Design of <=40bp primers adding restriction sites and also Gibson Assembly o - (Feb/18/2017 )

I'm interested in design on primers that add restriction sites as well as Gibson Assembly overlaps so that I can assemble multiple fragments by GA – then later easily go back and swap out one fragment later. An issue I've had with GA cloning is the lack of restriction sites between fragments and the need to perform a new reaction with all products if one fragment needs to be changed.

Normally 40bp primers for GA have ~20bp homology with the amplicon and another ~20bp overlapping with the fragment being cloned into an adjacent position. I'd like to squeeze in a 6bp cloning site in the middle without significantly reducing the efficiency of PCR or GA. I want to keep primer length below 40bp because after that the cost increases.

How much would you recommend I shorten either the amplifying or overlapping part of the primer to make room for a cloning site in the middle?

In theory you should have enough homology with very short sequences given that these DNA should be the only ones in the solution. However, I would not shorten either by any more than 5 bp, as this will lower the binding efficiency/energy, which is what you need to consider here rather than sequence similarity.