How exactly are you doing the cloning after you amplify the fragment?

Also, what do you mean by "deletion where the ligation occurs?" Is the deleted part in the backbone vector, the insert, are the deletions identical between clones or are they all different, etc?

I use a hifi polimerase. This generates blunt ends. I just simply perform a ligation and subsequent transformation. The tails when included properly, should just touch and then reconstituing my sequence, but i find that at the end of one tail a number of bases are missing. The deleted part is not in the backbone, it is part of the tail a I want to introduce. The deletions are different among clones, one clone has a 6-pb deletion, other 13 pb. Always where the primers should "touch"

Thanks for your interest and time

Regards

Oh okay, so you're doing inverse PCR on the whole vector then right? Sorry I misunderstood exactly what you were doing.

My first three guesses:

1. Your primers are simply bad. They could be poor quality containing lots of incomplete oligo synthesis products. Oligo synthesis is usually done in the 3'-5' direction and the closer you get to the 5' end typically the more failed products occur since nucleotides are added step-by-step. This isn't really a problem with shorter primers but it can really add up with longer ones.

In this case, you could just reorder them with additional size purification to make sure most of what you're getting is intact primer.

2. The primers don't work well and may be doing a lot of spontaneous dimerization with each other or with themselves, resulting in weird products.

This means you would have to probably mess with the annealing temps, put in additives like DMSO. In the worst case, you would probably just have to change your cloning method since you don't have a lot of options in terms of changing primer sequences.

3. Your proof reading polymerase's exonuclease activity is somewhat active at lower temperatures and it's chewing away bits of your oligos before you even get started. Single stranded DNA is typically considered an "error" by the polymerase as far as proof reading is concerned.

This would be an easy fix, you simply heat your reaction mix (without polymerase) up to your melting strand separation temp (98°C for Phusion) add the polymerase, then allow the reaction to proceed. This way it can't interact with unannealed primers as much and chew them up.

Something that I think is possible but maybe less likely:

*Your primers aren't 5' phosphorylated to allow ligation so the only way the vector can circularize is weird non-specific interactions at the ends.

In that case you'd just treat them with T4 PNK before using them in your PCR so that T4 ligase has a phosphorylated end to work with.

Wow!!! thanks a lot for your super detailed description. Yes, what I am doing is an inverse PCR, (sorry, this information would have clarified the things much better). I am tempted to say that the explanation 1 is the most likely (they offered me an HPLC purification step and I thought that this could be no necessary). One of the primers was fully integrated in the sequence and the other is the one that always appear as incomplete. Step three looks also possible.

Again, thanks a lot for your help!

regards