everybody, PLZ help me,

I was working on CpG methylation stuff, and from your guides, especially MethylNick, I finally could get very nice bands, my advisor and me became sooo happy and and I introduced your site to him too, we both thought that the most part of the work has been done, but... .but I can not get any sequencing result from PCR product,,, I like to tell a little more about my procedure;

I extracted PCR products by gel extraction kit (qiagen) and directly I took 9μl of that and 3.2μl of forward primer in one 0.5 ml tube and send to sequence, but just untreated samples(without any bisulfite treatment) had results. I thought maybe there is one agent in my samples whihc does not let sequencing does well job, so I washed my samples(before DNA is treated with bisulfite with zymo kit, then amplified by PCR, and PCR product was extracted by kit) then I put 9μl of that in to a 0.5 ml tube and I add 3.2μl of reverse primer instead of forward primer and sent for sequencing again, the same result, I mean no any sequencing result for bisulfite treated samples.

So, PLZ let me know what you think? it is important that my PCR products are very short, <200bp so if you think I have to amplify my PCR products before sending to sequence?

I'm disrupted too much, my great friends, PLZ give me your ideas,
Thanks