Paramyxovirus P genes are transcribed into two mRNAs which differ from each other by either one (measles and Sendai virus) or two (SV5 and mumps virus) G insertions, and which code for either the P or V proteins. The G insertions always occur within a short run of Gs, and a stuttering mechanism for the insertions has been suggested in which the viral polymerase reiteratively copies a template C residue during mRNA synthesis. Support for this mechanism was obtained by varying the reaction conditions during Sendai virus mRNA synthesis in vitro. A stuttering model is proposed which accounts for how the ratio of inserted to uninserted mRNAs is controlled, and why some paramyxoviruses insert one G and others two Gs when insertions occur.

Two forms of the Sendai virus P/C mRNA have been predicted: one an exact copy of the viral genome, and the other with a single G insertion within a run of three G's. We directly cloned the mRNA or portions of it containing the insertion site and screened the resulting colonies with oligonucleotides that could distinguish the presence of three or four G's at this position. We found that 31% of the mRNAs did in fact contain the predicted insertion, whereas the viral genomes contained no heterogeneity at this position. A smaller fraction

RNase mapping was used to estimate the levels of unencapsidated Sendai virus plus-strand RNAs which cross the leader-NP junction relative to NP mRNA. Significant amounts of leader readthrough RNAs were found in Z strain-infected cells, similar to that described for the poiR mutant of vesicular stomatitis virus, even though this strain is considered wild type. The levels of the readthrough RNAs detected fell sharply when progressively longer probes were used, unlike that of NP mRNA. These studies suggest that polymerases which read through

The role of glycosylation and of disulfide bonds in the formnation of the native structure of the Sendai virus hemagglutinin-neuraminidase (HN) and fusion (FO) glycoproteins was studied. In contrast to the HN and Fo synthesized in vivo, the proteins made from pSP6 transcripts in reticulocyte lysates, whether glycosylated or not, were not recognized by monoclonal antibodies or polyclonal rabbit sera raised against the native proteins; they efficiently reacted only with rabbit antisera raised against the reduced sodium dodecyl sulfate-denatured proteins. These in vitro-made proteins, however, did not contain disulfide bonds. The proteins made in vivo in the presence of tunicamycin, which were also not recognized by the anti-native protein antibodies, did contain disulfide bonds, but they were mainly incorrect interchain disulfide bonds. Moreover, while Fo acquired proper disulfide bonds as soon as it was synthesized under normal conditions in vivo, the disulfides were formed in HN only after a lag of 10 to 30 min. This lag coincides with the delay observed in HN native structure formation. We therefore conclude that the maturation of the HN and Fo proteins depends on the formation of proper intramolecular disulfide bonds, which in turn depends on the previous addition of high-mannose sugars.

No abstract