We made full length """"""""infectious"""""""" cDNA clones from a virulent dengue type 1 virus (DEN1 WP) and from a live attenuated DEN1 vaccine candidate, which had been adapted to grow in dog kidney cells (DEN1 PDK20). First, the left and right ends of DEN1 WP were amplified by RT-PCR using virion RNA as a template and synthetic oligonucleotide primers. The primers introduce an SP6 RNA polymerase promoter just upstream of the left-hand end. The cDNA products (approximately 3kb at the left end and 4kb at the right end) were cloned in proper orientation in the polylinker of the plasmid pRS424. Then, a 5kb cDNA was made from the middle of DEN1 WP by RT-PCR; this fragment overlaps both the left and right cloned ends. Homologous recombination between the linearized clone and the 5kb cDNA upon cotransfection of yeast creates a plasmid containing the full-length DEN1 WP cDNA. Plasmid DNA was prepared from E coli, and transcribed in vitro. Electroporation of RNA into LLCMK2 cells produced a dengue infection. Recovered virus was passaged on C6/36 cells. This virus behaved like the parent DEN1 WP in growth curves. Comparison of the nucleotide sequence of the cloned DEN1 WP cDNA to the sequence of the WP virion RNA (determined by sequencing uncloned RT-PCR products) revealed only 5 differences: 3 silent and 2 missense mutations. The DEN1 PDK20 infectious clone was made by replacing the WP sequences in the full-length clone with cDNA made from PDK-20 virion RNA. First, an RT-PCR product representing the right half of PDK-20 was used to replace the corresponding region of WP by homologous recombination in yeast. This chimeric plasmid was grown in E. coli, and transcripts from it were proven infectious. Then, the chimeric plasmid was used as a starting point to introduce the left half of PDK20 as above, and a PDK20 full-length clone which made infectious transcripts was identified. Comparison of the nucleotide sequence of the cloned DEN1 PDK20 cDNA with that of PDK20 virion RNA (determined by sequencing uncloned RT-PCR products) revealed 8 differences, including 4 coding changes. These coding changes were repaired in a further series of homologous recombination steps in yeast, resulting in a final infectious DEN1 PDK20 clone which has only 3 silent changes from its parent. Transcripts were electroporated into Vero cells to make virus, which was passaged on C6/36 cells. This virus has the same small plaque phenotype as its DEN1 PDK20 parent. Growth of these viruses in monkey and dog kidney cells is currently being investigated. In the future, we plan to map the mutations responsible for the adaptation of PDK20 to growth in dog cells, by making chimeras between the WP and PDK20 infectious clones. Also, in human clinical trials the PDK20 virus is slightly too reactogenic, yet a PDK26 virus is overattenuated. We plan to introduce further attenuating mutations into the PDK20 clone in an effort to recover a correctly attenuated vaccine strain, and at first we will use subsets of the mutations found in PDK26 which are not in PDK20.
