different copy # multimers with linker ligation? - (Mar/23/2010 )

Hey everyone. I need to incorporate between a # of copies of a 20 bp fragment into a single restriction site on a vector. Hopefully I can find a reliable way to produce 1, 3, 5 and 7 copies of this fragment.

There is one problem: The 20 bp sequence is semi-palindromic, which means that a forward oligo of this sequence would be able to bind to a copy of itself with very high binding energy, but with 2 bp mismatches -- which would make a traditional linker ligation very difficult.

Fortunately, another research group has already accomplished precisely what I want to do (even inserting it into the same restriction site!), but their paper is a 'letter' and the Method is not detailed enough for me to follow successfully.

Essentially, they put the 20 bp sequence between several bases on 2 oligos such that the oligos can anneal and form sticky ends that can ligate together, but only in one direction. The binding energy of the 2 oligos binding properly is much higher than that of other dimers that can form, though they are still relatively high energy (~-45 kcal/mol for the desired annealing, ~-15 kcal/mol or less for the others).

They then ligate the annealed oligos, use the Klenow fragment to fill in the 5' overhang, digest their recipient vector with XhoI and also fill the 5' overhang with Klenow fragment, and then they do a blunt-ended ligation.

So I ordered the exact same oligos (desalted, not 5 or 3' modified). To anneal them, I diluted to 100 uM in water, mixed equal volumes, brought the solution to 92C for 10 minutes, then lowered the temperature 0.1C/s to 4C.

Since I did not order 5' phosphorylated oligos, I then used NEB's T4 polynucleotide kinase (PNK) to phosphorylate the 5' ends. I followed their protocol of ~300 pmol of ends per 50 uL reaction and 10 units of enzyme (1 uL).

After this, I heat-inactivated the PNK at 65C and then added T4 DNA ligase (which is compatible with PNK's buffer). Not knowing what would work best, I put one sample at 16C overnight and another at 16C for 2 hours.

When I ran these ligations on a gel, my hope was that I would be able to see distinct bands for each multimer that can form. I ran a 1.5% gel for 35 minutes to increase the separation. However, all I saw was a very long and bright band from ~300bp all the way down the gel. Maybe I don't want to run this on a gel at all, maybe I want to purify it with a PCR purification kit and do a ligation with it directly -- I just don't know.

Now, I THINK my problem is that I should use PAGE-purified, 5' phosphorylated oligos, but I want a second opinion before I order these expensive oligos. However, I really am not sure what the problem is, nor am I sure what the correct protocol is, since the Method I am basing on this sucks. If anyone knows of a protocol that would work in this situation that would be ideal, but otherwise I am open to ideas.

The long and bright band may be due to the ligase, this is a problem I always have if I don't heat kill the enzyme. If that doesn't put it to your liking you can also like to add SDS to a final conc of 1% before loading.

In my experience, this result indicates that there is a variety of different length products. I suspect this is because I am only using desalted-grade oligos, which means that some significant fraction is not full length and so lots of different primer-dimers can form other than the one I desire, and all sorts of different length products will form during ligation. At least that's what I'm thinking is happening -- I would just like to find a confirmation before I spend $150+ on *2* oligos.

But now that you mention it, this would be the first time I've ever ran a ligation on a gel without purification (as my ligations are typically 10 uL and too low in concentration to see), so I will look into the effect of T4 DNA ligase on gel visualization to see if that could be it.