The last seven nucleotides of the 3 ' -end of rotavirus mRNAs, 5 ' -UGUGACC
-3 ', are highly conserved acid form a cis-acting signal that can promote t
he synthesis of (-) strand RNA to produce the viral dsRNA genome in vitro.
Previous studies have shown that the sequence, location, and strandedness (
single- versus double-stranded) of the 3 ' -consensus sequence of the mRNA
affect the efficiency of (-) strand synthesis. In this study. we have used
exhaustive mutagenesis of the SA11 gene 8 mRNA and an in vitro replication
system to define the importance of each of the residues in the consensus se
quence in (-) strand synthesis. The analysis showed that the CC of the cons
ensus sequence was the most critical for (-) strand synthesis. Furthermore,
the data revealed that other, but not all, residues of the consensus seque
nce contributed to efficient(-) strand synthesis in vitro. Mutant gene 8 RN
As supported an intermediate level of (-) strand synthesis when the 15 nt s
equence upstream of the CC was replaced with long tracts of poly(A) or poly
(U), but not with poly(G). Predictions of the secondary structure of the mu
tant RNAs suggested that the poly(G)-RNA could not replicate because its 3
' -terminus was largely basepaired, instead of extending as a single-strand
ed tail as is the case for the 3 ' -termini of the poly(A)- and poly(U)-RNA
s and wild-type gene 8 RNA. Subsequent experiments performed with complemen
tary oligonucleotides indicated that efficient RNA replication occurs in vi
tro only when the last four residues of the 3 ' -consensus sequence, and mo
st importantly the two terminal C's, existed in a single-stranded form. A s
ingle-stranded CC may be crucial for formation of an initiation complex for
(-) strand synthesis consisting of viral RdRP, mRNA, and the dinucleotide
pGpG.