Characterization of Provirus Clones of Simian Foamy Virus Type 1

Abstract

We have cloned proviral DNA of simian foamy virus type 1 (SFV-1) from linear unintegrated DNA (pSFV-1). Transfection of pSFV-1 induces cytopathology in several cell lines with supernatants from the transfected cell culture containing infectious viral particles. Electron microscopy of the transfected cells revealed foamy virus particles. Deletion analysis of pSFV-1 indicated that the transcriptional transactivator (tas) gene located between env and the long terminal repeat is critical for virus replication, whereas the second open reading frame (ORF-2) in this region is dispensable. Although the tas and ORF-2 regions of foamy viruses have significantly diverged, the results presented here suggested that the gene products have similar functions. Recombinant pSFV-1 containing the cat gene was able to transduce the heterologous gene, indicating the utility of SFV-1 as a vector. An infectious clone of SFV-1 which is distantly related to the human foamy virus will provide a means to understand the biology of this unique group of viruses.

Foamy viruses are a unique group of retroviruses that belong to the Spumavirus genus. These viruses are found in many mammalian species and are extremely cytopathic in cell culture. Foamy viruses appear to be nonpathogenic in naturally, as well as experimentally, infected animals. A human foamy virus (HFV) and several simian foamy viruses (SFVs) have been molecularly cloned, and their genomes have been completely sequenced (4, 8, 12, 14, 16, 21). In addition to the structural genes, the genome of foamy viruses contain regulatory genes. The human foamy virus has three open reading frames (ORFs) located between the env gene and the long terminal repeat (LTR). All of the other foamy viruses molecularly characterized thus far have only two ORFs in the corresponding region. The first ORF is a transcriptional transactivator (tas, formerly known as taf) which augments gene expression directed by the viral promoters (9, 11, 19, 22, 24, 30). For HFV, the tas gene has been shown to be critical for virus replication (2, 13). The functions of the other ORFs located between the env gene and the LTR are not known and are dispensable for virus replication (2, 13).

Infectious clones of the HFV and the SFVs (SFVcpz, SFV-6, and SFV-7) have been described (9, 13, 23, 29). The genomes of these foamy viruses are highly related and have homology ranging from 86 to 93% (9). Comparisons of SFV-1 or SFV-3 to SFVcpz or HFV show significant differences, especially in the tas and ORF-2 regions (9, 16). These ORFs of SFV-1 or SFV-3 are related by less than 40% to SFVcpz or HFV. Tas of SFV-1 does not activate gene expression directed by the HFV LTR promoter (17). Similarly, Tas of HFV does not transactivate SFV-1 LTR gene expression (24). Determining whether Tas or a gene product containing the ORF-2 region of SFV-1 is critical for virus replication has been hampered by the lack of an infectious clone. Recently, it has been suggested that foamy viruses may have a replication strategy different from those of other retroviruses, with features of both hepadnaviruses and retroviruses, as well as features distinct from both (3). Therefore, an infectious clone of SFV-1 will help elucidate the unique features of foamy virus replication. In addition, foamy viruses are currently being exploited for use as retroviral vectors; thus, infectious clones of SFV-1 isolates will be advantageous. In this report, we describe the construction of SFV-1 infectious clones, the effect of mutations in the regulatory genes, and gene transfer of a reporter gene by an SFV-1 vector.

Construction and characterization of SFV-1 infectious provirus DNA.

Total DNA from SFV-1-infected Cf2Th (canine fibroblast) cells was isolated in accordance with a standard protocol (27). Foamy viruses produce large amounts of unintegrated linear DNA. The plasmids containing the SFV-1 proviral DNAs were constructed by cloning blunt-ended SFV-1 linear unintegrated DNA into a pUC118 plasmid. DNA was subjected to agarose gel electrophoresis, and the region corresponding to the size of unintegrated SFV-1 linear DNA was eluted. The DNA was treated with T4 polymerase and Klenow fragment to create blunt ends and was cloned into the HindII restriction enzyme site of plasmid pUC118. We have identified three types of proviral clones (pSFV-1, pSFV-1/taorf2, and pSFV-1/enorf2; see Fig. 3A). The first type of clone had the full-length proviral genome (pSFV-1). The second type of proviral clone we identified had a deletion at positions 10,225 to 10,520 (pSFV-1/taorf2). This deleted region corresponds to a region that gets spliced out to generate bet transcripts (15, 20). The majority of the proviral clones screened were the pSFV-1/taorf2 type. Proviral DNA analogous to pSFV-1/taorf2 has also been described for HFV (26). It is proposed that this shorter provirus originated from singly spliced RNA with joined exons of bet, and upon reverse transcription of the singly spliced RNA, the shorter provirus accumulates in infected cells (26). However, it is not clear why this region is selectively spliced out at a higher rate to produce a shorter proviral DNA. The third type of proviral clone had a deletion from position 9,747 in the env gene to exactly the beginning of the 3′ LTR, at position 11,351 (pSFV-1/enorf2). This deletion removes 76 amino acids at the carboxy terminus of env and places the rest of the protein in frame with the carboxy-terminal 117 amino acids of ORF-2 located in the LTR.

FIG. 3.

Mutational analysis of infectious pSFV-1. (A) Genome organization of pSFV-1 and mutant derivatives. The hatched lines represent the deleted regions. pSFV-1/taorf2 and pSFV-1/enorf2 are proviral clones with natural deletions. (B) Effect of mutations on virus replication. Plus and minus signs represent the presence and the absence of a CPE, respectively. Samples for RT assays were harvested at day 9 posttransfection.

To determine the infectious potential of pSFV-1, several cell lines were transfected with 2 μg of the plasmid. Transfections were performed by the liposome-mediated method using Lipofectamine Reagent (Life Technologies, Inc., Gaithersburg, Md.). Transfected cells were observed for up to 10 days for the appearance of a cytopathic effect (CPE). Cells showed typical foamy virus cytopathology, which is generally characterized as the formation of intracellular vacuoles in multinucleated giant cells and, in some cases, balloon formation. The extent of cytopathology and the rate at which it occurred varied among the cell lines studied (Fig. 1A). In 293 (human fibroblast) cells, cytopathology was observed as early as 3 days posttransfection. L-929 (murine fibroblast) cells and COS-7 (African green monkey fibroblast) cells showed cytopathology beginning at days 6 and 8, respectively. Optimum cytopathology was observed at 10 days posttransfection in all of the cell lines tested. pSFV-1-transfected cells kept longer than 10 days in culture were completely destroyed as a result of massive infection. Supernatants from these cultures were monitored by reverse transcriptase (RT) assay for virus particle release, and the RT values corresponded to the level of cytopathology observed (Fig. 1B). The RT values from the later days posttransfection were equivalent to those from cells infected with virus particles. To establish whether the transfected cultures produced infectious virus particles, media from transfected cells were cleared by filtration and used to infect fresh cell lines permissive to SFV-1. Supernatants harvested from all cells transfected with pSFV-1 transmitted virus particles to uninfected cells as determined by the CPE on the infected cells and RT analysis (data not shown).

FIG. 1.

Infectivity of pSFV-1 in the 293, L-929, and COS-7 cell lines. (A) Levels of CPE on infected cells at different time points after pSFV-1 transfection. Levels of cytopathology were scored as follows: −, no CPE; +, 10 to 20% CPE; ++, 30 to 50% CPE; +++, 60 to 70% CPE; ++++, 80 to 90% CPE. Lysis refers to greater than 99% loss of cells in the infected cell culture. The numbers at the top indicate days after DNA transfection or virus infection. (B) RT analysis of supernatants from cell cultures transfected with pSFV-1 at different time intervals. RT was assayed under conditions optimized for foamy virus with Mn²⁺ cation (3). Samples were harvested at the indicated time intervals in triplicate, and RT activity was assayed. Less than 10% variation in replicate samples was observed.

EM examination of pSFV-1-transfected cells.

To demonstrate that pSFV-1 generates mature viral particles, transfected L-929 cells were examined by transmission electron microscopy. Cloned proviral SFV-1-transfected cells were prepared for electron microscopy (EM) 10 days posttransfection, when significant cytopathology was noted. Cells infected with wild-type SFV-1 were used as positive controls. These cells were prepared for EM 4 days after infection, when cytopathology was optimum. As shown in Fig. 2, viral particles were detected in samples from virus-infected, as well as pSFV-1-transfected, cells. The particles were spherical with the prominent envelope spike structure typical of foamy virus, demonstrating that transfection of the recombinant clone pSFV-1 produced mature viral particles.

FIG. 2.

Transmission EM analysis of virus particles from pSFV-1-transfected (A) and SFV-1-infected (B) L-929 cells. Virus particles are indicated with arrowheads (magnification, ×92,000). For ultrathin sectioning, cells were fixed with 2.5% glutaraldehyde in phosphate buffer at room temperature for 1 h, postfixed with 1% osmium tetroxide for 1 h, dehydrated in a graded series of ethanol, and embedded in Embed 812 (Polysciences, Warrington, Pa.). Ultrathin sections were stained with uranyl acetate and lead citrate. Stained sections were placed on grids and examined with a Hitachi H7000 transmission electron microscope (Hitachi Instruments, San Jose, Calif.).

Mutational analysis of pSFV-1.

For HFV, the tas (bel-1) gene product was shown to be critical for virus replication (2, 13). The Tas proteins of HFV and SFV-1 share 39% amino acid identity. Both proteins activate gene expression under the control of viral promoters. To demonstrate that the tas gene of SFV-1 is essential for virus replication, the region from positions 10,008 to 10,202 was deleted from pSFV-1 (pSFV-1/dtas) and tested for virus replication (Fig. 3). Cells transfected with pSFV-1/dtas were monitored for 20 days for cytopathology and for virus particles by RT assays. Transfected cells lacked cytopathology, and neither culture supernatant nor extracts from three cycles of frozen-thawed cells indicated RT activity above the background. This suggested that the tas gene of SFV-1, similar to HFV tas, is critical for virus replication. Cytopathology was noted in cells transfected with wild-type pSFV-1 beginning at day 6 and significantly increased at 10 days posttransfection. Studies comparing SFV-1 and SFV-3 with HFV showed that the U3 domains of the LTRs and the predicted amino acid sequences of the tas genes have greatly diverged (16). Consequently, Tas of either SFV-1 or HFV does not cross transactivate gene expression directed by the heterologous LTR (18, 25). Interestingly, however, the predicted proteins appear to have similar structures, with short stretches that are highly conserved. Several of these stretches have been determined to be critical for transactivation of gene expression directed by the viral LTR (7, 18, 31). These functional domains of Tas include a highly conserved carboxy-terminal hydrophobic activation domain and basic rich regions that include a nuclear localization signal. Therefore, although the tas genes of foamy viruses have diverged, there are regions that are conserved structurally and functionally, leading to a similar regulation of virus replication.

A deletion mutation was also introduced into the ORF-2 region, at positions 10,872 to 11,208, to examine the effect of this mutation on SFV-1 replication (pSFV-1/dorf2) (Fig. 3). Transfected cells showed cytopathology beginning 6 to 8 days posttransfection. The extent of the CPE and RT values were similar to those of cells transfected with wild-type pSFV-1. Virus replication was not observed in cells transfected with either pSFV-1/taorf2 or pSFV-1/enorf2. The bel-2 and bel-3 genes of HFV have been shown to be dispensable for in vitro virus replication (2, 13). The region encompassing ORF-2 (bel-2) also shows significant variation between HFV and SFV-1, which share only 38% homology. Comparisons of the ORF-2 amino acid sequences from different foamy viruses reveal short stretches of highly conserved regions, implying similar functions for the proteins encoded by ORF-2 (8, 16). SFV-1 studies presented in this report and HFV studies done by others (2, 9, 13) clearly indicated that foamy viruses can replicate in cell culture in the absence of the protein(s) encoded by the ORF-2 region. The postulated ORF-2 (bel-2)-encoded protein has been identified as a 44-kDa protein by some investigators (5, 13), while others have failed to find an ORF-2 product (2, 7, 32). However, a bet-encoded protein which is a product of a spliced message containing the first 88 amino acids of Tas fused to the last 390 amino acids of ORF-2 was found to be highly expressed in the cytoplasm of HFV-infected cells (7, 20). Proteins analogous to Bet have also been detected for SFV-1 (6, 15). The function of bet or ORF-2 remains to be determined. By using a sensitive assay, Yu and Linial (32) have determined that mutations in the bel-2 or bet gene decreases cell-free virus transmission about 10-fold, suggesting that either bet or bel-2 plays a role in efficient free virus transmission, similar to vif of the lentiviruses.

Recombinant SFV-1 with a reporter cat gene.

Since foamy viruses have been considered to be ideal vectors for gene transfer, we tested the ability of cloned SFV-1 proviral DNA to transduce a reporter cat gene. The cat gene was placed downstream from the internal promoter (pSFV-1/cat), replacing the tas and ORF-2 regions at positions 10,225 to 11,208 (Fig. 4). A second recombinant was also constructed by replacing the same region with the simian virus 40 (SV40) early gene promoter controlling cat expression (pSFV-1/svcat). To determine the ability of these recombinants to transduce the cat gene, an L-929 cell line expressing the tas gene was established. The SV40 promoter was placed upstream of the coding sequence of the tas gene, and this plasmid construct was cotransfected with a plasmid expressing the neomycin resistance gene (neo) into L-929 cells by electroporation as described previously (1). Transfected cells were grown in medium containing 1 mg of the neomycin derivative G418 per ml. Neomycin-resistant colonies were propagated for four rounds of single-cell cloning. To confirm the expression of tas, we examined the level of gene expression directed by either the viral LTR or the internal promoter by transfecting pSFV-1 LTR/CAT or pTM1/CAT, respectively, into the L-929-tas cell line (Fig. 4). The construction of plasmids pSFV-1 LTR/CAT and pTM1/CAT has been described previously (15, 17). For comparison, the same plasmids were also used to transfect L-929 cells not expressing the tas gene. Transfections were performed by the DEAE-dextran method (1). For each chloramphenicol acetyltransferase (CAT) assay experiment, duplicate cell cultures were transfected with 2 μg of the reporter plasmid and 3 μg of the effector or carrier plasmid. Cell lysates were prepared 48 h after transfection and assayed for CAT activity. Transient expression assays of L-929 cells demonstrated very low basal promoter activity from these constructs. In the established cell line containing the tas gene, the expression of CAT by either the LTR (75-fold) or the internal promoter (49-fold) was greatly increased. This observation indicated that the L-929-tas cells were expressing the tas gene.

FIG. 4.

Assays of CAT reporter transient expression in cells infected with recombinant virus particles containing the cat gene. Recombinant pSFV-1 containing the cat gene (pSFV-1/cat) or the cat gene under control of the SV40 promoter (pSFV-1/svcat) was constructed by replacing the indicated region of the SFV-1 genome. Filtered supernatants from cell cultures transfected with this plasmid were used to infect cells. pSFV-1 LTR/CAT41 and pTM1/CAT are positive controls transfected into cells to monitor the level of CAT activity. p22A2 is a plasmid containing a promoterless cat gene that was used as a negative control. The values shown are from reactions measuring the conversion of ³H-acetyl coenzyme A to ³H-acetylated chloramphenicol. Generally, less than 5% variation in replicate samples was observed. The CAT values in L-929 cells represent basal activity. Relative activity was calculated by dividing the CAT values in L-929-tas cells by the basal levels in L-929 cells. To calculate the relative activity for pSFV-1/svcat-infected cells, the CAT value of pSFV-1/cat in L-929 cells was used as the basal level.

Once we determined that the L-929-tas cell line expressed functional Tas, the pSFV-1 recombinant containing the cat gene (pSFV-1/cat or pSFV-1/svcat) was transfected into the L-929-tas cell line. Ten days after transfection, supernatant was collected and filtered through a 0.45-μm-pore-size membrane filter. Supernatant containing virus particles was used to infect fresh L-929-tas cells. Three days following infection, CAT assays were performed to determine transduction of the cat gene by SFV-1. As shown in Fig. 4B, cat values were significantly higher than basal-level activity. Cells infected with particles containing recombinant pSFV-1/cat or pSFV-1/svcat had CAT activity 19- or 25-fold higher than the background level, respectively. Cells infected with wild-type virus particles harvested from pSFV-1-transfected cells showed no CAT activity. Furthermore, no CAT activity was observed in L-929 cells infected with pSFV-1/cat transducing particles. The CAT value in L-929 cells infected with pSFV-1/svcat transducing particles was 12-fold higher than the background level. These results suggested that pSFV-1 can be used to develop a vector system for gene transfer. Similarly, vectors constructed based on HFV that encode heterologous genes were able to transduce a variety of cells (25, 28). The HFV studies done thus far indicated that the foamy virus vector titers were low. However, Russell and Miller (25) have shown that the efficiency of transduction of an HFV vector into stationary-phase cultures was higher than that of murine leukemia virus vectors. Furthermore, the efficiency of transduction by a foamy virus vector in primate hematopoietic cells compared favorably with those obtained with murine leukemia virus vectors (10).