Success with Direct Sequening of Bisulfite Pcr Products - (Mar/24/2006 )

Hi All,

I've been working on developing a system to check the methylation status of 85 CpG islands across a 700 bp region (1152 bp is actually sequence used for calculation of primers). It has been a bit of a headache getting things to work right and in some cases I ended up just being lucky.

So here is what I wanted to share, good and bad, dumb and even dumber at times .

First attempted to design primers, using manual conversion of Sequence to reflect Bisulfite conversion. I then used primer3 and went from there. First plan was to do 3 pcr reactions of around 300bp to cover the 700 bp region with some degree of overlap for primer binding locations.
Spent a good month, with bad sequencing, boss not happy, bad pcr products .

Primer set A (1-355) Fair Product, Fair Seq
Primer set B (300-600) Good Product, Good seq
Primer set C (550-750) Fair Prodcut, No seq
Primer set D (600-850) Multiple Band Product No seq
Primer set E (625-875) Multiple Band Product No seq

One afternoon after reading this forum on Nested pcr I went back to my PCR primers and found a few primer pairs that would yield larger pcr Products 700+ bp

Primer set A 5' + Primer Set D 3'
Primer set B 5' + Primer Set C 3'
Primer set A 5' + Primer Set E 3'

I then checked on primers locking the 5' Primer and searching for a 3' primer. Tested all my current primers. Dumb luck kicked in

Primer set A 5' + Primer Set B 3' 450 bp
Primer set B 5' + Primer set C 3' 545 bp

So here is actually what I do.
Round 1
PCR Bisulfite treated DNA with Large primer set (A5+D3), (A5+E3) With pcr Tm 1 degree below calc Tm, ext temp 72 , 6 min final 72 extension , Sigma Jumpstart ReadyMix Taq, 2x 3' primer
Round 2
Using 2 ul of product from Round 1 , 2x 3' primer PCR with primer set (A5+B3, B5+C3), Tm 1 Degree below calc Tm , ext temp 68 , 6 min final 68 extension, same taq.

Gel extract , Purify , Spec and submit for sequencing.
Here is where this is a bit differant for sequencing,

I use the following primers for sequencing,
Product (A5+B3) sequencing primers (A3, B5)
Product (B5+C3) sequencing primers (B3, C5)

Visual example
Pcr Product (A5+B3)
[A5-------------->-----------------------------------------------------------------<----------------B3]
Seq primers underlined
[----------------------------------------------{B5------->--<----------A3}---------------------------]
90% of the time the 5' seq primer has enough coverage to reach the end of the product.

Depending on the primer sets I do however have around a 15bp gap I simply can't sequencing results in, no matter how hard I try or with which set I try.

While this method is somewhat sloppy, and I learned alot from my mistakes. I am now getting good data. Given the budget of my project, Time, and other factors, It worked for me and I might work for someone else. So I thought I would share this with all of you.
as an added bonus the product from round one can yield up to 25 pcr reactions for round 2 , minimizing my need to do bisulifite conversion.
Now I score the CpG methylation by measuring height peak setting 100% to be equal height peaks and score on a scale of 0-4 I try to average the back ground C signal and subtract from peak height for calculation. (Im still debating on if this is a correct method to use)

0 = No Methylation signal
1= <25%
2=25-50%
3=50-75%
4=>75%
666872 peak 44 = 4, 49 =3
666873 peak 48= 2 (barely)

Open for discussion , feed back or better hints.
I'm still refining the system to yield better results.
The future plans include 5-aza treatment (1.0 uM appears to yield 20% death at 48hrs with 2.5 uM killing around 80% at 48hrs ) trypan blue and MTS assay.

Best of luck
AD5

great to hear that direct sequencing is working for you AD5!

I was wondering, both B5 and A3 primers worked for you in the A5+B3 pcr product? I have found that the reverse primer works and the forward never works and in your case it would mean A3 would work and B5 would not. However your method of sequencing is to use a different primer internal to your PCR product and I suppose this alleviates the problem. As for peak height scoring you method is a reasonable one and has been used in many publications. Does finchTV allow you to measure peak height by placing the souse over the peak? if not, might I suggest 4Peaks, it is a mac program, I am not too sure if you are using mac or PC, but I know 4peaks allows you to do this on screen!

have fun looking at DNA methylation!
Nick

The bisulfite conversion process creates non-complementary strands of DNA, because the non-methylated C's are converted to T's, while the G's on the complementary strand stay unchanged.

Unless the subsequent amplification process is strand selective, this results in a mixture of PCR products. When sequenced, these do not give a clear picture of the methylation status of the C's, instead giving reduced height mixtures of C and T, and similar reduced height mixtures of G and A.

The solution to this is to use primers for the post-conversion PCR which are selective for converted bases. Specifically, you can design primers which are selective for the forward strand, and a distinct set of primers selective for the reverse strand. This only works if you have some knowledge of the location of methylated bases. Ideally, you would want a converted base at the 3' end of each primer, and perhaps a few more in the middle. Designing these requires careful thought.

The non-strand-selective PCR primers can be used to amplify converted DNA, but to get accurate methylation status, the PCR product has to be cloned, which will select a single product molecule and amplify it sufficiently for subsequent sequencing, avoiding the problem of sequencing a mixed set of products. In principle, half of these clones will show the forward strand converted, and half the reverse strand converted.

You should be aware that PCR reactions for converted DNA have a template with unusually low GC content. Many times, this requires a PCR *extension* (not annealing!) temperature which is substantially lower than the normal PCR temperatures of 72C. I use 64-66 C for these and lengthen the extension time.

well put phage434, serious thought into primer design must be made, however I have to say with all bisulfite PCR, the primers are selected to be strand specific....ALWAYS or you won't get a PCR amplicon simply because the strands become non-compliementary.

So I have to say some of yor comments are incorrect with regard to non-strand specific primers because if they are not specific, then the PCR amplification process simply will not occur.

Nick

Thanks for all the comments.
"I was wondering, both B5 and A3 primers worked for you in the A5+B3 pcr product? I have found that the reverse primer works and the forward never works and in your case it would mean A3 would work and B5 would not."

The b5 primer works on the A5+B3 Product , But the b5 primer never worked for the B5+B3 Product, Now by working i mean I usually get around 150 bp sequenced before noise kicks in, alot of noise.

"You should be aware that PCR reactions for converted DNA have a template with unusually low GC content. Many times, this requires a PCR *extension* (not annealing!) temperature which is substantially lower than the normal PCR temperatures of 72C. I use 64-66 C for these and lengthen the extension time."

I tested each pcr reaction with the following extension temps 72 70 68 66 64 62 I found that 68 for the 2nd round of product seemed to yield the best product in terms of sequencing quaility. I will how ever review my data sets and double check it over.

If I remember right there is a poly t region and a 35 bp region that shares alot of homology with other DNA, that region interferes with a couple of my primers, (1st round primers) I found 64 did yield some product but experiment size on gel was 900bp when expected was 740bp, I extracted and sequenced, while the product did contain my sequence on the 3' end I noticed a poly T studder that then picked up sequence that was about a chromosome off.
It then moved a bit 5' of that region and retested everything with a new primer set. Expect size returned to experimental size, and sequence was right on.

Thanks for the feedback

Both direction primers work in sequencing my pcr reactions. I get four sequences, one forward and backward for the top strand, and one forward and backward for the bottom strand.

that is very interesting to hear this phage434.

in my hands, it is always the case, if direct sequencing actually works, the reverse primer seems to work and the forward never does, I put this down to the skewing of of bases depending on which strand you are sequencing with the primer, the forward would be rich in T, G and A, the reverse would be rich in T, C and A and I think it is because of this the reverse primer seems to work better. But that is from my observation, there is no scientific basis for this, all anecdotal.

Nick

Those are interesting findings. I just wanted to also share my first experience with direct sequencing. I just finished sequencing DNA from a single tumor sample using a single pair of primers. I had a lot of DNA so I did 5 different bisulfite reactions to see if it mattered how much DNA i used:

1) 2ug of DNA
2) 1ug
3) 0.5ug
4) 0.25ug
5) 25ng

I got clean data for direct sequencing in both the forward and reverse directions. The data also came out identical except for some evidence of methylation in the reverse direction only in my 0.5ug and 25ng samples. I thought this was strange, but i think it just might be a quirk since I didn't see in the forward directions for all 5 of the samples. In the end, I guess for this particular genomic region, I can get away with using a very small sample of DNA.

And just FYI, I used an extension time of 1 minute at 72 C degrees for each cycle and ending for 10 minutes for amplifying my BSP products. I also used the same forward and reverse primers from the BS-PCR reaction to also do the sequencing.

Phage434, did you mean 68C was the best temp for extension for the "second round" during the sequencing? or during the BS-PCR amplification?
Thanks

that is really interesting purplefetus, thank you for sharing this with everyone.

Can you please tell us, how you cleaned your PCR amplicons for direct sequencing?

Cheers

Nick