Is ligation really necessary? - (Apr/05/2013 )

I came across a new method of cloning that involves generating 15-17 bp complementary overhangs on PCR products (http://www.biomedcentral.com/1472-6750/11/92). In this method, the insert and the vector are PCR'd, treated with Dpn1 to remove the methylated template, and the pieces are directly transformed. The authors weren't sure of the method of exposing the stick ends, but suggested that the 3' exonuclease activity of high fidelity DNA polymerases expose these sites. If this worked well, then do we really need to ligate after a restriction digest?

Long matching overhangs can be directly transformed. Addition of PEG as a crowding agent is usually positive in making this work more efficiently. This does not work for overhangs of 4 bases.

Always helpful. phage, thanks

Would you recommend PEG for restriction ligations or is this usually included in buffers ( I used NEB T4, btw)?

PEG is the additon in so-called "quick ligase" buffer. It is very important in blunt end ligation, but tends to make more problems than it solves in cohesive end ligation (in my opinion). Also, reactions containing PEG cannot be heat-killed without dramatic loss of efficiency. I always do cohesive end, and avoid the quick ligase buffers.

One final question. My ligation efficiency is really low for a simple single digest of the vector and the insert. The insert is restriction digested out of a TOPO TA vector using the respective enzyme so I know it's cleaved. Unfortunately, there is no death cassette in the vector, so my plates are theoretically a mix of ligated and unligated constructs. I used t4 ligase (NEB) for 10 minutes at room temp (per their instructions) at a 1:3 vector to insert ratio. I've screened about 100 colonies so far and I haven't gotten a single insert. I know my procedure for analysing the presence of an insert works because I've successfully found one colony that did in a another reaction. Would letting the ligation go overnight help?

Probably not. Are you using a single enzyme or two different ones, for directional control?
You might consider making vector with PCR rather than by cutting it. The major difficulty with cutting vector is that some of it fails to cut. You can reduce the problem by linearizing with PCR instead cutting with a RE. You can remove most of the template DNA with a DpnI digest. You can test for background with a vector only ligation control.

So I have two seperate inserts, going into the same destination vector (I'm trying to make 2 different constructs). Inserts are cut from TOPO with EcoR1 and destination vector was cut with EcoRI. Believe me, I know there are better ways to do this but this is my last resort (time issue/rotation student). During my first screen, I found 1 of 10 colonies that contained my insert (PCR verified) for both constructs. I used restriction enzyme analysis to determine orientation. The one insert was in correctly, the other was flipped. So now I've just been screening for another good copy of the one that was flipped. I must have screened 100+ colonies by now and nothing is coming up. Mind you, I probably have 1000 colonies on my plate, but yeah this is getting ridiculous. Either way, I re-ligated for 24 hrs at 16c and retransformed. There's no reason this shouldn't work.

I'm using colony PCR now for speed, but I am using primers that flank the cloning site. So a hit should pop up. When I first found inserts, I grew up cultures and used 1 uL of a turbid culture for the PCR. I don't think there should be much difference between these 2 methods with reference to the results.

PCRing my vector wouldn't be easy, I don't think. It's 6kb long and we don't have any of those fancy lambda phage polymerases.