Recovery of infectious SV5 from cloned DNA and expression of a foreign gene

Generation of recombinant lentogenic Newcastle disease virus from cDNA

Recombinant lentogenic Newcastle disease virus (NDV) of the vaccine strain Clone-30 was reproducibly generated after simultaneous expression of antigenome-sense NDV RNA and NDV nucleoprotein, phosphoprotein and RNA-dependent RNA polymerase from plasmids transfected into cells stably expressing T7 RNA polymerase. For this purpose, the genome of Clone-30, comprising 15186 nt, was cloned and sequenced prior to assembly into a full-length cDNA clone under control of a T7 RNA polymerase promoter. Recombinant virus was amplified by inoculation of transfection supernatant into the allantoic cavity of embryonated specific-pathogen-free (SPF) chicken eggs. Two marker restriction sites comprising a total of five nucleotide changes artificially introduced into noncoding regions were present in the progeny virus. The recombinant NDV was indistinguishable from the parental wild-type virus with respect to its growth characteristics in cell culture and in embryonated eggs. Moreover, an intracerebral pathogenicity index of 0.29 was obtained for both viruses as determined by intracerebral inoculation of day-old SPF chickens, proving that the recombinant NDV is a faithful copy of the parental vaccine strain of NDV.

Involvement of the cytoplasmic domain of the hemagglutinin-neuraminidase protein in assembly of the paramyxovirus simian virus 5

Efficient assembly of enveloped viruses at the plasma membranes of virus-infected cells requires coordination between cytosolic viral components and viral integral membrane glycoproteins. As viral glycoprotein cytoplasmic domains may play a role in this coordination, we have investigated the importance of the hemagglutinin-neuraminidase (HN) protein cytoplasmic domain in the assembly of the nonsegmented negative-strand RNA paramyxovirus simian virus 5 (SV5). By using reverse genetics, recombinant viruses which contain HN with truncated cytoplasmic tails were generated. These viruses were shown to be replication impaired, as judged by small plaque size, reduced replication rate, and low maximum titers when compared to those features of wild-type (wt) SV5. Release of progeny virus particles from cells infected with HN cytoplasmic-tail-truncated viruses was inefficient compared to that of wt virus, but syncytium formation was enhanced. Furthermore, accumulation of viral proteins at presumptive budding sites on the plasma membranes of infected cells was prevented by HN cytoplasmic tail truncations. We interpret these data to indicate that formation of budding complexes, from which efficient release of SV5 particles can occur, depends on the presence of an HN cytoplasmic tail.

The signal for clathrin-mediated endocytosis of the paramyxovirus SV5 HN protein resides at the transmembrane domain-ectodomain boundary region

The hemagglutinin-neuraminidase (HN) glycoprotein of the paramyxovirus SV5 is internalized from the cell surface via clathrin-coated pits. However, the cytoplasmic domain of SV5 HN does not contain a previously characterized internalization motif. A cell-surface-expressed chimeric protein (APK), consisting of the cytoplasmic tail, transmembrane (TM) domain, and 12 residues of the ectodomain of HN joined to the cytoplasmic protein pyruvate kinase is internalized, indicating that the N-terminal region of HN contains an internalization signal. Although SV5 HN is internalized at a rate similar to that of influenza virus hemagglutinin (HA) mutant Y543, which contains a degenerate tyrosine-based signal in its cytoplasmic tail, the elimination of the majority of the HN cytoplasmic tail, or substitution of the HN TM domain with leucine residues, did not affect the rate of HN internalization. The HN protein of the closely related virus, Newcastle disease virus (NDV), is not internalized from the cell surface. Working under the usual convention that the TM domain consists of the hydrophobic residues bounded by two charged residues, analysis of internalization of mutant and chimeric NDV HN molecules indicates that the first seven SV5 HN ectodomain residues are critical for internalization of HN. A glutamic acid residue (E37) that abuts this presumptive HN TM domain/ectodomain boundary is important for SV5 HN internalization.

Mutations in the C, D, and V open reading frames of human parainfluenza virus type 3 attenuate replication in rodents and primates

Human parainfluenza virus type 3 (HPIV3) is a single-stranded negative-sense RNA virus belonging to the Respirovirus genus of the Paramyxoviridae family in the order Mononegavirales. The P gene encodes at least four proteins, including the C protein, which is expressed from an open reading frame (ORF) that overlaps the P ORF, and the D protein, which is encoded when the P ORF is fused to the D ORF by transcriptional editing. The P mRNA also contains a third ORF for the V protein, although it is unclear how or whether this ORF is accessed. We have used recombinant DNA technology to recover five mutant viruses that either interrupt or alter the C, D, and V ORFs. In one mutant virus, rC-KO, expression of the C protein was abrogated by changing the start codon from methionine to threonine and introducing two stop codons at amino acid positions 7 and 26 of the C ORF. In a second mutant virus, rF164S, a point mutation was introduced into the C ORF changing amino acid position 164 from phenylalanine (F) to serine (S), which corresponds to the F170S mutation described in the C protein of Sendai virus (Itoh et al., J. Gen. Virol. 78, 3207-3215). rC-KO was significantly attenuated in vitro and in vivo (rodents and primates), whereas rF164S was attenuated only in vivo. Interestingly, the rF164S mutant was more attenuated in the upper than in the lower respiratory tract of hamsters and monkeys. This pattern is the converse of that seen with temperature-sensitive attenuating mutations, and thus inclusion of this novel mutation in a recombinant live-attenuated vaccine candidate might prove useful in reducing residual virulence in the upper respiratory tract. Both rC-KO and rF164S conferred protection against challenge with wild-type HPIV3. In three other viruses, the D and V ORFs were interrupted singly or in combination. Although interruption of the D and V ORFs individually did not affect virus replication in vitro or in vivo, interruption of both together attenuated replication in vivo. These results indicate that the C, D, and V proteins of HPIV3 each has a role in virus replication in vitro, in vivo, or both, and define mutations that might be useful for the development of a vaccine against HPIV3.

Generation of influenza A viruses entirely from cloned cDNAs

We describe a new reverse-genetics system that allows one to efficiently generate influenza A viruses entirely from cloned cDNAs. Human embryonic kidney cells (293T) were transfected with eight plasmids, each encoding a viral RNA of the A/WSN/33 (H1N1) or A/PR/8/34 (H1N1) virus, flanked by the human RNA polymerase I promoter and the mouse RNA polymerase I terminator-together with plasmids encoding viral nucleoprotein and the PB2, PB1, and PA viral polymerases. This strategy yielded >1 x 10(3) plaque-forming units (pfu) of virus per ml of supernatant at 48 hr posttransfection. The addition of plasmids expressing all of the remaining viral structural proteins led to a substantial increase in virus production, 3 x 10(4)-5 x 10(7) pfu/ml. We also used reverse genetics to generate a reassortant virus containing the PB1 gene of the A/PR/8/34 virus, with all other genes representing A/WSN/33. Additional viruses produced by this method had mutations in the PA gene or possessed a foreign epitope in the head of the neuraminidase protein. This efficient system, which does not require helper virus infection, should be useful in viral mutagenesis studies and in the production of vaccines and gene therapy vectors.

Effect of inserting paramyxovirus simian virus 5 gene junctions at the HN/L gene junction: analysis of accumulation of mRNAs transcribed from rescued viable viruses

Simian parainfluenza virus 5 (SV5) is a prototype of the Paramyxoviridae family of nonsegmented negative-sense RNA viruses. The single-stranded RNA genomes of these viruses contain a series of tandemly linked genes separated by intergenic (IG) sequences flanked by gene-end (GE) and gene-start (GS) sequences. The viral RNA polymerase (vRNAP) complex is thought to enter the genome at its 3' end, and synthesis of mRNAs is thought to occur by a stop-start mechanism in a sequential and polar manner, with transcriptional attenuation occurring primarily at the intergenic regions. As a result, multiple nonoverlapping mRNA species are generated for each single entry of the vRNAP. To investigate the functions of GE, IG, and GS sequences in transcription, we constructed plasmids containing cDNAs of the full-length SV5 genome in which the gene junction sequences (GE, IG, and GS sequences) located between the hemagglutinin-neuraminidase (HN) and the polymerase (L) genes were replaced with the counterpart sequences from other gene junctions. By using reverse genetics, we recovered viable viruses from each cDNA construct, although their growth characteristics varied. Analysis of the HN and L mRNAs by quantitative RNase protection assay indicated that the ratios of HN to L mRNAs varied over a fourfold range. The alteration of the gene junction sequences also permitted examination of the hypothesized requirement for hexamer nucleotide position of the GS sites. The recovery of infectious viruses with transcription initiation sites that occurred at nucleotide positions 1, 2, 3, 5, and 6 of the hexamer suggest that the requirement is nonstringent.

Rescue of Newcastle disease virus from cloned cDNA: evidence that cleavability of the fusion protein is a major determinant for virulence

A full-length cDNA clone of Newcastle disease virus (NDV) vaccine strain LaSota was assembled from subgenomic overlapping cDNA fragments and cloned in a transcription plasmid between the T7 RNA polymerase promoter and the autocatalytic hepatitis delta virus ribozyme. Transfection of this plasmid into cells that were infected with a recombinant fowlpoxvirus that expressed T7 RNA polymerase, resulted in the synthesis of antigenomic NDV RNA. This RNA was replicated and transcribed by the viral NP, P, and L proteins, which were expressed from cotransfected plasmids. After inoculation of the transfection supernatant into embryonated specific-pathogen-free eggs, infectious virus derived from the cloned cDNA was recovered. By introducing three nucleotide changes in the cDNA, we generated a genetically tagged derivative of the LaSota strain in which the amino acid sequence of the protease cleavage site (GGRQGR downward arrowL) of the fusion protein F0 was changed to the consensus cleavage site of virulent NDV strains (GRRQRR downward arrowF). Pathogenicity tests in day-old chickens showed that the strain derived from the unmodified cDNA was completely nonvirulent (intracerebral pathogenicity index [ICPI] = 0.00). However, the strain derived from the cDNA in which the protease cleavage site was modified showed a dramatic increase in virulence (ICPI = 1.28 out of a possible maximum of 2.0). Pulse-chase labeling of cells infected with the different strains followed by radioimmunoprecipitation of the F protein showed that the efficiency of cleavage of the F0 protein was greatly enhanced by the amino acid replacements. These results demonstrate that genetically modified NDV can be recovered from cloned cDNA and confirm the supposition that cleavage of the F0 protein is a key determinant in virulence of NDV.

Enhanced measles virus cDNA rescue and gene expression after heat shock

Rescue of negative-stranded RNA viruses from full-length genomic cDNA clones is an essential technology for genetic analysis of this class of viruses. Using this technology in our studies of measles virus (MV), we found that the efficiency of the measles virus rescue procedure (F. Radecke et al., EMBO J. 14:5773-5784, 1995) could be improved by modifying the procedure in two ways. First, we found that coculture of transfected 293-3-46 cells with a monolayer of Vero cells increased the number of virus-producing cultures about 20-fold. Second, we determined that heat shock treatment increased the average number of transfected cultures that produced virus another two- to threefold. In addition, heat shock increased the number of plaques produced by positive cultures. The effect of heat shock on rescue led us to test the effect on transient expression from an MV minireplicon. Heat shock increased the level of reporter gene expression when either minireplicon DNA or RNA was used regardless of whether complementation was provided by cotransfection with expression plasmids or infection with MV helper virus. In addition, we found that MV minireplicon gene expression could be stimulated by cotransfection with an Hsp72 expression plasmid, indicating that hsp72 likely plays a role in the effect of heat shock.

Highly diverse intergenic regions of the paramyxovirus simian virus 5 cooperate with the gene end U tract in viral transcription termination and can influence reinitiation at a downstream gene

A dicistronic minigenome containing the M-F gene junction was used to determine the role of the simian virus 5 (SV5) intergenic regions in transcription. The M-F junction differs from the other SV5 junctions by having a short M gene end U tract of only four residues (U4 tract) and a 22-base M-F intergenic sequence between the M gene end and F gene start site. Replacing the 22-base M-F intergenic region with nonviral sequences resulted in a minigenome template (Rep 22) that was defective in termination at the end of the M gene. Efficient M gene termination could be restored to the mutant Rep 22 template in either of two ways: by increasing the U tract length from four to six residues or by restoring a G residue immediately downstream of the wild-type (WT) U4 tract. In a dicistronic SH-HN minigenome, a U4-G combination was functionally equivalent to the naturally occurring SH U6-A gene end in directing SH transcription termination. In addition to affecting termination, the M-F intergenic region also influenced polymerase reinitiation. In the context of the WT U4-G M gene end, substituting nonviral sequences into the M-F intergenic region had a differential effect on F gene reinitiation, where some but not all nonviral sequences inhibited reinitiation. The inhibition of F gene reinitiation correlated with foreign sequences having a high C content. Deleting 6 bases or inserting 18 additional nucleotides into the middle of the 22-base M-F intergenic segment did not influence M gene termination or F gene reinitiation, indicating that M-F intergenic length per se is not a important factor modulating the SV5 polymerase activity. Our results suggest that the sequence diversity at an SV5 gene junction reflects specific combinations which may differentially affect SV5 gene expression and provide an additional level of transcriptional control beyond that which results from the distance of a gene from the 3' end promoter.

Paramyxovirus replication and pathogenesis. Reverse genetics transforms understanding

A recent breakthrough in the field of nonsegmented negative strand RNA viruses (Mononegavirales), including paramyxoviruses, is the establishment of a system to recover an infectious virus entirely from complementary DNA and hence allow reverse genetics. Mutations can now be introduced into viral genomes at will and the resulting phenotypes studied as long as the introduced mutations are not lethal. This technology is being successfully applied to answer outstanding questions regarding the roles of viral components in replication and their contribution to pathogenicity, which are difficult to address using conventional virology. For instance, how the paramyxovirus accessory proteins V and C contribute to actual viral replication and pathogenesis has remained unanswered since their first description more than 20 years ago. Using Sendai virus, which causes fatal pneumonia in mice, it has been shown that the V protein is completely dispensable for viral replication in cell cultures but encodes a luxury function required for pathogenesis in vivo. The Sendai virus C proteins were also defined to be nonessential gene products which greatly contributed to replication both in vitro and in vivo. It is also now possible to design live vaccines by introducing predetermined or plausible attenuating mutations. In addition, the use of paramyxoviruses to express foreign genes has also become feasible. Paramyxovirus reverse genetics is thus renovating our understanding of viral replication and pathogenesis and will further mark an era in recombinant technology for disease prevention and gene therapy.

Nonsegmented negative-strand RNA viruses: genetics and manipulation of viral genomes

Protocols to recover negative-stand RNA viruses entirely from cDNA have been established in recent years, opening up this virus group to the detailed analysis of molecular genetics and virus biology. The unique gene-expression strategy of nonsegmented negative-strand RNA viruses, which involves replication of ribonucleoprotein complexes and sequential synthesis of free mRNAs, has also allowed the use of these viruses to express heterologous sequences. There are advantages in terms of easy manipulation of constructs, high capacity for foreign sequences, genetically stable expression, and the possibility of adjusting expression levels. Fascinating prospects for biomedical applications and transient gene therapy are offered by chimeric virus vectors carrying novel envelope protein genes and targeted to defined host cells.

Identification of mutations contributing to the temperature-sensitive, cold-adapted, and attenuation phenotypes of the live-attenuated cold-passage 45 (cp45) human parainfluenza virus 3 candidate vaccine

The live-attenuated human parainfluenza virus 3 (PIV3) cold-passage 45 (cp45) candidate vaccine was shown previously to be safe, immunogenic, and phenotypically stable in seronegative human infants. Previous findings indicated that each of the three amino acid substitutions in the L polymerase protein of cp45 independently confers the temperature-sensitive (ts) and attenuation (att) phenotypes but not the cold-adaptation (ca) phenotype (29). cp45 contains 12 additional potentially important point mutations in other proteins (N, C, M, F, and hemagglutinin-neuraminidase [HN]) or in cis-acting sequences (the leader region and the transcription gene start [GS] signal of the N gene), and their contribution to these phenotypes was undefined. To further characterize the genetic basis for the ts, ca, and att phenotypes of this promising vaccine candidate, we constructed, using a reverse genetics system, a recombinant cp45 virus that contained all 15 cp45-specific mutations mentioned above, and found that it was essentially indistinguishable from the biologically derived cp45 on the basis of plaque size, level of temperature sensitivity, cold adaptation, level of replication in the upper and lower respiratory tract of hamsters, and ability to protect hamsters from subsequent wild-type PIV3 challenge. We then constructed recombinant viruses containing the cp45 mutations in individual proteins as well as several combinations of mutations. Analysis of these recombinant viruses revealed that multiple cp45 mutations distributed throughout the genome contribute to the ts, ca, and att phenotypes. In addition to the mutations in the L gene, at least one other mutation in the 3' N region (i.e., including the leader, N GS, and N coding changes) contributes to the ts phenotype. A recombinant virus containing all the cp45 mutations except those in L was more ts than cp45, illustrating the complex nature of this phenotype. The ca phenotype of cp45 also is a complex composite phenotype, reflecting contributions of at least three separate genetic elements, namely, mutations within the 3' N region, the L protein, and the C-M-F-HN region. The att phenotype is a composite of both ts and non-ts mutations. Attenuating ts mutations are located in the L protein, and non-ts attenuating mutations are located in the C and F proteins. The presence of multiple ts and non-ts attenuating mutations in cp45 likely contributes to the high level of attenuation and phenotypic stability of this promising vaccine candidate.

Generation of bovine respiratory syncytial virus (BRSV) from cDNA: BRSV NS2 is not essential for virus replication in tissue culture, and the human RSV leader region acts as a functional BRSV genome promoter

In order to generate recombinant bovine respiratory syncytial virus (BRSV), the genome of BRSV strain A51908, variant ATue51908, was cloned as cDNA. We provide here the sequence of the BRSV genome ends and of the entire L gene. This completes the sequence of the BRSV genome, which comprises a total of 15,140 nucleotides. To establish a vaccinia virus-free recovery system, a BHK-derived cell line stably expressing T7 RNA polymerase was generated (BSR T7/5). Recombinant BRSV was reproducibly recovered from cDNA constructs after T7 RNA polymerase-driven expression of antigenome sense RNA and of BRSV N, P, M2, and L proteins from transfected plasmids. Chimeric viruses in which the BRSV leader region was replaced by the human respiratory syncytial virus (HRSV) leader region replicated in cell culture as efficiently as their nonchimeric counterparts, demonstrating that all cis-acting sequences of the HRSV promoter are faithfully recognized by the BRSV polymerase complex. In addition, we report the successful recovery of a BRSV mutant lacking the complete NS2 gene, which encodes a nonstructural protein of unknown function. The NS2-deficient BRSV replicated autonomously and could be passaged, demonstrating that NS2 is not essential for virus replication in cell culture. However, growth of the mutant was considerably slower than and final infectious titers were reduced by a factor of at least 10 compared to wild-type BRSV, indicating that NS2 provides a supporting factor required for full replication capacity.

Virus maturation by budding

Enveloped viruses mature by budding at cellular membranes. It has been generally thought that this process is driven by interactions between the viral transmembrane proteins and the internal virion components (core, capsid, or nucleocapsid). This model was particularly applicable to alphaviruses, which require both spike proteins and a nucleocapsid for budding. However, genetic studies have clearly shown that the retrovirus core protein, i.e., the Gag protein, is able to form enveloped particles by itself. Also, budding of negative-strand RNA viruses (rhabdoviruses, orthomyxoviruses, and paramyxoviruses) seems to be accomplished mainly by internal components, most probably the matrix protein, since the spike proteins are not absolutely required for budding of these viruses either. In contrast, budding of coronavirus particles can occur in the absence of the nucleocapsid and appears to require two membrane proteins only. Biochemical and structural data suggest that the proteins, which play a key role in budding, drive this process by forming a three-dimensional (cage-like) protein lattice at the surface of or within the membrane. Similarly, recent electron microscopic studies revealed that the alphavirus spike proteins are also engaged in extensive lateral interactions, forming a dense protein shell at the outer surface of the viral envelope. On the basis of these data, we propose that the budding of enveloped viruses in general is governed by lateral interactions between peripheral or integral membrane proteins. This new concept also provides answers to the question of how viral and cellular membrane proteins are sorted during budding. In addition, it has implications for the mechanism by which the virion is uncoated during virus entry.

Recombinant human respiratory syncytial virus (RSV) from cDNA and construction of subgroup A and B chimeric RSV

Infectious human respiratory syncytial virus (RSV) was produced from a cDNA clone that contains 15,222 nucleotides of RSV genome derived from the A2 strain of subgroup A. Recovery of infectious RSV from cDNA required cotransfection of only three expression plasmids encoding the nucleoprotein (N), the phosphoprotein (P), and the major polymerase protein (L). Inclusion of the M2-1 plasmid was not required in the transfection reaction and if included did not significantly increase the rescue efficiency. However, a single nucleotide substitution in the RSV leader region (C to G at position 4 in the antigenomic sense), greatly increased the amount of infectious virus recovered from cDNA. A recombinant RSVA2 virus that expresses an additional structural G protein derived from a subgroup B RSV was also obtained. Both A2 and B strain G glycoproteins were expressed in cells infected with the chimeric RSV. A chimeric RSV that expresses a heterologous subgroup antigen in a live attenuated vaccine candidate may be important for prevention of diseases associated with both RSV subgroup A and subgroup B infection.

The paramyxovirus SV5 small hydrophobic (SH) protein is not essential for virus growth in tissue culture cells

The SH gene of the paramyxovirus SV5 is located between the genes for the glycoproteins, fusion protein (F) and hemagglutinin-neuraminidase (HN), and the SH gene encodes a small 44-residue hydrophobic integral membrane protein (SH). The SH protein is expressed in SV5-infected cells and is oriented in membranes with its N terminus in the cytoplasm. To study the function of the SH protein in the SV5 virus life cycle, the SH gene was deleted from the infectious cDNA clone of the SV5 genome. By using the recently developed reverse genetics system for SV5, it was found that an SH-deleted SV5 (rSV5DeltaSH) could be recovered, indicating the SH protein was not essential for virus viability in tissue culture. Analysis of properties of rSV5DeltaSH indicated that lack of expression of SH protein did not alter the expression level of the other virus proteins, the subcellular localization of F and HN, or fusion competency as measured by lipid mixing assays and a new content mixing assay that did not require the use of vaccinia virus. The growth rate, infectivity, and plaque size of rSV5 and rSV5DeltaSH were found to be very similar. Although SH is shown to be a component of purified virions by immunoblotting, examination of purified rSV5DeltaSH by electron microscopy did not show an altered morphology from SV5. Thus in tissue culture cells the lack of the SV5 SH protein does not confer a recognizable phenotype.

Molecular basis for naturally occurring elevated readthrough transcription across the M-F junction of the paramyxovirus SV5

Transcription of the paramyxovirus RNA genome is thought to involve a sequential stop-start mechanism whereby monocistronic mRNAs are produced by polyadenylation and termination of 3' upstream gene followed by reinitiation at the downstream start site. For a number of paramyxoviruses, transcription across the M-F gene junction results in the synthesis of high levels of a dicistronic M-F readthrough RNA. In cells infected with the paramyxovirus SV5, 15% or less of the transcripts from the viral P, M, SH, HN, and L genes were detected as readthrough products with the 3' proximal gene. By contrast, approximately 40% of the SV5 F mRNA was detected as a dicistronic M-F transcript. A comparison of the individual SV5 gene junctions showed that elevated M-F readthrough transcription correlate with the M gene end having the shortest U tract for directing polyadenylation and a gene end sequence that differs from the consensus sequence. We have tested the hypothesis that elevated M-F readthrough transcription results from an inefficient termination signal at the end of the M gene. A reverse genetics system was established whereby SV5 transcription was reconstituted in transfected cells using cDNA-derived polymerase components and dicistronic minigenomes that encoded either the SV5 M-F or the SH-HN gene junction. Chimeric SV5 minigenomes were constructed to contain exchanges of a 10 base gene end sequence and the U tract from the M-F (approximately 40% readthrough) and SH-HN (approximately 15% readthrough) junctions. Northern blot analysis of RNA synthesized from these altered templates showed that, in the context of the M-F intergenic region, increasing the length of the M gene end U tract from four residues to six or eight U residues did not decrease M-F readthrough transcription. In contrast, chimeric minigenomes that contained the 10 base region from the end of the SH gene directed very efficient gene termination and a corresponding decrease in readthrough transcription. Mutational analysis showed that a single G to A substitution located five bases 3' to the M gene U tract was sufficient to convert the M gene end region to an efficient signal for polyadenylation-termination. These results demonstrate a role for the gene end region located immediately 3' to the U tract as a major determinant of transcription termination in the paramyxovirus genome. The possible role of M-F readthrough transcription in the paramyxovirus growth cycle is discussed.

A system for functional analysis of Ebola virus glycoprotein

Ebola virus causes hemorrhagic fever in humans and nonhuman primates, resulting in mortality rates of up to 90%. Studies of this virus have been hampered by its extraordinary pathogenicity, which requires biosafety level 4 containment. To circumvent this problem, we developed a novel complementation system for functional analysis of Ebola virus glycoproteins. It relies on a recombinant vesicular stomatitis virus (VSV) that contains the green fluorescent protein gene instead of the receptor-binding G protein gene (VSVDeltaG*). Herein we show that Ebola Reston virus glycoprotein (ResGP) is efficiently incorporated into VSV particles. This recombinant VSV with integrated ResGP (VSVDeltaG*-ResGP) infected primate cells more efficiently than any of the other mammalian or avian cells examined, in a manner consistent with the host range tropism of Ebola virus, whereas VSVDeltaG* complemented with VSV G protein (VSVDeltaG*-G) efficiently infected the majority of the cells tested. We also tested the utility of this system for investigating the cellular receptors for Ebola virus. Chemical modification of cells to alter their surface proteins markedly reduced their susceptibility to VSVDeltaG*-ResGP but not to VSVDeltaG*-G. These findings suggest that cell surface glycoproteins with N-linked oligosaccharide chains contribute to the entry of Ebola viruses, presumably acting as a specific receptor and/or cofactor for virus entry. Thus, our VSV system should be useful for investigating the functions of glycoproteins from highly pathogenic viruses or those incapable of being cultured in vitro.