Ligation-mediated PCR: Vent Polymerase, Why? - (Feb/17/2014 )
Hello everyone.
I'm attempting to amplify a region of unknown sequence, downstream of known sequence, starting from a genomic sample. For instance:
AAATGCATCGA-XXXXXXXXXXXXXXXXXXXXXXX
Ligation-mediated PCR is a technique that utilizes blunt-end ligation of a linker of known sequence to restricted genomic DNA. This "flanks" a chunk of unknown sequence, with two portions of known sequence, allow for PCR amplification and then sequencing of the amplicon. The process is as such:
1) Restrict the genomic DNA.
2) Creat a blunt end within your genomic DNA digest, and thermocycle using a primer specific to your known sequence. This is linear amplification, and utilizes one primer. This is where specificity arises - only your fragment of interest should generate a blunt end, and thus be the only eligible fragment for blunt-end ligation.
3) Ligate the linker to your restricted, now blunt-ended, genomic DNA. The linker is only blunt on one end, preventing linker-linker ligation.
4) Amplification, using two primers - one specific to the linker, one speific to the known sequence. Preferably, the primer to the known sequence will be different than that used in step 2, and closer to the unknown sequece. This increases specificity.
This protocol recommends using Vent polymerase. I'm not sure why this is preffered over a high-fidelity polymerase. Anyone have any insights?
I had a lot of background on my first try at this technique. I'm going to try to increase primer-specifity and annealing temperature. Any other ideas?
Your linker should use (on the non-linking end) a dephosphorylated 5' end, and (on the other strand) a 3' amino to block ligation. This will guarantee that no linker-linker ligation occurs. You could also design a single oligo which would fold to form a hairpin, blocking ligation at that end. What kind of background are you seeing? Neither of these will prevent blunt ligation of two linkers. A possibly better approach is to add a 3' A tail to your fragments, then ligate a hairpin oligo which has a phosphorylated 5' T overhang. This is what the Illumina sequencing adapters do. They use a pair of phosphothionated linkers at both ends, to inhibit chewback of the linker. You can probably leverage some of the end repair and A tailing kits from NEB (for example) designed to produce Illumina sequencing fragments.
phage434 on Mon Feb 17 20:32:47 2014 said:
Your linker should use (on the non-linking end) a dephosphorylated 5' end, and (on the other strand) a 3' amino to block ligation. This will guarantee that no linker-linker ligation occurs. You could also design a single oligo which would fold to form a hairpin, blocking ligation at that end. What kind of background are you seeing? Neither of these will prevent blunt ligation of two linkers. A possibly better approach is to add a 3' A tail to your fragments, then ligate a hairpin oligo which has a phosphorylated 5' T overhang. This is what the Illumina sequencing adapters do. They use a pair of phosphothionated linkers at both ends, to inhibit chewback of the linker. You can probably leverage some of the end repair and A tailing kits from NEB (for example) designed to produce Illumina sequencing fragments.
Hi Phage, thanks for your reply. The linked I'm using is only "blunt" on one end. I created it by ordering two oligos from iDT. The sequences were recommended in a LM-PCR protocol, developed Garrity et al. The linker looks something like this:
5-ATGCTGCGCTGGGCC-3
3-TACGACGC-5
I think you get the concept. I'll look into the Illumina technology. I've been trying to no avail to get an amplified chunk of this unknown region. I started with inverse PCR. Recently I designed a couple of primers that bind to repetitive elements, perhaps that'll work.
So, I assume your real linker has a 5' phosphate
5-<phosphate>ATGCTGCGCTGGGCC-3
3-TACGACGC-5
Otherwise, you won't get ligation. In any case, you will get a large number of ligation dimers with this technique.
Also, with single ended PCR, you end up with ssDNA, not dsDNA. This won't ligate in any case, which is probably the real problem you have. This protocol sounds broken.
I liked your inverse PCR idea quite a bit. Why didn't this work?
This is exactly the problem with making fragments for NGS sequencing. After end repair, they A tail then ligate an adapter with a T overhang.
Check out NEB E6050 and E6053.
Depending on your organism, the problem may lie in getting any amplification. Is it high GC or low GC? If high, add 5% 1M betaine solution. If low, lower the EXTENSION (not the annealing) temperature to 63-65, and extend the duration a bit.
phage434 on Tue Feb 18 23:53:37 2014 said:
So, I assume your real linker has a 5' phosphate
5-<phosphate>ATGCTGCGCTGGGCC-3
3-TACGACGC-5
Otherwise, you won't get ligation. In any case, you will get a large number of ligation dimers with this technique.
Also, with single ended PCR, you end up with ssDNA, not dsDNA. This won't ligate in any case, which is probably the real problem you have. This protocol sounds broken.
I liked your inverse PCR idea quite a bit. Why didn't this work?
This is exactly the problem with making fragments for NGS sequencing. After end repair, they A tail then ligate an adapter with a T overhang.
Check out NEB E6050 and E6053.
Depending on your organism, the problem may lie in getting any amplification. Is it high GC or low GC? If high, add 5% 1M betaine solution. If low, lower the EXTENSION (not the annealing) temperature to 63-65, and extend the duration a bit.
Hi Phage,
I probably did a poor job of explaining the technique. Here is an image for clarification:
I'm not sure how linker-dimer would arise, because only one end of the linker is blunt.
I just contacted iDT - my linker had no 5' phosphate. Thanks for the catch Phage.
My previous attempts at Inverse PCR are recorded within this thread: http://www.protocol-online.org/forums/topic/31194-troubleshooting-inverse-pcr/
I'd be happy to hear any feedback you might have on it.
Thanks for your help Phage!
So, your single primer "pcr" is really only a single cycle, leaving the same number of dsDNA molecules.
Your linkers can link back to back, ligating the blunt ends together.
Your diagram is a bit confusing. It shows extending the bottom strand, which is usually the 3' to 5' direction (which of course won't extend). It also shows that strand ligating to the short linker fragment, which is a 3' end. I think you meant this to be an extension of the top strand, and then a ligation to the top linker strand. Is that right?
phage434 on Wed Feb 19 17:02:33 2014 said:
So, your single primer "pcr" is really only a single cycle, leaving the same number of dsDNA molecules.
Your linkers can link back to back, ligating the blunt ends together.
Your diagram is a bit confusing. It shows extending the bottom strand, which is usually the 3' to 5' direction (which of course won't extend). It also shows that strand ligating to the short linker fragment, which is a 3' end. I think you meant this to be an extension of the top strand, and then a ligation to the top linker strand. Is that right?
Hi Phage.
Here is the actual sequence of the linker:
5'-GAATTCAGATC-3'
3'-CTTAAGTCTAGAGGGCCCAGTGGCG-5'
I believe I see what you mean by the dimer:
5'-GCGGTGACCCGGGAGATCTGAATTC-3' 5'-GAATTCAGATC-3'
3'-CTAGACTTAAG-5' 3'-CTTAAGTCTAGAGGGCCCAGTGGCG-5'
Correct? This process would end after one ligation. Adding linker in excess should compensate for this?
That image was the best I could find at the time - here's the actual image from the text. I'm following this almost exactly, except I am not labeling with "primer 3". Instead, if I can get successful amplification, I'm sending it away for sequencing using a third primer.
To answer your question about the other image, I believe the top strand must be 3' -> 5'. I just pulled the image off google - it's not the exact process I'm following, but rather a simplified version of LM-PCR.
Hi,
I am doing something similar. I wanted to ask about the ligation efficiency if the linker looks like:
0'-GAATTCAGATC-3'
3'-TCTTAAGTCTAGAGGGCCCAGTGGCG-5'
Basically, there is a T overhang and the target DNA has an A overhang. Secondly, the linker is 5' dephosphorylated on the ligating end (0').