Live, attenuated vaccines are best constructed by introducing stable genetic alterations using techniques of molecular biology. Here, dengue virus is our model system. Flavivirus genomic RNAs contain a conserved stem-loop structure within the 3'-noncoding region (3'-SL). We showed, by constructing a series of mutant viruses containing D2/West Nile [WN] chimeric 3~-SL nt sequences, that an 11-bp double-stranded segment of the D2 3~-SL was absolutely required for D2 replication. One mutant, containing a substitution of the 3~-terminal 7 bp of the D2 sequence by analogous WN nts, was severely restricted for growth in mosquito cells but replicated like wt in monkey kidney cells. This is a desirable property of a live, attenuated vaccine virus. (L Zeng, B Falgout, L Markoff  J Virol 72:7510-22). The mosquito cell growth-restricted mutant (mutF) was additionally characterized: (i) Nt sequence analysis showed that it had mutated during growth in monkey cells; an A residue was deleted in position 3 from the 3'-terminus of the genome, within the 7-bp substituted segment. (ii) The wt parent virus, DEN2 NGC strain, had been mouse-brain adapted prior to the derivation of the infectious cDNA genomic copy. Using neonatal outbred mice, we showed that the cDNA-derived wt parent virus was as neurovirulent as its ~parent~ mouse-brain adapted D2 NGC isolate and that mutF virus retained the mouse-brain adapted phenotype of the wt viruses. (iii) Replication of D2mutF virus in adult Aedes sp. mosquitoes was assessed, compared to wt cDNA-derived D2 NGC virus. In albopictus mosquitoes, mutF replication was severely restricted out to day 7 after trans-thoracic infection (>100,000-fold); in aeqypti mosquitoes, mutF virus titers were about 40-fold reduced compared to wt up to day 4 p.i., when titers of the mutant virus began to approach that of wt. These studies confirmed the potential usefulness of mutF in the context of live virus vaccine production, as a means to enhance vaccine safety and possibly as an attenuating mutation in itself. (iv) The original study defined other mutants that replicated in both monkey and mosquito cells in vitro. The 3~-termini of these """"""""viable"""""""" mutant virus genomes after replication in monkey cells were also sequenced [by BF and LM]. Two of them were shown to have undergone a point mutation at nt position 5 from the 3'-terminus. The significance of this finding is being investigated. (v) The genome of mutF virus isolated from monkey cells was further sequenced upstream from the 3~-SL, to the 5~-end of the NS5 (RDRP) gene, to detect additional spontaneous mutations related to replication competence. No difference from the wt parent was detected. (vi) MutF was re-created in the context of an infectious D1 full-length cDNA [provided by BF and SP]. This mutant was viable in monkey kidney cells. Its replication in mosquito cells is currently being assessed, as a proof of the concept that mutF will confer the same phenotype in the context of D1, D3, and D4 genomes as it does in the context of the D2 genome.