Infectious Murine Leukemia Virus Research Articles

The Inhibition of Gag-Pol Expression by the Restriction Factor Shiftless Is Dispensable for the Restriction of HIV-1 Infection.

The interferon-induced host cell protein Shiftless (SFL) inhibits -1 programmed ribosomal frameshifting (-1PRF) required for the expression of HIV-1 Gal-Pol and the formation of infectious HIV-1 particles. However, the specific regions in SFL required for antiviral activity and the mechanism by which SFL inhibits -1PRF remain unclear. Employing alanine scanning mutagenesis, we found that basic amino acids in the predicted zinc ribbon motif of SFL are essential for the suppression of Gag-Pol expression but dispensable for anti-HIV-1 activity. We have shown that SFL inhibits the expression of the murine leukemia virus (MLV) Gag-Pol polyprotein and the formation of infectious MLV particles, although Gag-Pol expression of MLV is independent of -1PRF but requires readthrough of a stop codon. These findings indicate that SFL might inhibit HIV-1 infection by more than one mechanism and that SFL might target programmed translational readthrough as well as -1PRF signals, both of which are regulated by mRNA secondary structure elements.

SERINC5 restricts influenza virus infectivity.

SERINC5 is a multi-span transmembrane protein that is incorporated into HIV-1 particles in producing cells and inhibits HIV-1 entry. Multiple retroviruses like HIV-1, equine infectious anemia virus and murine leukemia virus are subject to SERINC5 inhibition, while HIV-1 pseudotyped with envelope glycoproteins of vesicular stomatitis virus and Ebola virus are resistant to SERINC5. The antiviral spectrum and the underlying mechanisms of SERINC5 restriction are not completely understood. Here we show that SERINC5 inhibits influenza A virus infection by targeting virus-cell membrane fusion at an early step of infection. Further results show that different influenza hemagglutinin (HA) subtypes exhibit diverse sensitivities to SERINC5 restriction. Analysis of the amino acid sequences of influenza HA1 strains indicates that HA glycosylation sites correlate with the sensitivity of influenza HA to SERINC5, and the inhibitory effect of SERINC5 was lost when certain HA glycosylation sites were mutated. Our study not only expands the antiviral spectrum of SERINC5, but also reveals the role of viral envelope glycosylation in resisting SERINC5 restriction.

Dynamics and restriction of murine leukemia virus cores in mitotic and interphase cells

BackgroundMurine leukemia viruses (MLVs) naturally infect unsynchronized T and B lymphocytes, thus, the incoming virus encounters both interphase and mitotic cells. While it is well accepted that MLV requires cell division to complete its replication cycle, it is not known if ab initio infection of mitotic cells can result in productive infection. This question is highly relevant since the milieu of mitotic cells is markedly different from this of interphase cells; e.g. lacking radial microtubule network and intact nuclear envelope. To follow MLV infection in mitotic and interphase cells in real-time, we employed our recently developed infectious MLV particles with labeled cores, cellular models expressing fluorescence markers of different intracellular compartments and protocols for reversible mitotic arrest of MLV-susceptible cells.ResultsMulti-wavelength live cell imaging was employed to simultaneously visualize GFP-labeled MLV cores, DiD-labeled viral or cellular membranes, and fluorescently-labeled microtubules or chromosomes. Cells were imaged either at interphase or upon mitotic arrest with microtubule poisons. Analysis of virus localization and trajectories revealed entry by endocytosis at interphase and mitosis, and correlation between viral mobility parameters and presence or absence of polymerized interphase microtubules. The success of infection of viruses that entered cells in mitosis was evidenced by their ability to reverse transcribe, their targeting to condensed chromosomes in the absence of radial microtubule network, and gene expression upon exit from mitosis. Comparison of infection by N, B or NB -tropic viruses in interphase and mitotic human cells revealed reduced restriction of the N-tropic virus, for infection initiated in mitosis.ConclusionsThe milieu of the mitotic cells supports all necessary requirements for early stages of MLV infection. Such milieu is suboptimal for restriction of N-tropic viruses, most likely by TRIM5α.Electronic supplementary materialThe online version of this article (doi:10.1186/s12977-015-0220-2) contains supplementary material, which is available to authorized users.

Expression of YB-1 enhances production of murine leukemia virus vectors by stabilizing genomic viral RNA

Murine leukemia virus (MLV)-based retroviral vectors is widely used for gene transfer and basic research, and production of high-titer retroviral vectors is very important. Here we report that expression of the Y-box binding protein 1 (YB-1) enhanced the production of infectious MLV vectors. YB-1 specifically increased the stability of viral genomic RNA in virus-producing cells, and thus increasing viral RNA levels in both producer cells and virion particles. The viral element responsive to YB-1 was mapped to the repeat sequence (R region) in MLV genomic RNA. These results identified YB-1 as a MLV mRNA stabilizer, which can be used for improving production of MLV vectors.

Frequent detection of infectious xenotropic murine leukemia virus (XMLV) in human cultures established from mouse xenografts

Purpose: To investigate the frequency of xenotropic murine leukemia virus (MLV) presence in human cell lines established from mouse xenografts and to search for the evidence of horizontal viral spread to other cell lines. Methodology: We examined xenograft tumor cell lines from 7 independent laboratories and 128 non-xenografted tumor cell lines. Cell line DNA was examined for mouse DNA contamination, and by three Taqman qPCR assays targeting the gag, env or pol regions of MLV. Sequencing was used for viral strain identification. Supernatant fluids were tested for reverse transcriptase (RT) activity. Results: Six of 23 (26%) mouse DNA free xenograft cultures were strongly positive for MLV and their sequences had greater than 99% homology to known MLV strains. Four of five available supernatant fluids from these viral positive cultures were strongly positive for RT activity. Three of these supernatant fluids were studied to confirm the infectivity of the released virions for other human culture cells. Of the 78 non-xenograft derived cell lines maintained in the xenograft culture-containing facilities, 13 (17%) were positive for MLV, including XMRV, a virus strain first identified in human tissues. By contrast, all 50 cultures maintained in a xenograft culture-free facility were negative for viral sequences. Conclusions: Human cultures derived after mouse xenografting frequently contain and release highly infectious xenotropic MLV viruses. Laboratories working with xenograft-derived human cultures should be aware of the risk of contamination with potentially biohazardous human-tropic mouse viruses and their horizontal spread to other cultures.

Expression of IMP1 Enhances Production of Murine Leukemia Virus Vector by Facilitating Viral Genomic RNA Packaging

Murine leukemia virus (MLV)-based retroviral vector is widely used for gene transfer. Efficient packaging of the genomic RNA is critical for production of high-titer virus. Here, we report that expression of the insulin-like growth factor II mRNA binding protein 1 (IMP1) enhanced the production of infectious MLV vector. Overexpression of IMP1 increased the stability of viral genomic RNA in virus producer cells and packaging of the RNA into progeny virus in a dose-dependent manner. Downregulation of IMP1 in virus producer cells resulted in reduced production of the retroviral vector. These results indicate that IMP1 plays a role in regulating the packaging of MLV genomic RNA and can be used for improving production of retroviral vectors.

Expression of infectious murine leukemia viruses by RAW264.7 cells, a potential complication for studies with a widely used mouse macrophage cell line

The mouse macrophage-like cell line RAW264.7, the most commonly used mouse macrophage cell line in medical research, was originally reported to be free of replication-competent murine leukemia virus (MuLV) despite its origin in a tumor induced by Abelson MuLV containing Moloney MuLV as helper virus. As currently available, however, we find that it produces significant levels of ecotropic MuLV with the biologic features of the Moloney isolate and also MuLV of the polytropic or MCF class. Newborn mice developed lymphoma following inoculation with the MuLV mixture expressed by these cells. These findings should be considered in interpretation of increasingly widespread use of these cells for propagation of other viruses, studies of biological responses to virus infection and use in RNA interference and cell signalling studies.

Physiological properties of astroglial cell lines derived from mice with high (SAMP8) and low (SAMR1, ICR) levels of endogenous retrovirus

Previous studies have reported that various inbred SAM mouse strains differ markedly with regard to a variety of parameters, such as capacity for learning and memory, life spans and brain histopathology. A potential cause of differences seen in these strains may be based on the fact that some strains have a high concentration of infectious murine leukemia virus (MuLV) in the brain, whereas other strains have little or no virus. To elucidate the effect of a higher titer of endogenous retrovirus in astroglial cells of the brain, we established astroglial cell lines from SAMR1 and SAMP8 mice, which are, respectively, resistant and prone to deficit in learning and memory and shortened life span. MuLV-negative astroglial cell lines established from ICR mice served as controls. Comparison of these cell lines showed differences in: 1) levels of the capsid antigen CAgag in both cell lysates and culture media, 2) expression of genomic retroelements, 3) the number of virus particles, 4) titer of infectious virus, 5) morphology, 6) replication rate of cells in culture and final cell concentrations, 7) expression pattern of proinflammatory cytokine genes. The results show that the expression of MuLV is much higher in SAMP8 than SAMR1 astrocyte cultures and that there are physiological differences in astroglia from the 2 strains. These results raise the possibility that the distinct physiological differences between SAMP8 and SAMR1 are a function of activation of endogenous retrovirus.

Inhibition of infectious murine leukemia virus production by Fv-4 env gene products exerting dominant negative effect on viral envelope glycoprotein

Fv-4 is a mouse gene that confers resistance against ecotropic murine leukemia virus (MLV) infection on mice. While receptor interference by the Fv-4 env gene product, Fv-4 Env, that can bind to the ecotropic MLV receptor has been shown to play an important role in the resistance, other mechanisms have also been suggested because it confers extremely efficient, complete resistance in vivo. Here, we have examined the effect of Fv-4 Env on infectious MLV production. Infectious MLV titers in supernatants obtained after transfection with a Friend MLV (FMLV) Env-expressing plasmid from MLV gag– pol producer cells harboring a retroviral vector were largely reduced by coexpression of Fv-4 Env. Syncytia formation mediated by R-peptide-deleted FMLV Env in NIH 3T3 cells was impaired by Fv-4 Env coexpression. Similarly, Fv-4 Env inhibited infectious amphotropic MLV production and syncytia formation mediated by R-peptide-deleted amphotropic MLV Env. Immunoprecipitation analysis revealed interaction of Fv-4 Env with amphotropic MLV Env as well as FMLV Env. These results indicate that Fv-4 Env inhibits infectious ecotropic and amphotropic MLV production by exerting dominant negative effect on MLV Env, suggesting contribution of this inhibitory effect to the resistance against ecotropic MLV infection in Fv-4-expressing mice.

Detection of reverse transcriptase activity in human melanoma cell lines and identification of a murine leukemia virus contaminant

Stimulated by earlier reports on the presence of retroviruses in mouse and hamster melanoma cell lines, we addressed the question as to whether human melanoma cell lines might also harbour a retrovirus. The melanoma cell lines SK-MEL-25, SK-MEL-28, MEL-JUSO, MML-I, MeWo, A-375, Colo-38, BS-780 were confirmed to be human by human leucocyte antigen (HLA) typing, and supernatants were tested by the product-enhanced reverse transcriptase (PERT) assay for reverse transcriptase (RT) activity. Cell lines SK-MEL-25, SK-MEL-28, MEL-JUSO and MML-I were positive, whereas cell lines MeWo, A-375, Colo-38 and BS-780 were negative. The RT activity peaked at a buoyant density in sucrose typical for retroviruses. From this peak fraction an R-U5 sequence indistinguishable from murine leukemia virus (MLV) was identified by particle-associated retrovirus RNA amplification (PARRA). Semiquantitative MLV-specific RNA-PCR demonstrated colocalization of the MLV-like RNA and RT activity on the sucrose gradient of SK-Mel-25. MLV RNA and DNA were also detectable in culture supernatants of SK-MEL-28, MEL-JUSO and MML-I, but not of MeWo, A-375, Colo-38 and BS-780 by semiquantitative polymerase chain reaction (PCR). Sequence comparison revealed highest homology with the RET sequence previously identified in mouse myeloma SP2/0-AG14 cells. Taken together, our data strongly suggest that certain human melanoma cell lines are productively infected by a MLV which was probably introduced during tumour passage in mice or by laboratory contamination many years ago and subsequently spread to other lines. We recommend mandatory testing of melanoma and other human cell lines for contamination with infectious MLV or other animal retroviruses, similar to mycoplasma screening, in order to avoid artificial experimental data.

Analysis of murine leukemia virus replication complemented by yeast tRNA Phe reveals inherent preferences for the tRNA primer selected for reverse transcription

The replication of murine leukemia virus (MuLV) requires the capture of a cellular tRNA Pro as a primer for reverse transcription. To further study the specificity of primer selection, we have utilized a defective MuLV in which the primer-binding site (PBS) has been altered to be complementary to a nonmammalian tRNA, yeast tRNA Phe. Infectivity of the defective MuLV is dependent upon co-expression of yeast tRNA Phe in the cell. Defective MuLV genomes have been constructed in which the PBS was altered to be complementary to tRNA Phe that also encoded the cDNA for tRNA Phe. Transfection of these defective proviral genomes into cells resulted in the production of infectious MuLV as determined by a single-round assay. The amount of infectious virus produced using this complementation system, though, was approximately 6-fold lower than that produced following transfection of defective proviral genomes with a wild-type PBS complementary to tRNA Pro. The lower infectivity was not due to reduced expression of tRNA Phe in the transfected cells as compared to endogenous tRNA Pro or tRNA Lys,3. Serial passage of the MuLV genome with a PBS complementary to tRNA Phe that encoded tRNA Phe resulted in amplification of the virus. Using this rescue system, we have passaged the virus for four serial passages, after which time a revertant genome in which the PBS was altered to be complementary to tRNA Gln was detected that grew to high titers following subsequent serial passage. The results of these studies suggest that MuLV has preferences for the tRNA primer used in reverse transcription and are discussed with respect to the mechanism of primer selection.

Analysis of murine leukemia virus replication complemented by yeast tRNAPhe reveals inherent preferences for the tRNA primer selected for reverse transcription

The replication of murine leukemia virus (MuLV) requires the capture of a cellular tRNAPro as a primer for reverse transcription. To further study the specificity of primer selection, we have utilized a defective MuLV in which the primer-binding site (PBS) has been altered to be complementary to a nonmammalian tRNA, yeast tRNAPhe. Infectivity of the defective MuLV is dependent upon co-expression of yeast tRNAPhe in the cell. Defective MuLV genomes have been constructed in which the PBS was altered to be complementary to tRNAPhe that also encoded the cDNA for tRNAPhe. Transfection of these defective proviral genomes into cells resulted in the production of infectious MuLV as determined by a single-round assay. The amount of infectious virus produced using this complementation system, though, was approximately 6-fold lower than that produced following transfection of defective proviral genomes with a wild-type PBS complementary to tRNAPro. The lower infectivity was not due to reduced expression of tRNAPhe in the transfected cells as compared to endogenous tRNAPro or tRNALys,3. Serial passage of the MuLV genome with a PBS complementary to tRNAPhe that encoded tRNAPhe resulted in amplification of the virus. Using this rescue system, we have passaged the virus for four serial passages, after which time a revertant genome in which the PBS was altered to be complementary to tRNAGln was detected that grew to high titers following subsequent serial passage. The results of these studies suggest that MuLV has preferences for the tRNA primer used in reverse transcription and are discussed with respect to the mechanism of primer selection.

Generation of novel syncytium-inducing and host range variants of ecotropic moloney murine leukemia virus in Mus spicilegus.

Mus spicilegus is an Eastern European wild mouse species that has previously been reported to harbor an unusual infectious ecotropic murine leukemia virus (MLV) and proviral envelope genes of a novel MLV subgroup. In the present study, M. spicilegus neonates were inoculated with Moloney ecotropic MLV (MoMLV). All 17 inoculated mice produced infectious ecotropic virus after 8 to 14 weeks, and two unusual phenotypes distinguished the isolates from MoMLV. First, most of the M. spicilegus isolates grew to equal titers on M. dunni and SC-1 cells, although MoMLV does not efficiently infect M. dunni cells. The deduced amino acid sequence of a representative clone differed from MoMLV by insertion of two serine residues within the VRA of SUenv. Modification of a molecular clone of MoMLV by the addition of these serines produced a virus that grows to high titer in M. dunni cells, establishing a role for these two serine residues in host range. A second unusual phenotype was found in only one of the M. spicilegus isolates, Spl574. Spl574 produces large syncytia of multinucleated giant cells in M. dunni cells, but its replication is restricted in other mouse cell lines. Sequencing and mutagenesis demonstrated that syncytium formation could be attributed to a single amino acid substitution within VRA, S82F. Thus, viruses with altered growth properties are selected during growth in M. spicilegus. The mutations associated with the host range and syncytium-inducing variants map to a key region of VRA known to govern interactions with the cell surface receptor, suggesting that the associated phenotypes may result from altered interactions with the unusual ecotropic virus mCAT1 receptor carried by M. dunni.

Characterization of producer cell-dependent restriction of murine leukemia virus replication.

We previously reported that the human bronchocarcinoma cell line A549 produces poorly infectious gibbon ape leukemia virus-pseudotyped Moloney murine leukemia virus (MLV). In contrast, similar amounts of virions recovered from human fibrosarcoma HT1080 cells result in 10-fold-higher transduction rates (G. Duisit, A. Salvetti, P. Moullier, and F. Cosset, Hum. Gene Ther. 10:189-200, 1999). We have now extended this initial observation to other type-C envelope (Env) pseudotypes and analyzed the mechanism involved. Structural and morphological analysis showed that viral particles recovered from A549 (A549-MLV) and HT1080 (HT1080-MLV) cells were normal and indistinguishable from each other. They expressed equivalent levels of mature Env proteins and bound similarly to the target cells. Furthermore, incoming particles reached the cytosol and directed the synthesis of linear viral DNA equally efficiently. However, almost no detectable circular DNAs could be detected in A549-MLV-infected cells, indicating that the block of infection resulted from defective nuclear translocation of the preintegration complex. Interestingly, pseudotyping of A549-MLV with vesicular stomatitis virus glycoprotein G restored the amount of circular DNA forms as well as the transduction rates to HT1080-MLV levels, suggesting that the postentry blockage could be overcome by endocytic delivery of the core particles downstream of the restriction point. Thus, in contrast to the previously described target cell-dependent Fv-1 (or Fv1-like) restriction in mammalian cells (P. Pryciak and H. E. Varmus, J. Virol. 66:5959-5966, 1992; G. Towers, M. Bock, S. Martin, Y. Takeuchi, J. P. Stoye, and O. Danos, Proc. Natl. Acad. Sci. USA 97:12295-12299, 2000), we report here a new restriction of MLV replication that relies only on the producer cell type.

Bc3h1 myogenic cells produce an infectious ecotropic murine leukemia virus.

cDNAs representing an endogenous C-type ecotropic murine leukemia virus were isolated from a cDNA library constructed to represent mRNAs present in BC3H1 myogenic cells but not in C2C12 myogenic cells. RNA blot hybridization analysis using the cDNA inserts as probes revealed that BC3H1 cells produce MuLV-related transcirpts of at least three different size classes. A polymerase chain reaction enhanced assay for reverse transcriptase activity revealed the presence of reverse transcriptase in a viral pellet from medium conditioned by BC3H1 cells. A fungizone enhanced assay for syncitium formation provided further evidence of ecotropic retroviral particle production. Exposure of 3T3 cells to medium conditioned by BC3H1 cells, using conditions that facilitate infection, resulted in infection of the 3T3 cells, as confirmed by the syncitium formation assay. We conclude that BC3H1 cells produce an infectious ecotropic murine leukemia virus. Whether or not this feature of BC3H1 cells contributes to their inability to express some muscle-specific genes or to carry out myotube formation is unknown. Investigators will want to take into account that BC3H1 cells are virus producers when planning experiments that involve coculture of BC3H1 with other cell types, BC3H1 conditioned medium, retrovirally mediated transfection into BC3H1 cells, or study of the mCAT-1 amino acid transporter (the viral receptor) in BC3H1 cells. BC3H1 cells and the virus they produce may be of interest to those studying retroviral genomes and products and their effects.

Capsid-Targeted Viral Inactivation Can Eliminate the Production of Infectious Murine Leukemia Virus in Vitro

Capsid-targeted viral inactivation (CTVI), a promising gene-based antiviral strategy against retroviruses, was designed to disrupt the retroviral life cycle by incorporating a degradative enzyme (e.g., nuclease) into viral particles during assembly, thereby reducing or eliminating the production of infectious virus. The experimental system used to develop the CTVI strategy for retroviruses is designed to block the production of infectious Moloney murine leukemia virus (Mo-MLV). Two nucleases, Escherichia coli ribonulease HI and Staphylococcus nuclease, have been shown to be tolerated by the cell as Mo-MLV Gag-nuclease fusion polyproteins and still be active in the viral particles. The goal of this study was to determine what cellular and viral factors limit CTVI in cultured cells. The avian DF-1 cell line greatly expanded our ability to test the antiviral efficacy of CTVI in long-term assays and to determine the mechanism(s) of CTVI action. The CTVI antiviral effect is dependent on the level of Mo-MLV Gag-nuclease fusion polyprotein expressed. The Mo-MLV Gag-nuclease polyproteins produce a long-term prophylactic antiviral effect after a low- or high-dose Mo-MLV challenge. The Mo-MLV Gag-nuclease fusions have a significant therapeutic effect (∼1000-fold) on the production of infectious Mo-MLV. The therapeutic CTVI effect can be improved by a second delivery of the CTVI fusion gene. Both the prophylactic and the therapeutic CTVI antiviral approaches can virtually eliminate the production of infectious Mo-MLV in vitro and are only limited by the number of cells in the population that do not express adequate levels of the CTVI fusion polyprotein.

Initial binding of murine leukemia virus particles to cells does not require specific Env-receptor interaction.

The initial step of virus-cell interaction was studied by immunofluorescence microscopy. Single particles of murine leukemia virus (MLV) vectors and human immunodeficiency virus (HIV) were visualized by immunofluorescence. Fluorescent dots representing single virions could be localized by staining of capsid proteins (CA) or surface envelope proteins (SU) after fixation of virus supernatants. This technique can be used to determine particle concentration in viral supernatants and also to study virus-cell interaction. We investigated the role of the Env-receptor interaction for the initial binding event between the cell and the viral particles. Ecotropic MLV vector particles were shown to bind to human cells which do not express the specific viral receptor. In addition, MLV particles defective for Env were shown to bind the cells similarly to infectious MLV. Time course experiments of virus-cell binding and dissociation showed identical profiles for infectious and Env-defective MLV particles and suggested that MLV Env is not involved in the early phases of attachment of virus to cells. The possible implication of cellular factors in enhancing viral binding and infectivity is discussed.

Effect of substrate residues on the P2' preference of retroviral proteinases.

The substrate sequence requirements for preference toward P2' Glu residue by human immunodeficiency virus type 1 (HIV-1) proteinase were studied in both the matrix protein/ capsid protein (MA/CA) and CA/p2 cleavage site sequence contexts. These sequences represent typical type 1 (-aromatic*Pro-) and type 2 (-hydrophobic* hydrophobic-) cleavage site sequences, respectively. While in the type 1 sequence context, the preference for P2' Glu over Ile or Gln was found to be strongly dependent on the ionic strength and the residues being outside the P2-P2' region of the substrate, it remained preferable in the type 2 substrates when typical type 1 substrate sequence residues were substituted into the outside regions. The pH profile of the specificity constants suggested a lower pH optimum for substrates having P2' Glu in contrast to those having uncharged residues, in both sequence contexts. The very low frequency of P2' Glu in naturally occurring retroviral cleavage sites of various retroviruses including equine infectious anemia virus (EIAV) and murine leukemia virus (MuLV) suggests that such a residue may not have a general regulatory role in the retroviral life cycle. In fact, unlike HIV-1 and HIV-2, EIAV and MuLV proteinases do not favor P2' Glu in either the MA/CA or CA/p2 sequence contexts.

Envelope and long terminal repeat sequences of an infectious murine leukemia virus from a human SCLC cell line: implications for gene transfer.

Development of methods for gene transfer into specific cell types or tissues is important for experimental research as well as clinical therapeutical approaches. We report here the cloning and characterization of the envelope (env) gene and the U3 region of a retrovirus from an infected human Small Cell Lung Cancer (SCLC) cell line. The replication of this murine retrovirus is also fully supported by other lung cancer cell lines of different histological origin. We present evidence that a long terminal repeat (LTR)-beta-galactosidase (beta-Gal) reporter construct performed as well as an analogous cytomegalovirus (CMV) promoter beta-Gal construct in the human lung epithelial cell line A549 and in the human larynx carcinoma cell line HEp2.

Production of infectious recombinant Moloney murine leukemia virus particles in BHK cells using Semliki Forest virus-derived RNA expression vectors.

We describe a heterologous, Semliki Forest virus (SFV)-driven packaging system for the production of infectious recombinant Moloney murine leukemia virus particles. The gag-pol and env genes, as well as a recombinant retrovirus genome (LTR-psi (+)-neoR-LTR), were inserted into individual SFV1 expression plasmids. Replication-competent RNAs were transcribed in vitro and introduced into the cytoplasm of BHK-21 cells using electroporation. The expressed Moloney murine leukemia virus structural proteins produced extracellular virus-like particles. In these particles the gag precursor was processed into mature products, indicating that the particles contained an active protease. The protease of the gag-pol fusion protein was also shown to be active in a trans-complementation assay using a large excess of Pr65gag. Moreover, the particles possessed reverse transcriptase (RT) activity as measured in an in vitro assay. Cotransfection of BHK-21 cells by all three SFV1 constructs resulted in the production of transduction-competent particles at 4 x 10(6) colony-forming units (cfu)/ml during a 5-hr incubation period. Altogether, 2.9 x 10(7) transduction-competent particles were obtained from about 4 x 10(6) transfected cells. Thus, this system represents the first RNA-based packaging system for the production of infectious retroviral particles. The facts that no helper virus could be detected in the virus stocks and that particles carrying the amphotropic envelope could be produced with similar efficiency as those that carry the ecotropic envelope make the system very interesting for gene therapy.