We had a similar problem - adding 10% DMSO to the reactions solved it although the cloned PCR products did contain more base changes
