Superinfection exclusion of alphaviruses in three mosquito cell lines persistently infected with Sindbis virus.

Three Aedes albopictus (mosquito) cell lines persistently infected with Sindbis virus excluded the replication of both homologous (various strains of Sindbis) and heterologous (Aura, Semliki Forest, and Ross River) alphaviruses. In contrast, an unrelated flavivirus, yellow fever virus, replicated equally well in uninfected and persistently infected cells of each line. Sindbis virus and Semliki Forest virus are among the most distantly related alphaviruses, and our results thus indicate that mosquito cells persistently infected with Sindbis virus are broadly able to exclude other alphaviruses but that exclusion is restricted to members of the alphavirus genus. Superinfection exclusion occurred to the same extent in three biologically distinct cell clones, indicating that the expression of superinfection exclusion is conserved among A. albopictus cell types. Superinfection of persistently infected C7-10 cells, which show a severe cytopathic effect during primary Sindbis virus infection, by homologous virus does not produce cytopathology, consistent with the idea that cytopathology requires significant levels of viral replication. A possible model for the molecular basis of superinfection exclusion, which suggests a central role for the alphavirus trans-acting protease that processes the nonstructural proteins, is discussed in light of these results.

Multiple binding sites for cellular proteins in the 3' end of Sindbis alphavirus minus-sense RNA.

The 3' end of Sindbis virus minus-sense RNA was tested for its ability to bind proteins in mosquito cell extracts, using labeled riboprobes that represented different parts of this region. We found four domains in the first 250 nucleotides that could bind the same 50- and 52-kDa proteins, three with high affinity and one with low affinity, whereas tested domains outside this region did not bind these proteins. The first binding domain was found in the first 60 nucleotides, which represents the complement of the 5'-nontranslated region, the second in the next 60 nucleotides, the third in the following 60 nucleotides, and the fourth between nucleotides 194 and 249 (all numbering is 3' to 5'). The relative binding constants, Kr, of the first, second, and fourth sites were similar, whereas that of domain 2 was fivefold less. Deletion mapping of the first domain showed that the first 10 nucleotides were critical for binding. Deletion of nucleotides 2 to 4, deletion or replacement of nucleotide 5, or deletion of the first 15 nucleotides was deleterious for binding, deletion of nucleotides 10 to 15, 26 to 40, or 41 to 55 had little effect on the binding, and deletion of nucleotides 15 to to 25 increased the binding affinity. We also found that the corresponding riboprobes derived from two other alphaviruses, Ross River virus and Semliki Forest virus, and from rubella virus were also able to interact with the 50- and 52-kDa proteins. The Kr value for the Semliki Forest virus probe was similar to that for the Sindbis virus probe, while that for the Ross River virus probe was four times greater. The rubella virus probe was bound only weakly, consistent with the fact that mosquito cells are not permissive for rubella virus replication. We suggest that the binding of the 50- and 52-kDa proteins to the 3' end of alphavirus minus-sense RNA represents an important step in the initiation of RNA replication.

Flavivirus enzyme-substrate interactions studied with chimeric proteinases: identification of an intragenic locus important for substrate recognition.

The proteins of flaviviruses are translated as a single long polyprotein which is co- and posttranslationally processed by both cellular and viral proteinases. We have studied the processing of flavivirus polyproteins in vitro by a viral proteinase located within protein NS3 that cleaves at least three sites within the nonstructural region of the polyprotein, acting primarily autocatalytically. Recombinant polyproteins in which part of the polyprotein is derived from yellow fever virus and part from dengue virus were used. We found that polyproteins containing the yellow fever virus cleavage sites were processed efficiently by the yellow fever virus enzyme, by the dengue virus enzyme, and by various chimeric enzymes. In contrast, dengue virus cleavage sites were cleaved inefficiently by the dengue virus enzyme and not at all by the yellow fever virus enzyme. Studies with chimeric proteinases and with site-directed mutants provided evidence for a direct interaction between the cleavage sites and the proposed substrate-binding pocket of the enzyme. We also found that the efficiency and order of processing could be altered by site-directed mutagenesis of the proposed substrate-binding pocket.

Expression of the structural proteins of dengue 2 virus and yellow fever virus by recombinant vaccinia viruses.

Vaccinia virus recombinants were constructed which contained cDNA sequences encoding the structural region of dengue 2 virus (PR159/S1 strain) or yellow fever virus (17D strain). The flavivirus cDNA sequences were expressed under the control of the vaccinia 7.5k early/late promotor. Cultured cells infected with these recombinants expressed immunologically reactive flavivirus structural proteins, precursor prM and E. These proteins appeared to be cleaved and glycosylated properly since they comigrated with the authentic proteins from dengue 2 virus- and yellow fever virus-infected cells. Mice immunized with the dengue/vaccinia recombinant showed a dengue-specific immune response that included low levels of neutralizing antibodies. Immunization of mice with the yellow fever/vaccinia recombinant was less effective at inducing an immune response to yellow fever virus and in only some of the mice were low titers of neutralizing antibodies produced.

Sindbis virus ts103 has a mutation in glycoprotein E2 that leads to defective assembly of virions.

Sindbis virus mutant ts103 is aberrant in the assembly of virus particles. During virus budding, proper nucleocapsid-glycoprotein interactions fail to occur such that particles containing many nucleocapsids are formed, and the final yield of virus is low. We have determined that a mutation in the external domain of glycoprotein E2, Ala-344----Val, is the change that leads to this phenotype. Mapping was done by making recombinant viruses between ts103 and a parental strain of the virus, using a full-length cDNA clone of Sindbis virus from which infectious RNA can be transcribed, together with sequence analysis of the region of the genome shown in this way to contain the ts103 lesion. A partial revertant of ts103, called ts103R, was also mapped and sequenced and found to be a second-site revertant in which a change in glycoprotein E1 from lysine to methionine at position 227 partially suppresses the phenotypic effects of the change at E2 position 344. An analysis of revertants from ts103 mutants in which the Ala----Val change had been transferred into a defined background showed that pseudorevertants were more likely to arise than were true revertants and that the ts103 change itself reverted very infrequently. The assembly defect in ts103 appeared to result from weakened interactions between the virus membrane glycoproteins or between these glycoproteins and the nucleocapsid during budding. Both the E2 mutation leading to the defect in virus assembly and the suppressor mutation in glycoprotein E1 are in the domains external to the lipid bilayer and thus in domains that cannot interact directly with the nucleocapsid. This suggests that in ts103, either the E1-E2 heterodimers or the trimeric spikes (consisting of three E1-E2 heterodimers) are unstable or have an aberrant configuration, and thus do not interact properly with the nucleocapsid, or cannot assembly correctly to form the proper icosahedral array on the surface of the virus.

Nucleotide sequence of yellow fever virus: implications for flavivirus gene expression and evolution.

The sequence of the entire RNA genome of the type flavivirus, yellow fever virus, has been obtained. Inspection of this sequence reveals a single long open reading frame of 10,233 nucleotides, which could encode a polypeptide of 3411 amino acids. The structural proteins are found within the amino-terminal 780 residues of this polyprotein; the remainder of the open reading frame consists of nonstructural viral polypeptides. This genome organization implies that mature viral proteins are produced by posttranslational cleavage of a polyprotein precursor and has implications for flavivirus RNA replication and for the evolutionary relation of this virus family to other RNA viruses.

Amino-terminal amino acid sequences of structural proteins of three flaviviruses.

N-terminal amino acid sequences of structural proteins of three flaviviruses, yellow fever, St. Louis encephalitis, and dengue-2 viruses, have been obtained. The glycoproteins of these three viruses are 52-60% conserved in the region sequenced, depending upon which pair of viruses are compared, and 40% of the amino acids are invariant in all three viruses. Thus, flaviviruses are closely related and have in all probability descended from a common ancestor. Furthermore, residues important in the secondary structure of proteins are conserved, suggesting that the overall conformation of the glycoproteins is the same in all three viruses while considerable variation in the primary sequence can be accommodated. The N-terminal regions of the nucleocapsid proteins of yellow fever and St. Louis encephalitis viruses show markedly less homology (25%) and this region is highly basic with one-quarter (yellow fever) or one-third (St. Louis encephalitis) of the residues being lysine or arginine. N-terminal sequences for the M protein of yellow fever and for NV2(GP19) of St. Louis encephalitis viruses are also reported.

Sequence analysis of two mutants of Sindbis virus defective in the intracellular transport of their glycoproteins.

We have sequenced the complementary DNA corresponding to the genes encoding the viral glycoproteins of ts10 and ts23, mutants of Sindbis virus defective in the intracellular transport of their glycoproteins, and of revertants of these mutants. These studies have been augmented by direct amino acid sequencing of the amino-terminal regions of the glycoproteins of several virus strains. By comparing the deduced amino acid sequence with that of Sindbis HR virus, the parental strain of these mutants, and with the sequence of the revertants, we found ts23 to have a double mutation in glycoprotein E1, while ts10 was a single mutant in the same glycoprotein. In each case reversion to temperature insensitivity occurred by changes at the same site as the mutation, in two cases restoring the original amino acid and in the third case substituting an homologous amino acid (arginine in place of lysine). The three mutations were far apart from each other in the protein, suggesting that the three-dimensional conformation is very important for the correct migration of the glycoproteins from the rough endoplasmic reticulum to the plasma membrane. The sequence data also reveal that a number of other changes have occurred in the various virus strains during mutagenesis or passage.

Growth and release of several alphaviruses in chick and BHK cells.

The growth and release of several alphaviruses, including several strains of Sindbis virus (the wild-type strain, the large plaque and small plaque variants of the HR strain, and the HR mutant ts103), Semliki Forest virus(SFV) and Middelburg virus, and of the unrelated rhabdovirus, vesicular stomatitis virus (VSV), have been compared in chick cells and in BHK-21 cells as a function of the culture conditions for the host cell and the ionic strength of the medium. The small plaque strain of Sindbis HR, as well as SFV, grew better in BHK cells, whereas the large plaque strain of Sindbis HR showed a preference for chick cells. Wild-type Sindbis and VSV grew equally well in either cell. The optimum ionic strength for virus production as well as inhibition of virus release into the medium at low ionic strength depended upon both the virus and the host cell. Thus, VSV grown in medium of low ionic strength gave no additional release of virus on incubation with hypertonic medium (minimum effect), whereas ts103 released very little virus without exposure to hypertonic conditions (maximum effect). The viruses could be ordered as follows: minimum effect = vesicular stomatitis virus < Middelburg virus < Semliki Forest virus < Sindbis wt < Sindbis HR (large plaque) < Sindbis HR (small plaque) < Sindbis ts103 = maximum effect. After several passages in culture, chick cells required hypertonic conditions for optimum production and release of Sindbis virus. Furthermore, BHK cells cultured in different media responded differently to ionic strength for virus production and release. These results suggest that there is a charge-dependent stop in the maturation of alpha-viruses, possibly a configurational rearrangement of glycoprotein E2 upon its formation from the precursor PE2, which is sensitive to the ionic strength of the medium, to the composition of the host plasmalemma and to differences in the virus glycoproteins.