Molecular Events Accompanying Rous Sarcoma Virus Rescue from Rodent Cells and the Role of Viral Gene Complementation

ABSTRACT

Transformation of rodent cells with avian Rous sarcoma virus (RSV) opened new ways to studying virus integration and expression in nonpermissive cells. We were interested in (i) the molecular changes accompanying fusion of RSV-transformed mammalian cells with avian cells leading to virus rescue and (ii) enhancement of this process by retroviral gene products. The RSV-transformed hamster RSCh cell line was characterized as producing only a marginal amount of env mRNA, no envelope glycoprotein, and a small amount of unprocessed Gag protein. Egress of viral unspliced genomic RNA from the nucleus was hampered, and its stability decreased. Cell fusion of the chicken DF-1 cell line with RSCh cells led to production of env mRNA, envelope glycoprotein, and processed Gag and virus-like particle formation. Proteosynthesis inhibition in DF-1 cells suppressed steps leading to virus rescue. Furthermore, new aberrantly spliced env mRNA species were found in the RSCh cells. Finally, we demonstrated that virus rescue efficiency can be significantly increased by complementation with the env gene and the highly expressed gag gene and can be increased the most by a helper virus infection. In summary, Env and Gag synthesis is increased after RSV-transformed hamster cell fusion with chicken fibroblasts, and both proteins provided in trans enhance RSV rescue. We conclude that the chicken fibroblast yields some factor(s) needed for RSV replication, particularly Env and Gag synthesis, in nonpermissive rodent cells.

IMPORTANCE One of the important issues in retrovirus heterotransmission is related to cellular factors that prevent virus replication. Rous sarcoma virus (RSV), a member of the avian sarcoma and leukosis family of retroviruses, is able to infect and transform mammalian cells; however, such transformed cells do not produce infectious virus particles. Using the well-defined model of RSV-transformed rodent cells, we established that the lack of virus replication is due to the absence of chicken factor(s), which can be supplemented by cell fusion. Cell fusion with permissive chicken cells led to an increase in RNA splicing and nuclear export of specific viral mRNAs, as well as synthesis of respective viral proteins and production of virus-like particles. RSV rescue by cell fusion can be potentiated by in trans expression of viral genes in chicken cells. We conclude that rodent cells lack some chicken factor(s) required for proper viral RNA processing and viral protein synthesis.

INTRODUCTION

Retrovirus functions have been systematically studied by delineation of the retroviral genome structure and its individual genes and functional domains. However, it turned out that the host cell can alter expression of such genes and domains. Cellular factors may act in a dominant-negative way, efficiently suppressing viral functions in different steps of the virus replication cycle. Such factors have been isolated and characterized (1–3). The cell can also keep virus infection in check by the lack of cell functions required for virus replication. In such a case, it is more demanding to characterize the set of functions involved. One of the first models for the latter situation was provided by some mammalian cell lines transformed with avian Rous sarcoma virus (RSV) strains. These cell lines (designated originally as virogenic) harbor the integrated retrovirus genome indefinitely in every tested clonal cell population as integrated provirus (4). However, the viral genome is not fully expressed, and infectious virus production is not detectable. Such RSV-transformed cells can be forced to produce virus by cell fusion with permissive chicken fibroblasts (5), which was confirmed and extended (6–9). The RSV rescue studies also promoted HIV rescue experiments, which showed that despite adjusting rodent cells to early steps of HIV infection, these cells remained largely nonpermissive with regard to infectious virus production. However, infectious HIV synthesis was triggered when such cells were fused with permissive human cells (10–12). This indicated that permissive cells provided some function missing in nonpermissive cells that needs to be present in order to ensure full virus genome expression.

Despite that the cytological parameters of virus rescue have been clearly established and confirmed (5, 7, 13), we still lack molecular insight into this process. For our study, we employed the RSCh line of Chinese hamster fibroblasts transformed in vitro with the Schmidt-Ruppin RSV strain (SR-RSV), whose cytogenetic profile has been studied at regular intervals before, during, and after transformation (13). This cell line has also been thoroughly tested for the absence of any infectious RSV production and has been employed in quantitative virus-cell fusion experiments (5).

We show here that envelope (env) mRNA splicing, unspliced genomic RNA (gRNA) nuclear export, and Env and Gag protein formation occur after fusion of nonpermissive RSCh cells with chicken fibroblasts. We have demonstrated that cooperation of these molecular events is required for RSV rescue from nonpermissive cells. We were also able to compare the SR-RSV splicing pattern with the previously tested Prague RSV (PR-RSV) strain in mammalian cells (14), and we found additional aberrant env mRNA splice variants. Furthermore, we have documented that virus rescue efficiency can be increased by complementation via cell fusion with Env- or Gag-producing cells. However, the best results were achieved with chicken cells preinfected with avian leukosis virus (ALV) helper virus. These results are discussed in relation to the general problem of cell factor involvement in infectious retrovirus formation.

MATERIALS AND METHODS

Cell cultures.

RSCh is a Chinese hamster tumor cell line transformed with the Schmidt-Ruppin strain of RSV (SR-RSV-D). H-20 is a Syrian hamster cell line derived from a tumor induced by the Prague strain of RSV carrying only one provirus copy per genome (15). The DF-1 chicken cell line free of alpha endogenous retroviral (ev) loci was obtained from S. Hughes. Brown Leghorn (BL) chickens were selected for their sensitivity to RSV by J. C. Carr, who also kindly provided them to our laboratory. Avian leukosis virus (ALV) infection in this close breed was eliminated. Chicken embryo fibroblasts (CEFs) from these chickens were prepared by standard procedures from 10-day-old embryos and are denoted CEF-BL. Japanese quail embryo fibroblasts (QEFs) were prepared from 8-day-old embryos. 16Q is the QEF cell line transformed by the Bryan RSV strain lacking the envelope gene (BH-RSV) and was developed by H. Murphy (16). It represents a versatile tool for ALV env detection.

Cell treatment and transfection.

Cells were grown in 1:1 Dulbecco's modified Eagle's medium (DMEM) and F-12 medium (Life Technologies) supplemented with l-glutamine, 5% calf serum, 1 to 5% fetal calf serum, 1% chicken serum, and 10% tryptose phosphate broth (Life Technologies).

The cell suspension was X-irradiated with 100 Gy using Wolf-Medizin Technik RTG T-200. No replicated cells survived as measured by control seeding of X-irradiated cells, which were monitored for 4 weeks. Mitomycin C (Sigma) was applied for 2 h in the amount of 10 μg/ml medium. Cell cultures were then washed and left for 2 h in mitomycin-free medium.

Two antibiotics, cycloheximide (Calbiochem) and puromycin (Sigma), were used for proteosynthesis inhibition. Cycloheximide was applied for 24 h and puromycin for 15 h, both at a concentration of 10 μg/ml.

Actinomycin D (Sigma) was employed for transcription inhibition in a concentration of 1 μg per ml.

Transfection was performed using the Polyplus transfection protocol. Cells were exposed to Optimem (Gibco) 1 h before and during transfection (2 ml per 60-mm dish), and then 2 to 4 μg DNA per 60-mm dish was added. A plasmid containing the vesicular stomatitis virus G glycoprotein region (pVSV-G) was obtained from Clontech. The pcGagPol construct was described previously (17).

Cell fusion and infectious center assay.

For cell fusion, we followed the standard polyethylene glycol (PEG) procedure using BioUltra PEG 6000 (Fluka). Briefly, irradiated or mitomycin-treated cells were mixed, and 12 h after plating, the cells were treated with 3.5 ml of 50% PEG for 45 s and rinsed three times with DMEM without serum. Under our experimental conditions, polynuclear cell formation varied between 27% and 30% of total cell counts after fusion, as revealed by Hoechst 33342 (5 μM for 20 min) nuclear staining and phase-contrast microscopy.

The infectious center assay was performed as described previously (5). Proliferating foci were counted within 3 weeks, and virus titers were expressed as focus-forming units (FFU) per ml of medium.

Cell fractionation and RNA extraction.

Cell fractionation was performed according to reference 18. Cells were detached and washed in phosphate-buffered saline (PBS). The pellet was resuspended in 450 μl buffer RLN (50 mM Tris-HCl [pH 8.0], 140 mM NaCl, 1.5 mM MgCl₂, 0.5% Nonidet P-40 substitute, 1,000 U/ml RNase inhibitor [Promega], 1 mM dithiothreitol) and incubated for 5 min on ice. Debris and nuclei were pelleted (300 × g, 2 min, 4°C), and the cytoplasmic fraction was transferred to 1 ml of RNAzol (MRC). For RNA isolation from cell nuclei, the nuclear pellet was washed in 500 μl PBS, again pelleted for 3 min at 300 × g, and then resuspended in 1 ml of RNAzol. RNA was extracted from cultured cells and cell fractions with RNAzol according to the manufacturer's protocol.

qPCR and RT–semi-qPCR.

RNA samples were treated with DNase I (Roche) before reverse transcription-PCR (RT-PCR) to remove any contaminant DNA for 15 min in the Moloney murine leukemia virus (MMLV) reverse transcriptase buffer (Promega). One microgram of RNA was reversely transcribed into cDNA using random hexanucleotides and MMLV reverse transcriptase (Promega) according to the manufacturer's protocol.

One microliter of cDNA was used for the quantitative PCR (qPCR) based on the Mesa green qPCR MasterMix Plus for SYBR assay kit (Eurogentec) and a CFX96 system for qPCR detection (Bio-Rad). Quantifications of viral transcripts were performed with primers RSV_fw (CTTAGGAGGGCAGAAGCTGA) and unspliced_rv (GTTTTACACGCGGACGAAAT) for gRNA, RSV_fw and src_rv (GAGGCCACCAGCAGAGTC) for src mRNA, RSV_fw and envD_rv (TCGGAAATAGGAGACGGGATA) for env mRNA, dbl2_fw (GCCAGGGAACCTTTGGATTA) and src_rv for double-spliced env 2 plus cryptic spliced 2 (dbl2+cryptic2) mRNA, polend_fw (CAGCTGTGAAAAACAGGGACA) and src2_rv (GGGGTCCTTAGGCTTGCTC) for cryptic spliced 1 plus 2 mRNA, and dbl3_fw (TGCTTCTAACTCCACGGAACC) plus src_rv for double-spliced env 3 plus cryptic spliced 3 (dbl3+cryptic3) mRNA. Transcripts were normalized either to total viral RNA with primers srcdel_fw (GACTGAGCTGACCACCAAGG) and srcdel_rv (GCACTACTCAGCGACCTCCA) (results presented as percentages of total viral RNA), to the mammalian glyceraldehyde-3-phosphate dehydrogenase (mGAPDH) housekeeping gene transcripts with primers mGAPDH_fw (AACTTTGGCATTGTGGAAGG) and mGAPDH_rv (ATCCACAGTCTTCTGGGTGG), or to chicken glyceraldehyde-3-phosphate dehydrogenase (chGAPDH) with primers chGAPDH_fw (CATCGTGCACCACCAACTG) and chGAPDH_rv (CGCTGGGATGATGTTCTGG). The volume of the reaction mixture was 20 μl with a 400 nM final concentration of each primer. The external standards for gRNA, env, src, dbl2+cryptic2, cryptic2, dbl3+cryptic3, total viral RNA, and mGAPDH were constructed by PCR using RSCh cDNA and transcript-specific primer sets. The Resulting PCR fragments were cloned into pGEM-T Easy (Promega) and verified by sequencing. Calibration curves were prepared by PCR of diluted plasmid samples ranging from 10² to 10⁷ copies per reaction. Subcellular fractionation was also controlled with SYBR green-based RT-PCR, using primers preGAPDH_fw (TCATCCTGCCAGCAGTGG) and preGAPDH_rv (CAGAGGCAGGGAGTGAGGTC), homologous to intron 5 of the unprocessed pre-mRNA of the gene (pre-GAPDH RNA). One microgram each of extracted nuclear and cytoplasmic RNAs was subjected to RT-qPCR. This allowed calculation of the percentage of pre-GAPDH RNA in the cytoplasm relative to the nuclear pre-GAPDH RNA level. The cycling conditions are as follows: for gRNA, src, dbl2, dbl3, total viral RNA, and GAPDH, 5 min at 95°C followed by 40 cycles of 15 s at 95°C, 20 s at 61°C, and 15 s at 72°C; for env, 5 min at 95°C followed by 40 cycles of 15 s at 95°C, 20 s at 59°C, and 30 s at 72°C; for cryptic1 + 2, 5 min at 95°C followed by 40 cycles of 15 s at 95°C, 20 s at 62°C, and 40 s at 72°C; and for pre-GAPDH, 5 min at 95°C followed by 40 cycles of 15 s at 95°C, 20 s at 63°C, and 15 s at 72°C. We used water and samples without reverse transcriptase as negative controls.

For semiquantitative analysis of endogenous viral mRNAs, cDNA was used as a template, and primers RSV_fw and unspl_rv for the gag region and chENV233fwd (ACGGATTTCTGCCTCTCTACACA) and chENV1046rev (TTCCTTGCCATGCGCGATCCC) for the env region were employed. For identification of aberrant splicing of SR-RSV, we used two pairs of primers: (i) RSV_fw and src_rv and (ii) polend_fw and src2_rv.

Immunofluorescence.

For Env product detection, we employed immunoadhesin Tvb-mIgG. The tvb^S1 gene was equipped with the mouse IgG (mIgG) gene (gift from M. J. Federspiel), which was incorporated into DF-1 cells, where it produced the respective soluble immunoadhesin (19). The Gag product was detected by mouse anti-p27 monoclonal antibody (gift from E. Humphries, West Virginia University [unpublished]). Immunofluorescence microscopy was performed as described previously (20). Briefly, cells grown on coverslips were fixed for 20 min in 3% formaldehyde. Fixed cells were incubated for 60 min with Tvb-mIgG supernatant diluted 1:5 or anti-p27 antibody diluted 1:500. Binding of immunoadhesin Tvb-mIgG or anti-p27 antibody with their respective env or gag gene product was revealed by staining with Cy3-conjugated anti-mouse antibody (dilution, 1:1,000) (Jackson ImmunoResearch Laboratories, West Grove, PA). The preparations were mounted in Mowiol 4-88 (Calbiochem, San Diego, CA) supplemented with 4′,6-diamidino-2-phenylindole (DAPI) (Sigma) to label cell nuclei and examined with an Olympus AX-70 Provis microscope.

Western blot.

Cells and cell-free supernatants were harvested 4 days after fusion. The cells were lysed in SB buffer (2% sodium dodecyl sulfate [SDS], 10% glycerol, 0.625 M Tris-Cl [pH 6.8], 0.1% bromophenol blue, 5% β-mercaptoethanol). The viral particles were pelleted from the cell-free supernatant by ultracentrifugation through 25% sucrose cushion at 32,000 rpm for 1.5 h in a Beckman SW41 Ti rotor. The viral particles were lysed in SB buffer. The lysate of cells and viral particles was subjected to 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and separated protein bands were transferred to polyvinylidene difluoride (PVDF) membrane (Amersham). A rabbit anti-p27 IgG fraction conjugated to horseradish peroxidase (Charles River) was used at a 1:750 dilution to examine the Gag polyprotein in the Western blot assay. LumiGLO chemiluminescent substrate (Cell Signaling) was used to reveal the protein bands of unprocessed and processed Gag polyprotein. Equal protein loading and transfer were verified by immunodetection of GAPDH (dilution, 1:4,000; Invitrogen) on the same membrane.

RESULTS

Reevaluation of RSV-transformed rodent cells and selected chicken cells used for virus rescue.

In order to verify our stock of RSCh cells, we performed karyological analysis. The cells contained typical Chinese hamster chromosomes, and counting their numbers revealed that the highest representation of karyotypes corresponded to pseudodiploid counts—22 chromosomes (data not given), which agrees with our previous observation that repeatedly passaged RSCh cells show a clear tendency to acquire an aneuploid karyotype (13). As has been documented (21), RSCh cells harbor fully functioning provirus that can be transmitted to CEF-BL by DNA transfection, which historically represents the second independent and successful transfection experiment. Using qPCR for comparison of RSCh with H-20 cells carrying only one provirus copy (15), we found that RSCh cells contain an almost equal amount of proviral DNA (Fig. 1A). By terminal dilution, several monocellular clones were established, and one of them was employed in further experiments.

FIG 1.

Characterization of hamster and avian cells used for virus rescue. (A) The number of RSV proviral copies in hamster RSCh cells was compared with that in H-20 cells. Chromosomal DNA was isolated from RSCh and H-20 cells. Twenty-five nanograms of DNA was subjected to qPCR, and using virus-specific primers (SrcDel_fw and SrcDel_rv), quantification cycle (C_q) values were measured and compared. Results represent the means from triplicate samples ± standard deviations. (B and C) Expression of retroviral mRNAs in DF-1 and CEF-BL cells was analyzed by RT-PCR. For env mRNA detection (B), we employed primers designed to detect the conservative env region (chENV233fwd and chENV1046rev), and for gRNA (C), we used primers covering the leader and left part of the gag gene (RSV_fw and unspliced_rv).

For virus rescue experiments, we employed a stable line of chicken fibroblasts, DF-1, to eliminate any effect of endogenous genes. The previously used CEF-BL primary chicken fibroblasts contain four endogenous retroviral loci (ev) (22), while DF-1 cells are regarded as free of ev loci (23). Because we wanted to exclude the possibility that some viral protein synthesis activity remained undetected, we performed an additional control using RT-PCR with primers covering the conservative env region (24) and gag leader. As documented in Fig. 1B, in contrast to CEF-BL, DF-1 cells produced no expected env fragment. Similarly, no gag gene transcript was detected in DF-1 cells (Fig. 1C).

We retested the virus rescue capacity of RSCh cells by fusion with chicken CEF-BL cells using PEG. Then we established that cell line DF-1 can substitute for the primary chicken fibroblast CEF-BL used so far for virus rescue. RSCh cells, either mitomycin treated or X-irradiated, fused with DF-1 cells after 4 days produced 15 to 18 FFU/ml of virus in the culture medium (filtered or unfiltered), in contrast to untreated, treated, or self-fused RSCh cells, which produced no infectious virus. In the second step, we successfully isolated three foci and subcultured them on CEF-BL, and supernatants were tested in the focus assay 7 days later. All cultures produced 10⁴ to 10⁵ FFU/ml of virus.

Viral RNA splicing and synthesis in nonpermissive RSCh cells and fused cells.

As we learned from the past experience, cell fusion between nonpermissive RSV-transformed rodent cells and permissive chicken fibroblasts provides an efficient way to rescue infectious virus. Therefore, we were interested in the course of the events accompanying this process of virus rescue. Because of our previous finding that in nonpermissive RSV-transformed rodent cells env mRNA was undetectable by Northern blotting (15) and gRNA signals were far less intensive than src mRNA signal, we focused on RNA quantification using RT-qPCR. In order to measure individual viral RNA species, we designed the set of primers schematically represented in Fig. 2A. They allowed us to quantify different viral mRNA species.

FIG 2.

Comparison of levels of different viral RNAs' synthesis in RSV-transformed hamster RSCh cells, chicken DF-1 cells infected with SR-RSV, and RSCh cells fused with DF-1 cells. (A) A schematic summary of the SR-RSV-D genome and its transcripts is depicted. Different viral RNA species are shown together with the positions of primers for splicing-specific qPCR. LTR, long terminal repeat. (B) The amounts of viral RNAs were quantified. DF-1 cells were infected by SR-RSV-D. RSCh cells were exposed before fusion to the genotoxic agent mitomycin (mit) or X-irradiation. The fourth day after fusion, total RNA was isolated. cDNAs synthesized by RT from the RNA samples were measured by qPCR. The amounts of viral mRNAs are given as percentages of total viral RNA. The respective primers that were employed are indicated above. Error bars represent standard deviations. Results are from three independent experiments, each in two parallel dishes. The significance of differences between RSCh cells, RSCh cells exposed to mitomycin (mit) or X ray, and RSCh cells exposed to mitomycin or X ray and fused with chicken DF-1 cells are marked by asterisks (*, P = 0.01 to 0.001; **, P < 0.001).

As shown in Fig. 2B, control chicken DF-1 cells infected with SR-RSV-D synthesized mainly gRNA (62%), followed by env (23%) and src (15%) mRNAs. Contrary to that, the main RSCh viral RNA species was represented by src mRNA (66%), followed by a smaller amount of gRNA (20%), whereas the amount of env mRNA fell almost beyond detection. The situation changed after RSCh fusion with DF-1, when the env transcript was clearly detected, representing about 7.5% of total viral RNA. The src mRNA level was significantly decreased, and the gRNA amount did not change after fusion. Mitomycin treatment or X-irradiation used to stop replication after fusion did not change the splicing pattern of regularly spliced viral mRNAs.

New aberrantly spliced viral mRNA species in RSCh cells.

When we quantified env mRNA in RSCh cells, we encountered different expression levels, depending on the primer position. When the reverse primer was positioned close to the env splice acceptor, the amount of env mRNA was larger. This indicated that aberrant splicing using cryptic splicing sites might have been involved. Because aberrant splicing of the PR-RSV strain has been described (25), we decided to find out whether SR-RSV is processed in the same way. Using two pairs of primers, we identified five aberrantly spliced viral mRNA species—cryptic double-spliced env mRNAs 1 and 2 (Dbl1 and Dbl2) and cryptic single-spliced mRNAs 1, 2 and 3 (Fig. 3A and B). Aberrantly spliced RNAs were separated, extracted from the gel, and sequenced. Three different cryptic splicing donor (SD) sites were identified in the 5′ half of the env gene (Fig. 3A); thus, splicing products coding for truncated proteins. Dbl1 and Dbl2 arose by normal splicing from the SD site in gag to the splicing acceptor (SA) site upstream of the env gene and by aberrant splicing from cryptic SD sites to the SA between the env and src genes. Cryptic spliced 1, 2, and 3 RNAs differ from Dbl by the absence of gag-env splicing. Their primary nucleotide sequence fit with their position given in Fig. 3A, with the exception of Dbl3, which was not detected by PCR, and therefore its structure is only putative. Furthermore, cryptic SD sites utilized by splice variants were evaluated by their capability to form H bonds with U1 snRNA (26). The H-bond scores of cryptic sites 1, 2, and 3 were found to be 15, 16.9, and 13.8, respectively (http://www.uni-duesseldorf.de/rna/html/hbond-score.php), which corresponds to 14 hydrogen bonds found as a proper consensus. In addition, aberrantly spliced RNA representation is in accordance with their H-bond scores (Fig. 3B and C). The amount of aberrantly spliced viral RNAs measured by RT-qPCR in RSCh cells exceeded that in chicken cells, reaching 5.6% of total viral RNA in the former and 1.4% in the latter (Fig. 3C). We point out that we used the RSCh cell line in our study, which was shown to contain only one proviral copy, and therefore cryptic splicing cannot result from additional aberrant provirus transcription.

FIG 3.

Identification of cryptic spliced viral mRNA species in RSV-transformed hamster RSCh cells, chicken DF-1 cells infected with SR-RSV, and RSCh cells fused with DF-1 cells. (A) A schematic representation of RSV proviral DNA with depicted splicing sites (splicing donor site [SD] or splicing acceptor site [SA]) and the structure of newly defined aberrantly spliced viral mRNAs is shown. Asterisks indicate cryptic splicing donor sites 1, 2, and 3 at nucleotides (nt) 5222, 5348, and 5805 according to GenBank accession no. D10652.1. Positions of PCR primers and product sizes are shown. The positions of primers used for qPCR are depicted below. Dbl3 was not detected, and its structure is only putative. (B) RSCh cells were exposed to mitomycin before fusion. Total RNA was isolated on the fourth day after fusion and then subjected to RT-PCR. Amplified PCR products obtained using the described primer sets were visualized in ethidium bromide-stained agarose gels. Size markers in base pairs are indicated on the left. (C) The amounts of cryptic viral mRNAs were quantified. DF-1 cells were infected by SR-RSV-D. RSCh cells were exposed before fusion to genotoxic agent mitomycin or X-irradiation. The fourth day after fusion, total RNA was isolated. cDNAs synthesized by RT from the RNA samples were measured by qPCR. The amounts of viral mRNAs are given as percentages of total viral RNA. The respective primers that were employed are indicated above. Error bars represent standard deviations. Results are from four independent experiments, each in two parallel dishes. Significant differences between RSCh cells, RSCh cells exposed to mitomycin (mit) or X ray, and RSCh cells exposed to mitomycin or X ray and fused with chicken DF-1 cells are marked by asterisks (*, P = 0.01 to 0.001; **, P < 0.001).

For the cell fusion experiment, we employed cells exposed to genotoxic treatments such as mitomycin C and X-irradiation in order to prevent cell replication after fusion. This was not without consequences for splicing. Aberrant splicing, especially that of Dbl2+Cryptic2, was highly increased (Fig. 3C). This is not surprising in the light of other studies performed on other models (27, 28).

In summary, we identified new cryptic SD sites in SR-RSV-D env and observed the respective aberrantly spliced mRNA species in RSCh cells. The level of aberrant splicing significantly increased after the genotoxic treatment, but it changed only moderately after fusion with chicken cells.

Rescue of env mRNA and envelope glycoprotein after fusion with permissive cells.

Because comparison of viral mRNA species in RSV-transformed hamster cells and in RSV-transformed hamster cells fused with chicken cells revealed a striking difference in the env mRNA levels, which increased 7-fold on the fourth day after fusion (Fig. 2B), we decided to study env mRNA and the respective protein synthesis in more detail. First, we analyzed the time course of viral envelope glycoprotein synthesis using the immunoadhesin Tvb/d receptor equipped with a mouse IgG (mIgG) domain, detecting the respective glycoprotein. Figure 4 demonstrates multinuclear cells exhibiting fluorescence on the fourth day after fusion. The positive signal was achieved in 30% of polykaryons, whereas cells containing single nuclei were negative. The kinetics of envelope glycoprotein formation after fusion is given in Fig. 5A. There is a good correlation between the increase in proportion of positively stained multinuclear cells (bar graph) and env mRNA synthesis (line graph), as assayed by RT-qPCR.

FIG 4.

RSCh × DF-1 polykaryons producing envelope glycoprotein. The 1:1 mixture of mitomycin (mit)-treated RSCh and DF-1 cells was fused. Cells were fixed on the fourth day after fusion, and Env glycoprotein was detected by immunoadhesin Tvb-mIgG and visualized by goat anti-mouse antibody labeled with Cy3 (red). Nuclear DNA was stained with DAPI (blue). Polykaryons producing envelope glycoprotein are marked by arrows.

FIG 5.

Expression of env mRNA depending on the time after fusion and the presence of proteosynthesis inhibitors in partner cells. (A) The levels of expression of envelope glycoprotein, env mRNA, and three other viral mRNAs were determined during 4 days after the fusion. The line graph shows the levels of four different viral mRNAs (gRNA, src mRNA, env mRNA and dbl2+cryptic2 mRNA) as measured by RT-qPCR. Ratios of individual viral mRNAs to total viral RNA relative to RSCh cells treated with mitomycin (mit) were calculated. Error bars indicate standard deviations. Results are from two independent experiments. The bar graph shows determination of the relative amount of viral envelope glycoprotein-positive cells. At least 500 polykaryon cells were counted for each sample. (B) In two independent experiments, RSCh or DF-1 cells were treated before fusion, in addition to mitomycin, with an inhibitor of proteosynthesis, puromycin (10 μg/ml, 15 h) or cycloheximide (10 μg/ml, 24 h). The second and fourth days after fusion, the cells were lysed, total RNA was isolated, and cDNA was prepared. The amount of env mRNA was measured by RT-qPCR.

Furthermore, we pretreated partner cells separately with protein synthesis inhibitors (puromycin or cycloheximide) before fusion to see whether the decreased amount of proteins with a shorter half-life influences the effect of fusion on splicing. We observed that the increase in env mRNA synthesis after fusion was significantly lower when the permissive chicken partner cells were exposed to puromycin or cycloheximide (Fig. 5B). We assume that the level of the factor required for env mRNA synthesis in heterokaryons decreased due to proteosynthesis inhibition in chicken fibroblasts, and thus the fusion was not so effective in triggering env mRNA splicing. Treatment of RSCh cells with proteosynthesis inhibitors did not produce any significant effect on the env mRNA level after fusion. This indicates that at least for env mRNA synthesis, some chicken protein factors are required.

Nuclear export and stability of gRNA.

Although we did not detect any considerable change in the total amount of gRNA after fusion with chicken cells, cell fusion might have exerted its influence even at a more subtle level related to gRNA egress from the nucleus. We therefore compared the levels of gRNA export in RSCh versus the positive control 16Q (quail cells harboring a defective RSV genome producing gRNA and src mRNA). In the latter case, cytoplasmic gRNA reached 20% of nuclear gRNA, in contrast to RSCh, where gRNA representation in the same fraction dropped 10-fold (Fig. 6A). Based on a repeated experiment, in which we measured the gRNA quantity present in the cytoplasmic fraction, we obtained evidence that the RNA egress from RSCh nuclei increased 3 times after fusion with DF-1 cells (Fig. 6B). The cytoplasmic fraction purity was also tested, and it was established that contamination with nuclear RNA did not exceed 4% (Fig. 6C) and did not change after the fusion (Fig. 6D).

FIG 6.

Nuclear export and stability of gRNA. (A) Viral gRNA was quantified in cytoplasmic and nuclear fractions. The amount of gRNA was determined in fractionated RSCh and 16Q cells by RT-qPCR. The cytoplasmic RNA levels are given as percentages of nuclear RNA levels. The mean values ± standard deviations for two to three independent experiments are shown. (B) The gRNA level in the cytoplasmic fraction of RSCh cells after mitomycin (mit) treatment and fusion with DF-1 cells was analyzed 3 or 4 days after the fusion. The data represent means ± standard deviations for three independent experiments. (C) The purity of the isolated fraction is shown by pre-GAPDH levels. Cellular pre-mRNA of GAPDH in RSCh was detected via RT-qPCR. The cytoplasmic RNA levels are given as percentages of nuclear RNA levels. The results are means ± standard deviations of three separate experiments. (D) The purity of RSCh cytoplasmic fractions after mitomycin treatment and fusion with DF-1 cells was measured by pre-GAPDH in three independent experiments. (E) The stability of viral RNAs in DF-1 cells infected with SR-RSV-D and RSCh cells was measured. Cells were incubated with actinomycin D (1 μg/ml) for each designated time period followed by RNA isolation and RT-qPCR. Graphs show levels of viral mRNAs at various time points after addition of actinomycin D, presented as the percentage of the RNA level at the time of addition of actinomycin D. Viral RNAs were normalized to GAPDH mRNA levels. The data represent the means ± standard deviations of three independent experiments.

Finally, we checked gRNA stability after RNA synthesis inhibition using exposure to actinomycin D. In these experiments, we compared RSCh cells with SR-RSV-infected DF-1 cells and found that in hamster cells, the gRNA stability dropped during the first hours after actinomycin addition, while in chicken cells, gRNA remained stable (Fig. 6E).

These results show that gRNA in RSCh cells is unstable, its level is low, and although its total amount is not altered by fusion with chicken cells, gRNA nuclear export significantly increases.

Gag expression after cell fusion.

The viral gRNA also serves as the mRNA encoding the Gag polyprotein. Thus, we decided to confirm the above-mentioned data concerning gRNA with direct Gag protein measurement using immunology approaches. Using antisera against Gag for RSCh cell staining, weak fluorescence diffused in the cytoplasm was found. The fluorescence intensity was clearly lower than that obtained in heterokaryons (Fig. 7A).

FIG 7.

Gag expression. (A) Determination of protein Gag by immunofluorescence labeling in RSCh cells and RSCh cells fused with DF-1 cells. Cells were fixed on the fourth day after fusion and stained with anti-p27 antibody and Cy3-conjugated secondary antibody. (B) Expression of Gag precursor Pr76 and product p27 was determined by Western blotting with anti-p27 antibody of total cell lysates and VLPs. Comparison of Gag production in RSCh cells and RSCh cells mixed with DF-1 cells with or without fusion is shown. (C) Gag expression was analyzed by Western blotting with anti-p27 antibody in chicken DF-1 cells with or without preinfection with helper virus or transfection with pcGagPol, quail QEF, and 16Q cells.

By Western blotting, we again observed that RSCh cell lysates produce almost undetectable amounts of the Gag precursor protein (Fig. 7B). Furthermore, when RSCh cells were fused with DF-1 cells, expression of Gag protein increased, and the regularly processed Gag p27 product appeared. By employing Western blotting, we also verified the presence of processed Gag synthesis products in DF-1 cells transfected with pcGagPol, DF-1 cells infected with the ALV helper viruses, and the 16Q cell line (Fig. 7C).

Finally, in order to check the release of virus-like particles (VLPs), the supernatant from RSCh cell culture was spun through sucrose cushion, and the sediment was also subjected to Gag analysis. As documented in Fig. 7B, the sediment obtained from RSCh cells was devoid of the presence of Gag. Not surprisingly, the sediment from RSCh cells fused with DF-1 cells gave rise to the clear p27 band, thus reinforcing the importance of permissive cells for virus particle formation.

We have shown that fusion of mammalian RSCh cells with chicken cells induces not only RSV env mRNA splicing, Env glycoprotein synthesis, and gRNA nuclear export but also Gag protein synthesis and processing, as well as VLP formation and release from the fused cells.

Complementation with env and gag genes.

Our results pointed to the importance of envelope glycoprotein and Gag protein in the facilitation of virus rescue. We therefore decided to test env and gag gene complementation in RSV rescue experiments. In order to test virus rescue efficiency, RSCh cells were fused with DF-1 cells producing Env or Gag, fused cells were plated on indicator cells, and the rescued transforming virus formation was measured as foci by the infectious center assay. We employed CEF-BL cells as the indicator, because DF-1 as a modified and stabilized cell line cannot be used for focus screening (our and others' repeated observations).

As Env, we employed vesicular stomatitis virus G glycoprotein (VSV-G), which is known to efficiently provide a heterologous envelope to retroviruses (29). As shown in Fig. 8A, RSCh fused with VSV-G-transfected DF-1 produced 5 times more foci than RSCh fused with untransfected DF-1 cells. In order to examine the influence of processed Gag protein on virus rescue, we transfected DF-1 cells with the plasmid pcGagPol containing a selectable marker and the gag gene. After transfection and selection, DF-1 cells expressed small amounts of Gag (Fig. 7C). After fusion with RSCh cells, we observed no increase in RSV rescue in comparison with untransfected DF-1 cells. These results indicate that a low level of Gag cannot complement RSV replication in nonpermissive mammalian cells.

FIG 8.

Virus rescue efficiency after complementation with Gag, Env, or both. RSCh cells were complemented by fusion with avian cells synthesizing either VSV-G or Gag or infected with helper virus. A total of 10⁴ fused cells was plated on indicator cells, and after 21 days, infection centers were counted. The data show means ± standard deviations of two to four independent experiments, and differences after complementation are depicted. (A) Virus rescued after fusion of RSCh cells with DF-1 chicken cells (VSV-G-transfected DF-1, DF-1 transfected with pcGagPol, RAV-1-infected DF-1) and QEF quail cells (VSV-G-transfected QEF and 16Q) was tested on CEF-BL cells. DF-1 and QEF cells were transfected with VSV-G 1 day before fusion. RSCh cells mixed with DF-1 cells without fusion were used as a negative control. (B) Tissue culture fluids (supernatants [sup]) taken the second day after VSV-G transfection of RSCh and 16Q cells, as well as untransfected RSCh and 16Q supernatants were centrifuged (13,000 × g, 20 min, 4°C) and tested in the focus assay on two parallel dishes on CEF-BL.

In order to check whether larger amounts of Gag can influence virus rescue efficiency, we used the 16Q cell line, which is derived from QEF cells transformed by BH-RSV (the RSV strain lacking the env gene) and therefore rich in Gag protein (Fig. 7C). RSCh fusion with quail QEF cells was slightly less effective than fusion with DF-1 cells (Fig. 8A), but when we employed 16Q cells as partner cells, the virus rescue efficiency markedly increased. In order to confirm our results with VSV-G complementation, RSCh cells were fused with VSV-G-transfected QEF cells. Virus rescue efficiency was 13 times enhanced (Fig. 8A).

A direct way to equip mammalian cells with both env and gag gene products is provided by their fusion with ALV-infected partner cells. In such a way, DF-1's rescue activity in fusion experiments with RSCh was increased by 2 orders of magnitude when DF-1 cells were exposed to RAV-1, a helper virus replicating in such cells (Fig. 8A).

Mixture of RSCh and DF-1 cells without PEG fusion was tested to determine whether the background control of the virus rescue was likely caused by spontaneous cell fusions. Mixture of 10⁵ cells resulted in no foci or one focus, which corresponds with previously published data (5). Furthermore, we tested infectious virus formation in culture media from the tested cells (Fig. 8B). No virus activity was recorded, with the exception of the positive control represented by 16Q cells transfected with VSV-G, which as expected produced 6 × 10³ FFU/ml.

In summary, these findings show that env gene complementation can significantly increase efficiency of virus rescue. The increase after gag gene complementation was evident only in the case of high expression of this gene. The best results were achieved when RSCh cells were fused with chicken cells infected with a helper virus providing both env and gag gene products.

DISCUSSION

For decades RSV and avian leukosis viruses have been a cornerstone of retrovirus research (30). Furthermore, RSV was shown to transgress class barriers under experimental conditions, which produced new challenges related to provirus integration, its expression, and rescue of infectious virus (31–33). In this report, we concentrated on the molecular characterization of the RSV-transformed rodent cell line RSCh and the consequences of RSCh cell fusion with avian cells leading to infectious RSV rescue. We are aware of the fact that virus rescue is only one side of a more general problem involving provirus expression in heterologous cells, including involvement of known cellular factors. This issue was thoroughly upgraded (34), and we shall return to it in our next study.

In order to extend and update the model of SR-RSV-transformed rodent cells, we selected the original well-characterized cell line RSCh, which had been studied systematically and quantitatively in a series of previous communications (5, 13, 33). Our present data document that this tumor cell line produces very small amounts of env mRNA, confirming our previous Northern blot analysis of three rodent tumor cell lines (15). However, when fused with DF-1 chicken cells lacking endogenous alpharetroviral loci, synthesis of both env mRNA and glycoprotein is initiated quickly. Similarly, Gag protein synthesis is increased, its regular cleavage takes place, and in the tissue culture supernatant fraction corresponding to VLP, the processed Gag is present. In such a way, the main molecular steps required for virus rescue are accomplished thanks to rodent cell complementation with the chicken cell factor(s). The complementation hypothesis is supported by the finding that protein synthesis inhibitors block increase of the env mRNA level when applied to DF-1 cells and not to RSCh cells before the fusion.

One of the factors responsible for the env mRNA threshold synthesis might be related to aberrant splicing. We found that quantified aberrant splicing of the env gene represents only 5.6% of total viral RNAs in nonpermissive cells, but it is also detectable in permissive cells, where it reaches 1.3%. Aberrantly spliced RNAs utilized different cryptic splice donor (SD) sites and the same regular splice acceptor (SA) site. The splice variants designated Dbl1 and Cryptic1 have already been described in PR-RSV (25), but the new ones—Dbl2, Cryptic2, and Cryptic3—are characteristic of SR-RSV, where a single nucleotide substitution created a new cryptic SD site. From both our data and reference 25, it seems highly unlikely that aberrant splicing alone might play any decisive role in diminishing the pool of full-length env transcripts. Rather, the env gene represents the most recent innovation in the retrovirus evolution, and therefore we can expect its variability.

Quite a lot of attention has been paid to the molecular mechanisms playing a role in RSV gRNA export from the cell nucleus to the cytoplasm. Unique elements called direct repeats (DRs) constituting the 3′ end of the viral genome were shown to represent a genomic “mark” licensing gRNA export and stability. RSV contains two DRs, which are structurally highly similar. A series of subsequent articles using various upstream or downstream DR modifications provided data indicating their role in the virus life cycle (35–41). The most convincing are observations obtained in viruses lacking both DRs that provided evidence for the drop in gRNA stability, egress, and possibly also Gag product synthesis (35). Our study of RSCh cells revealed some similarity to viruses lacking DR. This concerns gRNA instability, decreased export from the nucleus, and low efficiency of Gag product synthesis. Furthermore, as reported previously (35) and confirmed in our experiments (unpublished observation), DR deletion does not impart any effect on the RSV replication in rodent cells. On the other hand, it was demonstrated that DR facilitates gRNA egress from the nucleus in mammalian cells (42). Still, there remains the possibility that rodent cells are devoid of some factor(s) needed for interaction with DRs or produce an inhibitor paralyzing DR function, especially at later stages of gRNA processing that follow its export. Other factors, such as Gag import in the nucleus (43) and nonstructural Gag protein cleavage product p10, which is equipped with the nucleus localizing signal (44), should also be taken into account.

Interesting results were obtained by studying the consequences of partner cell complementation with gag and env gene products or helper virus infection. So far, it is known that RSV-transformed rodent cells are not complemented when different independently obtained cell lines are fused one with the other (45). Similarly it was revealed that infection of transformed cells with rodent retroviruses does not lead to RSV rescue in spite of the fact that rodent retroviruses complete their replication cycle (46). Here we utilized a different strategy in which we did not complement nonpermissive rodent cells but instead complemented permissive avian partner cells with gag, env, or both gene products. Under such conditions, a clear increase in RSV yield after cell fusion occurred; however, it was dependent on the degree of complementing gene expression. This supported our conclusion that the deficiency in rodent cells lies in part in their inability to synthesize or process gag or env gene products. This deficiency can be diminished by supplying these products in trans. However, this transcomplementation needs to be supported by fusion with permissive cells. gag or env gene products are not sufficient to ensure virus rescue in nonpermissive rodent cells by themselves, but they can increase RSV rescue efficiency after fusion with chicken cells. This indicates that not only complementation with gag and env gene products but also other factors from permissive cells are required.

We are finishing this article with a quotation from a recent HIV review (47) stating that “the long-standing puzzle of the cell-type-dependent requirement … for viral replication remains an exciting area of research.” In the case of RSV strains, this area has been with us for long years but acquired only humble attention.