In nucleotide sequencing of the cDNA of the influenza virus PB2 polyme
rase gene by the dideoxy method using a modified T7 DNA polymerase, Se
quenase, the sequence of the promoter region, 5'-AGCGAAAGCAGG, was sho
wn to be misread as 5'-AGCGAAACGAGG, i.e., a GC doublet at positions 8
and 9 was read in reverse. This misreading was also found both when t
he sequence of BsmI restriction site upstream from the PB2 promoter se
quence was exchanged by that of the promoter of T7 RNA polymerase and
when the downstream region was substituted with the nonstructural (NS)
protein gene. These results indicated that the misreading by Sequenas
e was attributed specifically to the PB2 promoter region, independent
of the upstream and downstream sequences. The misreading, however, did
not occur when dGTP in the labeling mixture was substituted with anot
her nucleotide analog, dITP. Furthermore, the reversion did not occur
in the NS gene promoter region, where the nucleotide sequence was 5'-A
GCAAAAGCAGG. Since the nucleotide difference between the PB2 and NS pr
omoter regions was only at the fourth residue, i.e., G for PB2 and A f
or NS, the G residue followed by a triplet AAA in the PB2 promoter reg
ion was suggested to be a signal responsible for the misreading by Seq
uenase T7 DNA polymerase. The finding warns of possible misreading in
determining DNA sequences, in addition to compression of the sequencin
g ladder.