MODELING VIRAL EVOLUTION IN-VITRO USING EXO(-) KLENOW POLYMERASE - CONTINUOUS SELECTION OF STRAND DISPLACEMENT AMPLIFIED DNA THAT BINDS AN OLIGODEOXYNUCLEOTIDE TO FORM A TRIPLE-HELIX

Evolution comprises cycles of amplification, mutagenesis and selection . To study evolutionary phenomena, isothermal strand displacement ampl ification (SDA) of double-stranded DNA as an in vitro model for rollin g-circle replication of viruses has been coupled to a positive selecti on procedure. First, two subsequent amplification reactions utilizing exo(-) Klenow polymerase were performed under direct observation using the fluorescent dye thiazole orange. Under the chosen conditions, the mutation rate was 1.5 x 10(-3) and 0.4 x 10(-3) for base substitution s and deletions, respectively Then, a 16mer oligodeoxynucleotide with an acridine moiety coupled to its 5' end was used to select for double strands that retained their ability to form a triple-helix with the o ligodeoxynucleotide. Conditions for triple-helix formation were chosen such that only 10 to 40% of the SDA products were allowed to bind the third strand. Non-denaturing polyacrylamide gel electrophoresis was u sed to-separate triple-helices from unmodified double strands, and onl y tripler strands were used to initiate a new round of error-prone amp lification and selection. Nine such rounds with about 270 molecular ge nerations were performed. The final mutant spectrum was characterized and compared with those of amplification reactions without additional selection pressure. While without selection pressure base substitution s and deletions throughout the initial wild-type rapidly produce a div erse mutant distribution, the consensus after nine selection rounds cl early shows two mutational hotspot positions. Using gel shift assays a nd a newly developed non-radioactive DNase I footprinting technique, i t could be shown that both the initial wild-type and the final consens us do not differ significantly in their tripler formation ability As o pposed to this, they do show different amplification efficiencies. The final consensus sequence is amplified with the highest rate in the ex ponential reaction phase, while the most abundant clone, which is char acterized by two additional point deletions, is the sequence with the highest amplification rate in the linear growth phase. (C) 1995 Academ ic Press Limited