Elementary aspects of autointerference and the replication of defective interfering virus particles

The Impact of Defective Viruses on Infection and Immunity

Defective viral genomes (DVGs) are generated during viral replication and are unable to carry out a full replication cycle unless coinfected with a full-length virus. DVGs are produced by many viruses, and their presence correlates with alterations in infection outcomes. Historically, DVGs were studied for their ability to interfere with standard virus replication as well as for their association with viral persistence. More recently, a critical role for DVGs in inducing the innate immune response during infection was appreciated. Here we review the role of DVGs of RNA viruses in shaping outcomes of experimental as well as natural infections and explore the mechanisms by which DVGs impact infection outcome.

Rescue of a Sendai virus DI genome by other parainfluenza viruses: implications for genome replication

Using a defective interfering Sendai virus stock (DIH4) freed of nondefective helper virus, we found that the closely related parainfluenza viruses 1 and 3 could substitute for the Sendai virus helper in replicating DIH4, creating chimeric nucleocapsids. The morbillivirus measles and the rhabdovirus VSV could not substitute. When DIH4 is incubated intracellularly for 5 days in the absence of help, the ability of PIV3 to rescue DIH4 at this time depended on fresh Sendai virus polymerase. The PIV3 polymerase apparently can only copy the chimeric template, but not that wrapped in the homologous Sendai NP protein. These results suggest that the cis-acting RNA sequences important for genome replication, e.g., the promoter and the encapsidation site, have been conserved among these viruses, but that the interactions between the polymerase and the template protein NP are unique for each virus.

Replication of the genome RNAs of defective interfering particles of vesicular stomatitis and Sendai viruses using heterologous viral proteins

We have tested the ability of heterologous viral proteins to support the in vivo and in vitro replication of the RNA of defective interfering (DI) particles of two serotypes of VSV and of Sendai virus. In all the combinations of heterologous coinfections in vivo, DI particle replication was observed only in the coinfection with the VSV-Indiana DI particle and wild-type VSV-New Jersey. By quantitating RNA synthesis in reconstitution experiments we showed that with DI nucleocapsids isolated from infected cells, however, the soluble protein fraction from heterologous wild-type virus-infected cells could substitute in vitro to varying degrees for the homologous proteins in the elongation reaction of RNA replication and encapsidation. In these cases successful replication was confirmed by demonstrating the specific association of the heterologous N protein with the product nucleocapsid RNA. The initiation step, that is, the initial binding of the nucleocapsid protein to the leader RNA, in contrast, requires the homologous protein, since heterologous viral proteins could not support RNA replication and encapsidation from purified DI particles.

The 5'-terminal sequence of VSV(NJ) (Ogden): is the interaction of the NS protein with the NS binding site responsible for heterotypic interference activity?

The 5'-terminal sequence of VSV(NJ) (Ogden) and VSV(NJ) (Hazelhurst) was compared in an attempt to understand why the defective interfering particle, DI-LT, heterotypically interferes with VSV(NJ) (Ogden) but not with VSV(NJ) (Hazelhurst). The 5'-terminal sequence of VSV(NJ) (Ogden) genomic RNA was determined by direct RNA sequencing and by DNA sequencing of cDNA clones of the 3'-terminal sequence of VSV(NJ) (Ogden) DI particle genome. Primer extension analysis of the 5'-terminus of VSV(NJ) (Ogden) standard genomic RNA confirmed these data. Within the last 47 nucleotides, equivalent to the negative-strand leader RNA, the only nucleotide changes between VSV(NJ) (Ogden) and VSV(NJ) (Hazelhurst) occur between nucleotides 19 and 26, representing part of the putative NS binding region described by Isaac and Keene (J. Virol. 43, 241-249 (1982] for VSV(IND) DI particles. The spacer (S) region, located between the polyadenylation signal of the L gene and the 47th nucleotide of the leader RNA, contains more differences. The polyadenylation signal of the L gene is fully conserved, but the remainder of the L gene region (177 nucleotides) has highly diverged between VSV(NJ) (Ogden) and VSV(NJ) (Hazelhurst). The changes in the NS binding region of the negative-strand leader RNA provide further evidence for the divergent evolution of VSV(NJ) (Ogden) and VSV(NJ) (Hazelhurst). The NS binding region has been implicated as a crucial site for the initiation of RNA transcription and replication. The interaction of the NS protein with this site may determine the ability of DI particles to interfere heterotypically.

Defective interfering particles of VSVNJ (Ogden), generated by heat treatment, contain multiple internal genomic deletions

Defective interfering (DI) particles have been isolated from a heat-resistant strain of the New Jersey (Ogden) serotype of vesicular stomatitis virus (VSV). Most of these DI particles contain various portions of all five cistrons of VSV. The two largest DI particles, NJ-121 and NJ-PG2, represent approximately 60% of the standard virus genome and contain both the positive and negative strand leader RNA templates. These two DI particles are transcriptionally active and synthesize both the positive and negative strand leader RNAs in vitro. Virion RNA probe-mRNA hybridizations and cDNA probe-virion RNA hybridizations have shown that NJ-121 contains mainly sequences from the L and G genes. In contrast, NJ-PG2 has portions of the sequences from all five genes of VSV. Smaller DI particles, NJ-121a, NJ-121b, NJ-PG1, and NJ-JM2 representing approximately 50, 38, 28, and 25% of the standard virus genome respectively, were also generated. These DI particles did not have sequences complementary to the positive strand leader RNA template. The mRNA hybridization patterns and results of the genomic RNAs hybridizing to cDNAs of N, NS, M, and G genes of these DI particles showed that they contain parts of information from all five cistrons. Most of the DI particles appear to be generated by multiple deletions throughout the standard virus genome. None of these DI particles interfered heterotypically with VSVIND-HR in BHK21, R(B77), or L2 cells. However, they interfered well with infection by VSVNJ (Hazelhurst).

Frequent generation of new 3'-defective interfering particles of vesicular stomatitis virus

We have isolated and partially characterized a number of different genome types of defective interfering (DI) particles newly generated by a highly heat-resistant strain of vesicular stomatitis virus in either Rat(B77) or Vero cells. Northern blot analyses revealed that many of these DI genomes contain N gene sequences and/or sequences of the NS, M, and G genes. One type contains NS sequences without any indication for the presence of either N, M, or G sequences. Another type of DI particle genomes did not contain any detectable sequences of N, NS, M, or G, but contain panhandle-type sequences and, thus, most likely resembles the 5'-panhandle-type DI particles. Unlike previously assumed, these data demonstrate that DI genomes which have the 3'-terminal N, NS, M, and G genes or portions of these genes conserved do frequently arise together with 5'-DI particle genomes after serial undiluted passages of the heat-resistant strain of vesicular stomatitis virus.

Specific interactions of vesicular stomatitis virus L and NS proteins with heterologous genome ribonucleoprotein template lead to mRNA synthesis in vitro

Two dissociable proteins, L and NS, and N-RNA template were purified from two serologically distinct vesicular stomatitis viruses, Indiana [VSV(IND)] and New Jersey [VSV(NJ)]. Requirements for RNA synthesis in heterologous reconstitution reactions in vitro were studied. The L and NS proteins of VSV(NJ) failed to synthesize full-length leader RNA and mRNAs in vitro when reconstituted with N-RNA(IND) template. However, when purified homologous NS(IND) was added to the reaction mixture, mRNA synthesis ensued. The requirements for transcription of N-RNA(NJ) template were different from those for N-RNA(IND). For RNA synthesis, transcription specifically required L(NJ), but the NS(NJ) and NS(IND) proteins were interchangeable. This suggests that there are specific domains on the L(NJ) protein, at which NS proteins of both serotypes may interact to form an active RNA polymerase complex, whereas L(IND) lacked such domains for interaction with NS(NJ). The function of the L protein appeared primarily to initiate RNA chains, and the NS protein was required for chain elongation. The results of these in vitro complementation experiments are discussed in light of previous in vivo complementation studies.

A replication-efficient mutant of West Nile virus is insensitive to DI particle interference

A previous report described the isolation of a mutant of West Nile virus (WNV) from culture fluid obtained from persistently infected genetically resistant C3H/RV mouse cells that replicates significantly more efficiently in cultures of C3H/RV cells than does the parental virus. This replication-efficient mutant, designated RE-WNV, has now been found to be insensitive to interference by WNV defective interfering (DI) particles. This characteristic was demonstrated by several means. The RE-WNV mutant was able to superinfect persistently infected cultures that were no longer producing detectable parental virus, while the parental virus was not. Good yields of the mutant virus were produced during six serial undiluted passages of RE-WNV in both resistant C3H/RV and congenic susceptible C3H/HE cells. In contrast, during passage of parental virus in C3H/RV cells, progeny virus could not be detected after the third passage, due to an enhanced interference by WNV DI particles with standard virus replication in these cells. The RE-WNV was also insensitive to interference by a pool of parental virus enriched for DI particles. Analysis of the mutant genome by oligonucleotide fingerprinting indicated that the genome RNA of the mutant differs by two unique spots from the parental RNA. The relevance of this mutant to the eventual understanding of the mechanism by which C3H/RV and C3H/HE cells manifest their flavivirus-specific difference in the efficiency of progeny virus production is discussed.

In vitro and in vivo inhibition of primary transcription of vesicular stomatitis virus by a defective interfering particle

A unique defective interfering (DI) particle, generated by a heat-resistant (HR) mutant of Indiana serotype vesicular stomatitis virus, was capable of inhibiting primary transcription by heterologous New Jersey serotype virions. The correlation between this phenomenon and the lowering of viral yields from doubly infected cells was investigated by the construction of chimeric DI particles containing the HR DI particle genome with a thermolabile polymerase. At the nonpermissive temperature, these DI particles were unable to self-transcribe, inhibit virion primary transcription, or reduce virion yield, but were able to be replicated. These results suggested that self-transcription of the HR DI particle genome was a prerequisite for heterotypic interference, but not for its own replication. Inhibition of virion primary transcription by HR DI ribonucleocapsids was also observed in vitro. At low HR DI to virion ribonucleocapsid ratios, the extent of inhibition was concentration dependent, whereas at high ratios, the amount of inhibition was concentration independent, approaching a limiting maximum value. A speculative mathematical model, which quantitatively accounts for these data, is presented. According to this model, the higher affinity for polymerase molecules by the HR DI ribonucleocapsids is explained in terms of dissociation events during transcription, which are more frequent in the longer virion ribonucleocapsids.

Inhibition of cellular DNA synthesis by vesicular stomatitis virus

DNA synthesis in mouse myeloma (MPC-11) cells and L cells was rapidly and progressively inhibited by infection with vesicular stomatitis virus (VSV). No significant difference in cellular DNA synthesis inhibition was noted between synchronized and unsynchronized cells, nor did synchronized cells vary in their susceptibility to VSV infection after release from successive thymidine and hydroxyurea blocks. Cellular RNA synthesis was inhibited to about the same extent as DNA synthesis, but cellular protein synthesis was less affected by VSV at the same multiplicity of infection. The effect of VSV on cellular DNA synthesis could not be attributed to degradation of existing DNA or to decreased uptake of deoxynucleoside triphosphates, nor were DNA polymerase and thymidine kinase activities significantly different in VSV-infected and uninfected cell extracts. Analysis by alkaline sucrose gradients of DNA in pulse-labeled uninfected and VSV-infected cells indicated that VSV infection did not appear to influence DNA chain elongation. Cellular DNA synthesis was not significantly inhibited by infection with the VSV polymerase mutant tsG114(I) at the restrictive temperature or by infection with defective-interfering VSV DI-011 (5' end of the genome), but DI-HR-LT (3' end of genome) exhibited initially rapid but not prolonged inhibition of MPC-11 cell DNA synthesis. DNA synthesis inhibitory activity of wild-type VSV was only slowly and partially inactivated by very large doses of UV irradiation. These data suggest that, as in the effect of VSV on cellular RNA synthesis (Weck et al., J. Virol. 30:746-753, 1979), inhibition of cellular DNA synthesis by VSV requires transcription of a small segment of the viral genome.

Vesicular stomatitis virus defective interfering particle containing a muted internal leader RNA gene

The RNA of a unique long defective interfering particle (DI-LT2) derived from the heat-resistant strain of vesicular stomatitis virus (VSV) contains 70 nucleotides at its 3' end that are complementary to the 5' end of the VSV RNA. Following this region of terminal complementarity, there is a precise copy of the 3' end of the nondefective VSV RNA. The sequence homology between the DI-LT2 RNA and the 3' end of VSV RNA extends for at least 60 bases and probably for most of the length of the DI-LT2 RNA. The DI-LT2 particle is capable of transcription in vitro but produces only a short RNA [defective interfering (DI) particle product], which is encoded by the extreme 3' terminus of the DI RNA. Neither leader RNA nor capped VSV mRNAs are synthesized by DI-LT2, although competent templates for these are present. These data suggest that the 3'-terminal initiation is a prerequisite of the production of competent transcripts and that the sequence coding for leader RNA is not, by itself, sufficient for initiation. We propose a model for the origin of this DI particle, involving specific termination and resumption of replication, which is similar to that described previously for another class of DI particle RNAs.

Defective interfering particles of Sindbis virus do not interfere with the homologous virus obtained from persistently infected BHK cells but do interfere with Semliki Forest virus

Defective interfering particles derived from wild-type Sindbis virus no longer interfere with the infectious virus cloned from BHK cells persistently infected with Sindbis virus for 16 months. These particles do interfere with the replication of Semliki Forest virus.

Defective interfering particle generated by internal deletion of the vesicular stomatitis virus genome

The genome structure of the long, truncated defective interfering particle derived from the heat-resistant strain of vesicular stomatitis virus has been examined. Stocks of this defective interfering particle are shown to contain several different species having information primarily from the 3' half of the vesicular stomatitis virus genome; the proportions of these components vary depending on the passage history of the stock. The two most abundant types have been identified and characterized. One has complementary 5' and 3' termini and consequently appears as a circular molecule when examined by electron microscopy. The other cannot circularize and remains linear. The circular forms are consistently 8 to 10% longer than the linear molecules. Rapid sequencing analyses reveal that both forms retain the 5' parental viral terminal sequence, but only the linear form retains the parental 3'-terminal sequence which is the complement of the 5' end. Hybridization experiments and electron microscopic analyses indicate that the linear form has retained 320 to 350 nucleotides of the 5' parental sequence and was probably generated by an internal deletion of the vesicular stomatitis virus genome.