AbstractRestriction enzymes are the workhorses of molecular biology. We introduce a new problem which arises in the course of our project to design virus variants to serve as potential vaccines: we wish to modify virus-length genomes to introduce large numbers of unique restriction enzyme recognition sites while preserving wild-type function by substitution of synonymous codons. We show that the resulting problem is NP-Complete, give an exponential-time algorithm, as well as well-performing heuristics, and give excellent results for five sample viral genomes. Our resulting modified genomes have several times more unique restriction sites and reduce the maximum gap between adjacent sites by three to nine-fold.