The vast majority of eukaryotic mRNAs are monocistronic and express a single open reading frame (ORF) which initiates from the ATG codon nearest the capped 5Ј end. This initiation site is chosen by a scanning mechanism in which initiation factors, 40S ribosomal subunits, and initiator Met-tRNA bind to the capped 5Ј end of the mRNA and linearly scan the nucleotide sequence for the first start codon (the 5Ј scanning model; for a review, see reference 29). Some eukaryotic viral mRNAs, like the Sendai virus (SeV) P/C mRNA, however, express two ORFs which overlap (Fig. 1). Such mRNAs therefore also initiate proteins at start codons which are not 5Ј proximal. To account for initiation on these bicistronic mRNAs, a modified scanning model which allows for leaky scanning, i.e., scanning in which the scanning complex can bypass the first start codon to some extent when it is not in an optimal context for initiation and thereby can initiate at a downstream start codon, was proposed (23). This model followed from the realization that the nucleotide sequence surrounding the start codon (the context) was important for the efficiency with which ribosomes initiate protein synthesis, with the strongest determinant for efficiency being a purine at position Ϫ3, followed by a G at position ϩ4.Eukaryotic mRNAs known to initiate proteins at two start codons are nevertheless uncommon, and those that initiate at more than two start codons are rare. The most extreme example of an mRNA known to initiate proteins at multiple start sites remains the SeV P/C mRNA, which uses five start codons located 81 to 201 nucleotides (nt) from the 5Ј end and a start codon located more than 1500 nt from the 5Ј end to generate eight primary translation products (CЈ, P, C, Y1, Y2, V, W, and X) (Fig. 1). The second start site (and 5Ј-proximal ATG codon) generates three proteins (P, V, and W) containing the same N-terminal 317 amino acids (aa) but different C-terminal regions, as this mRNA is cotranscriptionally edited by the SeV polymerase via the programmed insertion of zero, one, or two G residues (45). The X protein, which represents approximately the C-terminal 95 aa of the 568-aa-long P protein, is thought to be initiated in a cap-dependent but scanning-independent manner (5). The first ribosomal start site (an unusual ACG start codon) and the third, fourth, and fifth start sites (all ATG codons) generate a nested set of four C proteins (termed CЈ, C, Y1, and Y2, of 215, 204, 183, and 175 aa, respectively) with a common C terminus, irrespective of whether the mRNA is edited (6,18,33). The Y proteins do not appear to be generated by proteolytic cleavage of the C or CЈ protein, because a recombinant SeV which does not express either C or CЈ (due to mutation of their start codons) overexpresses the Y proteins (25a). The available evidence suggests that whereas the first three start sites (for the CЈ, P, and C proteins) are accessed by leaky scanning, it is unlikely that the last two start sites (for the Y proteins) are accessed by a mechanism involvi...

We have studied the relationship between the Sendai virus (SeV) C proteins (a nested set of four proteins initiated at different start codons) and the interferon (IFN)-mediated antiviral response in IFN-competent cells in culture. SeV strains containing wild-type or various mutant C proteins were examined for their ability (i) to induce an antiviral state (i.e., to prevent the growth of vesicular stomatitis virus [VSV] following a period of SeV infection), (ii) to induce the elevation of Stat1 protein levels, and (iii) to prevent IFN added concomitant with the SeV infection from inducing an antiviral state. We find that expression of the wild-type C gene and, specifically, the AUG114-initiated C protein prevents the establishment of an antiviral state: i.e., cells infected with wild-type SeV exhibited little or no increase in Stat1 levels and were permissive for VSV replication, even in the presence of exogenous IFN. In contrast, in cells infected with SeV lacking the AUG114-initiated C protein or containing a single amino acid substitution in the C protein, the level of Stat1 increased and VSV replication was inhibited. The prevention of the cellular IFN-mediated antiviral response appears to be a key determinant of SeV pathogenicity.

No abstract

Recombinant Sendai viruses were prepared which cannot express their Cprime, C, or Cprime plus C proteins due to mutation of their respective start codons ([Cprime-minus], [C-minus] and [double mutant], respectively). The [Cprime-minus] and [C-minus] stocks were similar to that of wild-type (wt) virus in virus titer and plaque formation, whereas the double-mutant stock had a much-reduced PFU or 50% egg infective dose/particle ratio and produced very small plaques. Relative to the wt virus infection, the [Cprime-minus] and [C-minus] infections of BHK cells resulted in significantly greater accumulation of viral RNAs, consistent with the known inhibitory effects of the Cprime and C proteins. The double-mutant infection, in contrast, was delayed in its accumulation of viral RNAs; however, once accumulation started, overaccumulation quickly occurred, as in the single-mutant infections. Our results suggest that the Cprime and C proteins both provide a common positive function early in infection, so that only the double mutant undergoes delayed RNA accumulation and exhibits the highly debilitated phenotype. Later in infection, the same proteins appear to act as inhibitors of RNA accumulation. In infections of mice, [Cprime-minus] was found to be as virulent as wt virus whereas [C-minus] was highly attenuated. These results suggest that the Cprime and C proteins cannot be functionally equivalent, since C can replace Cprime for virulence in mice whereas Cprime cannot replace C.

No abstract