Abstract
Studies with a murine cytomegalovirus mutant tsm5 suggested two possible approaches to producing a live attenuated human cytomegalovirus vaccine. One approach would be to use a combination of five to six mutants where an attenuating mutation in the gene of one mutant is compensated by the wild-type version in a second mutant, which in turn has a mutation in a different gene compensated by the wild-type version in a third mutant, etc. Important genes in this approach could include those involved in DNA replication. The importance of the carboxy terminase of the primase gene (M70/UL70) for its function suggested a second approach where some of the natural codons in this region could be substituted with synonymous non-preferred (minor) codons that would reduce the replication fitness of the mutant.