Feline Immunodeficiency Virus Orf-A Is Required for Virus Particle Formation and Virus Infectivity

Abstract

The orf-A (orf-2) gene of feline immunodeficiency virus (FIV) is a small open reading frame predicted to encode a 77-amino-acid protein that contains putative domains similar to those of the ungulate lentiviral Tat protein. Orf-A is reported to be critical for efficient viral replication in vitro and in vivo. A series of FIV-pPPR-derived proviruses with in-frame deletions and point mutations within orf-A were constructed and tested for replication in feline lymphoid cells. Orf-A mutant proviruses were also tested for viral gene and protein expression, viral particle formation, and virion infectivity. Deletions within orf-A severely restricted FIV replication in feline peripheral blood mononuclear cells (PBMC) and interleukin-2-dependent T-cell lines. In addition, substitutions of alanines for leucines in the putative leucine-rich domain, for cysteines in the putative cysteine-rich domain, and for a tryptophan at position 43 in Orf-A restricted the replication of FIV mutants. Deletions and point mutations in orf-A imposed a small effect or no effect on FIV long-terminal-repeat-driven viral gene expression and had no effect on viral protein expression. However, release of cell-free, virion-associated viral RNA in supernatants from cells transfected with orf-A mutant proviruses was severely restricted but was rescued by cotransfection with a wild-type Orf-A expression vector. In addition, virions derived from orf-A mutant proviruses expressed reduced infectivity for feline PBMC. Our findings suggest that Orf-A functions involve multiple steps of the FIV life cycle including both virion formation and infectivity. Furthermore, these observations suggest that Orf-A represents an FIV-encoded analog more similar to the accessory gene vpr, vpu, or nef than to the regulatory gene tat encoded by the primate lentiviruses.

Feline immunodeficiency virus (FIV), a member of the lentivirus subfamily of retroviruses, is the etiological agent of a chronic, progressive disease found in cats that is similar to AIDS caused by human immunodeficiency virus (HIV) infection (28, 51). Immunological dysfunctions associated with infection by either virus include decreased CD4 T-cell counts, decreased CD4/CD8 T-cell ratios, and increased susceptibility to infectious agents and opportunistic infections (1-3, 24, 49). FIV exhibits a cell tropism similar to, although broader than, that described for HIV and has been shown to infect CD4 T cells, macrophages, and dendritic cells in vivo (3, 9, 10, 14, 15, 26). As an immunodeficiency-inducing lentivirus, FIV provides an animal model for the development of HIV vaccines and antiviral therapies (4). Elucidation of viral genes necessary for viral replication, cell tropism, and pathogenesis is important for designing efficacious virus-targeted interventions. Accordingly, characterizing functions of specific viral accessory genes shared by the primate and feline lentiviruses will facilitate the use of the FIV animal model in studies testing antiviral strategies for use in HIV infection.

The genomic organization of FIV is similar to that of other members of the lentivirus family, in which three large open reading frames, gag, pol, and env, encode structural proteins (27, 43). Three additional open reading frames include vif, which encodes an accessory protein necessary for viral infectivity (21, 37, 47); rev, which expresses a viral posttranscriptional regulatory protein (30, 48); and a third accessory gene, orf-A, also designated orf-2, whose function has not been fully characterized. FIV genomic organization reveals that orf-A occupies a position within the genome similar to that of the tat genes of other lentivirus family members, implicating Orf-A as a transcriptional transactivator (Tat) of FIV (12, 46).

Two types of lentiviral Tat proteins have been described. Type 1 Tat proteins are expressed by HIV type 1 (HIV-1), HIV-2, simian immunodeficiency virus (SIV), equine infectious anemia virus, and bovine immunodeficiency virus. The ungulate lentiviruses maedi-visna virus (VV) and caprine arthritis-encephalitis virus (CAEV) express type 2 Tat proteins (45). Type 1 Tat-mediated upregulation of viral gene transcription requires the interaction of Tat with the Tat response region (TAR), an RNA stem-loop structure encoded by the long terminal repeat (LTR) (34). The functional domains characteristic of type 1 Tat include a basic region for nuclear localization and binding to TAR (38), an N-terminal acidic region, a conserved core region, and a cysteine-rich region; the latter two are important for transactivation activity (18). Unlike the potent type 1 transcriptional transactivators, type 2 Tat proteins are weak transactivators (25, 35, 41, 44). Moreover, lentiviruses that express a type 2 Tat do not have LTRs containing a hairpin RNA loop structure characteristic of the HIV TAR region; instead, they include LTR promoters capable of high basal levels of transcription (25, 35, 41, 44). Therefore, type 2 Tat is thought to mediate transactivation in a TAR-independent manner. Although functional domains for type 2 Tat have not been well characterized, putative domains described for these proteins differ from those mapped for type 1 Tat proteins and suggest that type 2 Tat proteins may function by distinct and different mechanisms (18, 45). Based on the amino acid sequence of FIV Orf-A and the strong promoter activity of the FIV LTR, FIV Orf-A has been designated a type 2 Tat protein (45).

The orf-A gene product is necessary for efficient viral replication in interleukin-2 (IL-2)-dependent feline lymphoid cell lines (MCH5-4), in peripheral blood mononuclear cells (PBMC), and in vivo (9, 17, 23, 32, 46, 50). Amino acid sequence alignments of FIV Orf-A with VV and CAEV Tat proteins reveal a similar organization of conserved putative domains including N-terminal acidic and hydrophobic, central leucine-rich, and C-terminal cysteine-rich regions (45). In addition, FIV Orf-A encodes previously unrecognized conserved tryptophans at positions 43 and 66, positioned similarly to conserved tryptophan residues 63 and 85 of VV Tat. However, previous studies have produced conflicting data regarding the ability of Orf-A to transactivate the FIV LTR (7, 12, 22, 41, 44, 50). Two recent studies have demonstrated a moderate upregulation of FIV LTR promoter activity by coexpression of Orf-A in transient reporter gene expression assays (7, 12). In contrast, earlier studies revealed either a small effect or no effect imposed by Orf-A on FIV LTR-directed gene expression (22, 41, 44, 50). Based on these findings, the role of Orf-A as a critical transactivator of viral gene expression remains questionable. Therefore, other functions through which Orf-A mediates efficient viral replication warrant further investigation.

To elucidate the primary function of Orf-A in the viral life cycle and the mechanisms through which Orf-A mediates optimal viral replication and production, we constructed proviruses containing either deletions or point mutations in orf-A. Mutant orf-A proviruses were compared for the ability to replicate in feline primary PBMC and cell lines and were tested for viral gene expression, viral protein expression, and virus particle formation. Furthermore, Orf-A mutant viruses were tested by infectivity assays. Our findings suggest that late steps in the viral life cycle involving virion formation and early steps associated with virus infectivity are restricted by mutations in orf-A. Lastly, domains shared by ungulate lentivirus Tat proteins and FIV Orf-A are functionally important for viral replication and virus production.

MATERIALS AND METHODS

Construction of FIV orf-A mutant proviruses.

All orf-A mutant provirus plasmids (Fig. 1) were derived from the previously described provirus pSV-pPPR (pSVWT) (5). Plasmid pSVWT is an FIV-pPPR-based provirus (31) directed by the simian virus 40 (SV40) early promoter. The 5′ region of pSVWT is characterized by a chimeric SV40 promoter-FIV LTR containing the enhancer and TATA box of the SV40 early promoter in addition to 16 nucleotides (nt) of FIV U3 (nt −1 to −16) and the RU5 domains of the FIV LTR. Designations of orf-A mutant provirus constructs containing the SV40 promoter-FIV LTR (SV40pr/R/U5) hybrid have a “pSV” prefix. A second set of orf-A mutant provirus constructs, retaining the original wild-type (WT) U3 and RU5 sequences of the FIV LTR, was constructed; these designations have an “FIV-pPPR” prefix (Table 1). pSVWT and pSV orf-A mutant provirus constructs were assayed in replication experiments. WT FIV-pPPR and FIV-pPPR orf-A mutant provirus constructs were used to assess WT FIV LTR-directed viral mRNA expression and release of virion-associated viral RNA.

FIG. 1.

FIV-pPPR orf-A mutants. All orf-A mutant proviruses were derived from pSV-pPPR, which is an FIV-pPPR-based provirus. The N-terminal acidic or hydrophobic, central leucine, and C-terminal cysteine regions, as well as conserved tryptophans, are indicated by solid rectangles. All deletions of Orf-A amino acids are indicated by dashes and are in frame. Dots represent amino acids identical to those present in WT Orf-A. Deletion of 13 residues (residues 24 to 36) and replacement of phenylalanine 23 by an aspartic acid (D) residue (introduced by the inserted ClaI restriction enzyme site) are shown for the ΔorfAmid mutant. All other amino acid substitutions (asterisks) are alanine substitutions.

TABLE 1.

Description of FIV orf-A mutant proviruses

Namea	Amino acid position	Mutation
ΔorfA	1-72	210-bp in-frame deletion
ΔorfAmid	24-36	39-bp in-frame deletion
	23	F→D
orfA-trp43stop	43	W→stop
orfA-L-A	34, 37, 38, 41, 42, 45	L→A
orfA-W43A	43	W→A
orfA-W66A	66	W→A
orfA-WW/AA	43, 66	W→A
orfA-C-A	55, 57, 60, 61	C→A

^aThese designations are preceded either by the prefix pSV, indicating that the provirus contains a 5′ SV40 promoter-FIV LTR hybrid, or by the prefix FIV-pPPR, indicating that the provirus contains a WT 5′ LTR.

orf-A mutant provirus constructs pSVΔorfA and pSVΔorfAmid encode large and small deletions, respectively, spanning the orf-A gene (Fig. 1). The pSVΔorfA construct contains a 210-bp in-frame deletion and a unique ClaI restriction enzyme site introduced into pSVWT between nt 5991 and 6202 (Table 2). For construction of pSVΔorfA, a subgenomic 739-bp fragment spanning the 3′ half of vif was PCR amplified from pSVWT by using primers OrfASalU5272 (containing a SalI site) and OrfAClaL5991 (containing a ClaI site), each designed to include terminal SalI and ClaI restriction enzyme sites (Table 2), and then cloned into pSP72 (Promega Biotech, Madison, Wis.) to yield plasmid pAB. Next, a subgenomic fragment spanning the terminal 21 nt of orf-A and the first 1,203 nt of env was PCR amplified using primers OrfAClaU6202 and OrfABglL7283 (Table 2) and then cloned into pAB downstream of the 739-bp subgenomic fragment to create plasmid pAD. Finally, replacement of the Bsu361-BclI subgenomic fragment of pSVWT with the Bsu361-BclI fragment of pAD produced the orf-A deletion mutant provirus plasmid pSVΔOrfA. A similar strategy utilizing primers OrfASalU5272, OrfABglL7283, OrfAClaL6054, and OrfAClaL6100 (Table 2) was employed to create pSVΔorfAmid, which contains a 39-bp in-frame deletion and a substitution of aspartic acid for phenylalanine 23 within orf-A as a result of a ClaI restriction enzyme site insertion (Fig. 1).

TABLE 2.

Oligonucleotides used for plasmid construction

Primer	Oligonucleotide sequence	Genomic position
OrfASalU5272	ACGCGTCGACTTGCAGTGCTCCAAGGAGGG	5272-5301
OrfAClaL5991	CCATCGATTCACAGTTTTCCTCGCCATAACATGC	5991-6023
OrfAClaU6202	CCATCGATACATTATCATAGATACTGCTTAG	6202-6231
OrfABgL7283	GAAGATCTCATACATGTTTGATTAGGCCC	7283-7310
OrfAClaL6054	CCATCGATCTTCTGCAAGGACTACTTTGG	6054-6081
OrfAClaU6100	CCATCGATCTTCTGCAAGGACTACTTTGG	6100-6127
TrpstopU6105	GCAAGGACTTTAGAGACTGAG	6105-6125
TrpstopL6134	CTCAGTCTCTAAAGTCCTTGC	6134-6153
LeuU6098	GAGCTGCGCAAGGAGCAGCTTGGAGAGCGAGATTTAG	6098-6134
LeuL6120	CCAAGCTGCTCCTTGCGCAGCTCTAATCGCTTTATCCC	6120-6155
TrpL6171	GCAAAACGCACATAAACAATCTTTTGATTTAGGTTTTCTAAATCTCAGTCTCCAAAG	6171-6224
TrpL6202	GTAGACTGCAACTGCGCATAATATAATCTCGCAGC	6202-6234
TrpU6151	GATTGTTTATGTTGGTTTTGCTGCAGATTATATTATGCGCAG	6151-6190
CysL6180	TAATCTGGCGGCAAACCAAGCTAAAGCATCTTTTGATTTAGG	6137-6180
CysU6151	GATGCTTTAGCTTGGTTTGCCGCCAGATTATATTATTTGC	6151-6190

Mutant pSVWT provirus constructs pSVorfA-trp43stop, pSVorfA-L-A, pSVorfA-W43A, pSVorfA-W66A, pSVorfA-WW/AA, and pSVorfA-C-A (Table 1), carrying point mutations within orf-A, were produced by PCR-mediated overlap extension as described previously (16) to produce specific amino acid changes within Orf-A. Briefly, four nucleotide primers used for construction of each mutant included an external primer set (OrfASalU5272 and OrfABgL7283) with nucleotide sequences common to all of the mutants and two internal primers containing nucleotide sequence changes specific to each mutant (Table 2). Analogous orf-A mutants constructed with the WT FIV-pPPR provirus were identical to pSV orf-A mutants but retained the WT FIV 5′ LTR domain (Table 1). For construction of the FIV-pPPR orf-A mutants, the NotI-Bsu361 fragment containing the SV40-FIV LTR hybrid promoter from each of the pSV orf-A mutants was replaced with the analogous NotI-Bsu361 fragment of WT FIV-pPPR carrying the WT FIV 5′ LTR. All mutant provirus plasmids were confirmed by restriction enzyme analysis and automated nucleotide sequencing.

Cells.

Crandell feline kidney (CrFK) cells, a feline adherent cell line (ATCC CCL 94), were cultivated in 50% minimum essential medium (MEM) (Invitrogen Corp., Grand Island, N.Y.)-50% Leibovitz L-15 supplemented with 10% heat-inactivated fetal bovine serum (FBS), l-glutamine, and penicillin-streptomycin. 3201-B cells, a feline lymphoid cell line (39), were maintained in RPMI-Leibovitz L-15 medium (1:1) supplemented with 20% heat-inactivated FBS. MCH5-4 cells, an IL-2-dependent feline T-cell line (kindly provided by J. Elder, Scripps Research Institute) (19), were maintained in RPMI medium supplemented with 10% heat-inactivated FBS, 100 U of IL-2 per ml, glutamine, penicillin-streptomycin, sodium pyruvate, and MEM vitamin solution (Life Technologies, Grand Island, N.Y.). Primary feline PBMC were isolated from peripheral blood of specific-pathogen-free cats by density centrifugation in Ficoll-Histopaque (Sigma Chemical, St. Louis, Mo.) and cultured as previously described (40).

Replication of Orf-A mutants in feline PBMC and MCH5-4 cells.

To assess the replication of orf-A mutant provirus constructs, CrFK cells were transfected by electroporation with a pSV orf-A mutant provirus plasmid or with pSVWT (10 μg of plasmid DNA) as described previously (5). Because CrFK cells are nonpermissive for FIV-pPPR infection, 24 h after transfection, CrFK cells were cocultivated with either primary feline PBMC or MCH5-4 cells for an additional 24 h. Cocultivated PBMC and MCH5-4 cells were then removed from adherent CrFK cells, resuspended in fresh culture media, and cultivated up to 4 weeks after transfection. Cell culture supernatants were harvested every 3 to 4 days and tested for virus production by an FIV p24^gag antigen capture enzyme-linked immunosorbent assay (ELISA) (8).

Preparation of virus stocks.

Virus stocks prepared from plasmids pSVorfA-L-A, pSVorfA-WW/AA, pSVorfA-C-A, and pSVWT were generated by plasmid electroporation of CrFK cells followed by cocultivation with specific-pathogen-free feline PBMC as previously described (5). Cells infected with either WT or orf-A mutant viruses were maintained for 14 and 30 days after cocultivation, respectively. Culture supernatants were collected every 3 to 4 days, clarified by centrifugation at 1,500 × g for 10 min, and stored in 1.0- to 1.5-ml aliquots at −70°C. WT viral stocks were assayed for infectious titers (50% tissue culture infective doses [TCID₅₀]) on feline PBMC by using the method of Reed and Muench as previously described (5, 32a). The concentrations of virion-associated FIV p24^gag were measured for 100 and 500 TCID₅₀ of WT FIV virus stock by an FIV p24^gag antigen capture ELISA. By use of reverse transcription (RT)-PCR (13, 29), cDNA was prepared from virion RNA extracted from orf-A mutant virus stocks as described below and was used for PCR amplification of orf-A genes. Amplified orf-A genes were directly sequenced to assess for reversions of introduced mutations.

Expression of intracellular viral mRNA by FIV-pPPR provirus plasmids.

To assess expression of viral mRNA from orf-A mutant proviruses, feline 3201-B cells were electroporated as previously described (41) with either an FIV-pPPR orf-A mutant provirus plasmid (5 μg) or WT FIV-pPPR. All provirus plasmids were cotransfected with a green fluorescent protein (GFP) expression plasmid, pNDeGFP (kindly provided by G. Rhodes, University of California, Davis) (5 μg), to normalize transfection efficiency. Briefly, pNDeGFP was constructed by using a eukaryotic expression vector, pND, that contains the human cytomegalovirus IE1 promoter-enhancer region, the IE1 intron (6), and the bovine growth hormone polyadenylation signal cloned into pUC19. The gene encoding enhanced GFP (EGFP) was digested from plasmid pEGFP-N1 (Clontech Laboratories, Palo Alto, Calif.) by using enzymes BamHI and NotI and was then inserted into pND to produce pNDeGFP. Equal numbers of transfected cells, based on 200,000 GFP-positive cells assayed by flow cytometry, were harvested at 20 h after transfection for quantitation of viral mRNA copies. Harvested cells were first incubated with DNase (10 U/ml) (Roche Molecular Biochemicals, Indianapolis, Ind.)-10 mM MgCl₂-20 mM Tris-HCl (pH 8.0) for 30 min at 37°C to eliminate residual transfected plasmid DNA in cell culture supernatants. Cells were next incubated in 10 mM trypsin for 1 h at 4°C to remove virions bound to the surfaces of cells. Total cell-associated RNA was extracted using the RNeasy kit (Qiagen Inc., Valencia, Calif.) and similarly treated with DNase to eliminate cellular genomic DNA, as well as any residual transfected proviral DNA. cDNA was prepared from total cellular RNA in an RT reaction (10 μl) under conditions previously described (13, 29). The RT reaction volume (10 μl) was increased to 100 μl with double-distilled H₂O, 5 μl of which was assayed for the FIV gag RNA copy number in a real-time TaqMan PCR assay using primers and a probe described previously (13, 29). As a control for possible DNA contamination, total cellular RNA was also tested in the real-time PCR assay. All RNA samples were normalized for quantitation of mRNA transcripts by an assay for expression of feline glyceraldehyde-3-phosphate dehydrogenase (GAPDH) using primers and probes described previously (20). Analysis was performed on an ABI PRISM 7700 sequence detector (Applied Biosystems, Foster City, Calif.).

Assay for virus particle formation.

Supernatants were collected from 3201-B cells 24 h after electroporation with either WT FIV-pPPR or an FIV-pPPR orf-A mutant provirus plasmid in order to measure virus particle formation based on quantitation of cell-free, virion-associated viral RNA concentrations. After clarification of cell culture supernatants, virus particles were pelleted from supernatants by centrifugation at 20,000 × g for 1 h and then treated with DNase as described above. Viral RNA was extracted from pelleted virions by using the QIAamp viral RNA Mini Kit (Qiagen Inc.) and was assayed for the FIV gag RNA copy number in a one-tube, real-time TaqMan RT-PCR assay as previously described (29).

Complementation assay.

To assess the ability of Orf-A to restore the release of virion-associated viral RNA by a provirus with deletions in orf-A, an Orf-A expression vector (pEGFP-Orf-A) was constructed as previously described (M. C. Gemeniano, E. T. Sawai, and E. E. Sparger, submitted for publication). Briefly, the complete orf-A gene was PCR amplified and inserted into pEGFP-C1 (Clontech Laboratories) to produce pEGFP-Orf-A, which fuses EGFP to the N terminus of Orf-A. Expression plasmid pEGFP-Orf-A produced a GFP-Orf-A fusion protein recognized by a rabbit polyclonal anti-Orf-A antibody in Western blot analysis (Gemeniano et al., submitted). 3201-B cells (6 × 10⁶) were cotransfected with 5 μg of FIV-pPPRΔorfAmid and 5 μg of either pNDeGFP (used as a control) or pEGFP-Orf-A. For a positive control, 3201-B cells were cotransfected with 5 μg of WT FIV-pPPR and 5 μg of pNDeGFP. Viral RNA was extracted from virions pelleted from the supernatant harvested from transfected cells and was assayed for viral RNA copy number by a one-tube real-time TaqMan RT-PCR assay as described above.

Virus infectivity assays on feline PBMC.

To measure the infectivities of FIV orf-A mutant viruses, feline PBMC were inoculated with viral stocks generated from pSVWT and pSV orf-A mutant proviruses and then assayed for newly synthesized viral cDNA postinfection. Prior to inoculation, all virus stocks were treated with DNase as described above. As a control for plasmid or genomic viral DNA contamination of virus stocks, a second set of viral stock inocula was heat inactivated by incubation at 100°C for 1 h followed by DNase treatment. Cultivated healthy feline PBMC (5 × 10⁶) were plated in triplicate wells of six-well plates and inoculated with 2.5 ng of virion-associated FIV p24^gag, a viral antigen concentration associated with 500 TCID₅₀ of WT FIV-pPPR virus. Inocula were removed 24 h after infection, and infected cells were harvested, washed twice with phosphate-buffered saline, and treated with DNase. Genomic DNA was extracted from infected cells by using the DNeasy Mini Kit (Qiagen Inc.) and was assayed for newly synthesized FIV cDNA copies by a real-time PCR assay described previously (20, 29). The viral DNA copy number measured for cells infected with heat-inactivated inocula was subtracted as background. Genome equivalents per assay were determined by targeting a single-copy gene by using a feline CCR5-specific real-time TaqMan PCR assay with probe (6-carboxyfluorescein [FAM]-6-carboxytetramethylrhodamine [TAMRA]) fC5.68p (3′-FAM-CCGACGTGAGGCAAATCGCAGC5′-TAMRA), forward primer fC5.48f (TCGGAGCCCTGCCAGAA), and reverse primer fC5.102r (GTAGAGCGGAGGCAGGAGTCT).

Expression of viral proteins by FIV-pPPR provirus plasmids.

To rule out possible disruption of important sequences downstream or upstream of newly introduced mutations, expression of the FIV structural proteins Gag and Env was assessed. CrFK cells were transfected by electroporation as previously described with either a pSV orf-A mutant provirus plasmid construct or pSVWT (10 μg) and were plated onto Thermanox tissue culture slides (Nunc Inc., Naperville, Ill.) as previously described (21). Twenty-four to 48 h after transfection, adherent CrFK cells were assessed by a previously described immunocytochemical assay for expression of both FIV Gag and FIV Env (21).

To further assess viral protein expression from orf-A mutant proviruses, either an FIV-pPPR orf-A provirus plasmid (5 μg) or WT FIV-pPPR was cotransfected with pNDeGFP (5 μg) (control for transfection efficiency) into 3201-B cells (6 × 10⁶) as described above. Approximately 24 h after transfection, viral proteins were immunoprecipitated from cell extracts with FIV-infected cat sera by using protocols described previously (21, 36). Immunoprecipitates were analyzed for FIV p24^gag expression by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis and Western blot analysis using a mouse monoclonal antibody specific for FIV p24^gag (a gift from N.C. Pedersen, University of California at Davis), as described previously (21). Briefly, after SDS-polyacrylamide gel electrophoresis, proteins were transferred to a nitrocellulose membrane that was probed with an FIV p24^gag-specific monoclonal antibody. FIV p24^gag was detected by using a horseradish peroxidase-conjugated rabbit anti-mouse antibody (Pierce, Rockford, Ill.). Bands were visualized by use of SuperSignal West Dura Substrate (Pierce) and exposure of the nitrocellulose membrane to X-ray film.

RESULTS

Replication of orf-A mutants in feline primary PBMC and cell lines.

Previous experiments using FIV proviruses containing mutations in orf-A demonstrated that orf-A is important for efficient viral replication (9, 32, 50). To investigate the requirement of orf-A for efficient replication of the molecular clone FIV-pPPR in feline T cells, provirus plasmids pSVΔorfA and pSVΔorfAmid were constructed to contain large (210-bp) and moderate-sized (13-bp) deletions, respectively, within orf-A (Fig. 1; Table 1). The molecular clone FIV-pPPR was utilized for these studies because this virus has been shown to replicate efficiently in primary feline lymphocytes in vitro and to produce a moderately robust viremia in vivo (9, 40). The orf-A gene carried by pSVΔorfA was characterized by deletion of amino acid residues 1 to 71 and retention of codons for the carboxy-terminal 7 amino acids. pSVΔorfAmid contained a deletion that replaced amino acid residues 23 to 36 with an aspartic acid residue and retained Orf-A N-terminal amino acids 1 to 22 and C-terminal amino acids 37 to 77. For these replication studies, CrFK cells were transfected with pSV orf-A mutant proviruses that carried a 5′ chimeric LTR (SV40pr/R/U5) for expression of FIV virions containing genomes with orf-A mutations and were then cocultivated with either feline PBMC or MCH5-4 cells. PBMC or MCH5-4 cells were separated from transfected CrFK cells 24 h later and maintained in culture up to 27 days. As shown in Fig. 2A and C and Table 3, based on the lack of detection of FIV p24^gag in cell culture supernatants, replication of ΔorfA and ΔorfAmid mutant viruses in PBMC and MCH5-4 cells appeared to be completely restricted compared to that of WT FIV-pPPR. Expression of the structural viral proteins FIV Gag and Env was verified by immunocytochemical analysis of CrFK cells transfected with either pSVΔorfA or pSVΔorfAmid to rule out possible disruptions of splicing leading to the absence of rev or structural gene expression (data not shown).

FIG. 2.

Replication of pSV orf-A mutants in feline PBMC and MCH5-4 cells. CrFK cells were electroporated with pSV orf-A mutant provirus constructs or pSVWT and then cocultivated with either primary feline PBMC or MCH5-4 cells as described in the text. Cocultivated PBMC were separated from CrFK cells, maintained in culture up to 4 weeks posttransfection, and monitored by an FIV p24^gag antigen capture ELISA as described in Materials and Methods. Shown is the replication of all pSV orf-A-generated mutants in feline PBMC (A), of pSV orf-A mutants that are severely restricted in feline PBMC (B), of all pSV orf-A-generated mutants in feline MCH5-4 cells (C), and of pSV orf-A mutants that are severely restricted in MCH5-4 cells (D). Data are representative of three or more experiments.

TABLE 3.

Replication of FIV orf-A mutants^a

Name	Replicationb	Transcriptionc	Viral protein expressionc	Virus particle formationd	Infectivity
WT	++++	++++	++++	++++	++++
ΔorfA	—	ND	ND	ND	ND
ΔorfAmid	—	+++	+++	—	ND
orfA-trp43stop	— to +++e	ND	ND	ND	ND
orfA-L-A	+ (variable)	++++	++++	++++	—
orfA-W43A	+	ND	ND	ND	ND
orfA-W66A	++++	ND	ND	ND	ND
orfA-WW/AA	+	+++	++++	—	—
orfA-C-A	+	++++	++++	+	+

^a++++, equivalent to WT; +++, levels similar to WT; ++, intermediate levels; +, low levels; —, not detectable; ND, not determined.

^bOrfA mutant proviruses carrying a 5′ SV40-FIV LTR hybrid were used in PBMC and MCH5-4 replication studies as described in the text.

^cOrfA mutant proviruses carrying a WT 5′ FIV LTR were used in studies testing cell-associated viral mRNA by real-time RT-PCR and protein expression by immunoprecipitation and immunoblotting.

^dOrfA mutant proviruses carrying a WT 5′ FIV LTR were used for assay of virion-associated viral RNA produced in supernatants from provirus-transfected cells.

^eOrf-A mutation in orfA-trp43stop reverts to WT sequence.

To further demonstrate that orf-A is essential for replication in primary feline PBMC and lymphoid cell lines, pSVorfA-trp43stop was constructed to produce a virus carrying a truncated variant of orf-A that encodes only the N-terminal 42 amino acids and is similar to the orf-A variant carried by molecular clone FIV-34TF10 (43). The replication of the virus produced by transfection of pSVorf-A-trp43stop was severely restricted for as long as 15 to 20 days posttransfection in both PBMC and MCH5-4 cells (Fig. 2A and C; Table 3). However, by 24 days after transfection, FIV p24^gag production in cell culture supernatants increased. Assessment of orf-A nucleotide sequences from virions harvested from PBMC cell culture supernatants 30 days after transfection revealed a reversion of the stop codon at position 43 to a tryptophan codon. Reversion of the introduced mutation to the WT sequence suggested the presence of a selective pressure to maintain an intact orf-A gene. Taken together, these experiments demonstrated that a full-length orf-A is essential for efficient FIV replication.

Based on amino acid alignments of FIV Orf-A with other lentiviral Tat proteins, Orf-A is more similar to the ungulate lentiviral Tat proteins than to HIV or SIV Tat proteins (45). To characterize the functional significance of domains conserved between the CAEV and VV Tat proteins and FIV Orf-A, FIV proviruses encoding alanine substitutions for either leucines within the leucine-rich domain, cysteines within the cysteine-rich domain, or conserved tryptophans were constructed and tested for replication in feline PBMC and MCH5-4 cells (Fig. 2).

orfA-W66A was the only FIV orf-A mutant capable of virus replication comparable to that of WT FIV in both PBMC and MCH5-4 cells (Fig. 2A and C; Table 3). Although replication of all other orf-A point mutants was restricted, low concentrations of FIV p24^gag were usually still detectable. To demonstrate this very low yet measurable virus production, replication curves for specific orf-A mutants are presented in graphs that do not show data for WT FIV-pPPR (Fig. 2B and D). In contrast to the findings for the orfA-W66A mutant, virus production by the orfA-W43A and orfA-WW/AA mutants was severely restricted (Fig. 2B and D). No differences in replication between primary feline PBMC and MCH5-4 cells were observed for the orfA-W43A and orfA-WW/AA mutants (Table 3). These findings suggest that W43 is critical for optimal FIV replication whereas W66 is not.

Replication curves for orf-A mutant viruses produced by pSVorfA-C-A and pSVorfA-L-A were different in primary feline PBMC versus MCH5-4 cells. Virus replication of the orfA-C-A and orfA-L-A mutants was moderately reduced in PBMC compared to that of WT FIV (Fig. 2A and B). However, virus replication of the orfA-L-A mutant was either not detectable (Fig. 2C and D) or severely reduced (data not shown) in MCH5-4 cells. The replication of the orfA-C-A mutant was also severely restricted in MCH5-4 cells (Fig. 2C and D; Table 3). Results from these replication studies reveal that putative domains conserved between FIV Orf-A and ungulate lentivirus Tat proteins encode functions required for efficient virus replication. Moreover, replication data for the ΔorfAmid mutant indicate that amino acid residues 23 to 36, a region previously unmapped as a putative domain within Orf-A, are also critical for FIV replication.

Analysis of cell-associated FIV mRNA expression.

Previous studies used transient expression assays with reporter genes to investigate the role of Orf-A in viral LTR-directed gene expression and reported either a very small or a moderate effect of Orf-A on FIV LTR promoter activity (7, 12, 22, 41, 44, 50). For our studies, we used FIV proviruses carrying either deletions or point mutations within orf-A to investigate the requirement of Orf-A for FIV LTR-directed proviral gene expression. Either WT FIV-pPPR or FIV-pPPR orf-A mutant provirus plasmids carrying a WT FIV 5′ LTR (Table 1) were transfected into 3201-B cells and assayed for viral gene expression. The FIV mRNA copy number was quantitated from cellular RNA harvested from cells 20 h posttransfection by using a real-time RT PCR assay designed to measure gag-containing mRNA. Since 3201-B cells are nonpermissive for FIV-pPPR infection (11, 19), quantitated viral RNA represented viral RNA species expressed from transfected provirus only, rather than RNA from new virus infections of cells. All FIV-pPPR orf-A mutant constructs were cotransfected with pNDeGFP to normalize transfection efficiency, and cells transfected with pNDeGFP alone served as mock controls. Although replication of the ΔorfAmid mutant virus appeared completely restricted, the FIV-pPPRΔorfAmid provirus showed only a 3.8-fold decrease in the cell-associated viral RNA copy number relative to that with the WT FIV-pPPR provirus (Fig. 3A; Table 3). Repetition of this experiment revealed that the reduction in viral RNA expression by FIV-pPPRΔorfAmid compared to expression by WT FIV-pPPR was variable but was never greater than ninefold. In addition, the FIV-pPPRorfA-L-A and FIV-pPPRorfA-WW/AA proviruses showed very small decreases (1.3- and 1.8-fold, respectively) in viral RNA copy numbers relative to that with WT FIV-pPPR (Fig. 3B and D; Table 3). No decrease in viral RNA expression was detected after transfection with FIV-pPPRorfA-C-A compared to WT FIV-pPPR transfection (Fig. 3C; Table 3). These observations do not support restricted provirus-directed viral RNA expression as a cause for the severe restriction of virus replication observed for the orfA-WW/AA and ΔorfAmid mutants. In contrast, these results demonstrate that FIV-pPPR orf-A mutant proviruses express concentrations of viral RNA transcripts slightly less than or equal to those expressed by WT FIV-pPPR, and they suggest that orf-A encodes alternative functions in FIV replication.

FIG. 3.

Quantitation of intracellular FIV mRNA expressed by orf-A mutant proviruses. 3201-B cells were transfected with either an FIV-pPPR orf-A mutant provirus or WT FIV-pPPR. All proviruses were cotransfected with pNDeGFP to normalize transfection efficiencies. By 24 h posttransfection, total RNA was extracted from harvested cells and reverse transcribed to cDNA, and the copy number was determined by a real-time PCR assay as described in the text. The FIV RNA copy number per 200,000 GFP-positive cells was determined from cells transfected with FIV-pPPRΔorfAmid (A), FIV-pPPRorfA-L-A (B), FIV-PPRorfA-C-A (C), or FIV-pPPRorfA-WW/AA (D) and was compared to the FIV RNA copy number from cells transfected with WT FIV-PPR. Values shown are means of triplicate transfections except for the experiment showing FIV-pPPRΔorfAmid, which was performed in duplicate. Error bars, standard deviations. Data are representative of three experiments.

Quantitation of FIV particle-associated viral RNA in cell culture supernatants.

To investigate the role of orf-A in virion formation or release from cells, 3201-B cells were again transfected with either FIV-pPPR orf-A mutant provirus plasmids or WT FIV-pPPR. Virus particle formation and/or release into cell culture supernatants from transfected cells was measured based on quantitation of virion-associated FIV RNA by using a real-time TaqMan RT-PCR assay. Virion-associated viral RNA was not detected in supernatants harvested from cells transfected with either FIV-pPPRΔorfAmid or FIV-pPPRorfA-WW/AA (Fig. 4A, D, and E; Table 3). Transfection with FIV-pPPRorfA-C-A produced moderate amounts of viral RNA in cell culture supernatants, although a 70-fold decrease in viral RNA copy number was observed relative to that in supernatants harvested from WT FIV-pPPR transfection (Fig. 4C; Table 3). Based on our observations described above for FIV gene expression by orf-A mutants, the markedly reduced release of virion-associated viral RNA observed with transfections of ΔorfAmid and orfA-WW/AA proviruses was not due to a block in viral RNA transcription. Interestingly, transfection with FIV-pPPRorfA-L-A produced concentrations of virion-associated FIV RNA in cell supernatants similar to that measured for WT FIV-PPR transfection (Fig. 4B; Table 3). These observations suggest that Orf-A expression is required for some step in virus formation such as viral RNA encapsidation, particle assembly, release, or maturation and that multiple regions of Orf-A, including residues 23 to 36 (deleted in ΔorfAmid), the putative cysteine-rich domain, and residue W43, are important for this Orf-A function.

FIG. 4.

Orf-A mutation alters virus particle formation and is complemented by coexpression of GFP-Orf-A. 3201-B cells were transfected with either an FIV-pPPR orf-A mutant provirus or WT FIV-pPPR. All proviruses were cotransfected with pNDeGFP to normalize transfection efficiencies. Within 24 h posttransfection, total RNA was extracted from supernatants harvested from transfected cells, reverse transcribed to cDNA, and measured for FIV gag copy number by real-time RT-PCR as described in the text. Total RNA was also tested in the real-time PCR assay as a control for possible DNA contamination. The FIV RNA copy number per milliliter of cell culture supernatant from cells transfected with FIV-pPPRΔorfAmid (A), FIV-pPPRorfA-L-A (B), FIV-pPPRorfA-C-A (C), or FIV-pPPRorfA-WW/AA (D) was assayed and compared to the copy number per milliliter of supernatant from cells transfected with WT FIV-pPPR. Similarly, the FIV RNA copy number per milliliter of cell culture supernatant from cells transfected with either FIV-pPPRΔorfAmid or WT FIV-pPPR was assayed and compared to that measured from cells cotransfected with FIV-pPPRΔorfAmid and pEGFP-OrfA (E). 3201-B cells transfected with pNDeGFP alone served as a mock control. Data shown are representative of three or more experiments.

Expression of GFP-Orf-A complements an orf-A deletion mutant.

The observation that an FIV provirus with a deletion in orf-A could not produce detectable virion-associated viral RNA after transfection of cells presented an opportunity to assay the complementation activity of a recombinant Orf-A expressed as a fusion protein with GFP, designated GFP-Orf-A (Gemeniano et al., submitted). To test the functional activity of recombinant GFP-Orf-A, 3201-B cells were cotransfected with FIV-pPPRΔorfAmid and either pEGFP-Orf-A or pNDeGFP. Supernatants harvested from cell cultures posttransfection were then assayed for FIV particle formation based on the virion-associated RNA copy number as described above. As in the experiments described above, viral RNA was not detected in supernatants from cells cotransfected with FIV-pPPRΔorfAmid and the negative control pNDeGFP (Fig. 4E). However, concentrations of virion-associated RNA in supernatants from cells cotransfected with pEGFP-Orf-A and FIV-pPPRΔorf-Amid were comparable to those measured from cotransfection of WT FIV-pPPR and pNDeGFP (Fig. 4E). These findings demonstrate that GFP-Orf-A is functional and capable of complementing an FIV provirus with a deletion in orf-A. Importantly, these data also verify that the virus replication defects observed with an FIV provirus carrying a deletion in orf-A resulted solely from mutation of orf-A and not from changes in other FIV genes possibly affected by mutagenesis of the provirus.

Infectivity of orf-A mutant viruses.

To investigate a possible role for FIV Orf-A in virus infectivity, the cell-free virus infectivities of WT and orf-A mutant virus preparations for feline PBMC were tested. Virus infectivity was assayed by the detection and quantitation of FIV cDNA newly synthesized in cells inoculated with either WT or orf-A mutant virus inocula containing equal concentrations of virion-associated FIV p24^gag. FIV DNA copy numbers measured for cells inoculated with heat-inactivated preparations of either WT FIV or orf-A mutant virus inocula were usually below or close to the limits of detection of this assay and were subtracted as background from values measured for live virus infections. The FIV DNA copy number measured for orfA-C-A-inoculated cells was reduced 200-fold compared to that for cells infected with WT FIV-pPPR (Fig. 5B). However, FIV DNA was not detected in cells 24 h after inoculation with FIV-pPPRorfA-WW/AA or -orfA-L-A virus stocks (Fig. 5A and C; Table 3). These findings provide an explanation for the defective virus replication observed for the orfA-L-A mutant despite an absence of defects in viral RNA expression and virus particle release. Furthermore, these observations indicate that Orf-A is important for early steps of virus infection of PBMC involving either virus binding, entry, or reverse transcription, and they characterize the leucine-rich domain and residue W43 as possible orf-A-encoded determinants for virus infectivity.

FIG. 5.

Infection of feline PBMC with orf-A mutant viruses. Feline PBMC (5 × 10⁶) were infected with WT FIV or FIV orf-A mutant inocula containing 2.5 ng of virion-associated FIV p24^gag. Mock infections served as negative controls. At 24 h postinfection, the inoculum was removed, genomic DNA was extracted from harvested cells, and the copy number was determined by a real-time PCR assay as described in the text. FIV DNA copies per 5 × 10⁶ cells infected with FIV-pPPRorfA-L-A (A), FIV-pPPRorfA-C-A (B), or FIV-pPPRorfA-WW/AA (C) inocula are shown. Genomic DNA samples were normalized by quantitation of cellular CCR5 DNA copy number. Values shown are means of triplicate infections. Error bars, standard deviations. Data shown are representative of three experiments.

FIV p24^gag expression of FIV-PPR orf-A mutant proviruses.

To investigate a possible role for FIV Orf-A in viral protein expression, 3201-B cells were transfected with either WT FIV-pPPR or FIV-pPPR orf-A mutant provirus constructs and were cotransfected with pNDeGFP to normalize for transfection efficiency. Transfected cells were assayed for FIV p24^gag expression by immunoprecipitation and Western blot analysis. FIV p24^gag expression in cells transfected with WT FIV-pPPR was similar to that in cells transfected with FIV-pPPR orf-A mutants including ΔorfAmid, orfA-WW/AA, orfA-L-A, and orfA-C-A (Fig. 6; Table 3). Similarly, expression of FIV p55^gag was also comparable in cells transfected with WT FIV-pPPR versus FIV-pPPR orf-A mutants but was partially obscured by the immunoglobulin G heavy chain banding at a similar but slightly lower position in the SDS-polyacrylamide gel (data not shown). These findings revealed that an Orf-A-imposed restriction of viral protein expression is not a likely cause for the defective virus particle formation and restricted virus replication observed for FIV-pPPR mutants ΔorfAmid, orfA-WW/AA, orfA-L-A, and orfA-C-A. Moreover, this observation further supports consideration of Orf-A as an FIV accessory protein important for virus particle formation and infectivity.

FIG. 6.

Expression of FIV p24^gag by FIV-pPPR orf-A mutant proviruses. 3201-B cells were transfected with either WT FIV-pPPR (lane 1), FIV-pPPRΔorfAmid (lane 2), FIV-pPPRorfA-WW/AA (lane 3), FIV-pPPRorfA-C-A (lane 4), or FIV-pPPRorfA-L-A (lane 5). All proviral plasmids were cotransfected with pNDeGFP to normalize for transfection efficiency. Cells transfected with pNDeGFP alone served as a negative control. Within 48 h posttransfection, immunoprecipitations were performed on cellular extracts by using FIV-infected cat sera, and proteins were electrophoretically separated on SDS-12% polyacrylamide gels. Immunoblot analysis was performed by using a mouse monoclonal antibody against FIV p24^gag.

DISCUSSION

Previous studies testing FIV isolates carrying either a truncation, a deletion, or both within this accessory gene have demonstrated that orf-A is important for efficient viral replication in feline PBMC (9, 23, 32, 46, 50). For our studies, we first characterized the replication of the infectious molecular clone FIV-pPPR modified by either a large or a small deletion, or by point mutations within orf-A, in both feline PBMC and T-cell lines. In agreement with previously reported in vitro studies (11, 32, 46, 50), we showed that orf-A is essential for efficient replication in feline PBMC and a feline T-cell line. Replication of deletion mutants ΔorfA and ΔorfAmid in feline PBMC and MCH5-4 cells was completely restricted after transfection with mutant proviruses.

Orf-A domains critical for efficient viral replication in feline PBMC and T-cell lines have not been well characterized. Previous reports revealed that truncation of Orf-A resulting from a premature stop codon carried by molecular clone FIV-34TF10 at tryptophan residue 44 was responsible for the restricted replication observed for this virus in PBMC. In agreement with earlier studies, our in vitro studies demonstrated that the replication of a virus produced from an FIV-pPPR provirus encoding an Orf-A protein truncated at residue W43 (pSVorfA-trp43stop) was severely restricted in both PBMC and MCH5-4 cells (Fig. 2). Importantly, a reversion of the stop codon introduced at residue 43 to a tryptophan was associated with emerging virus production observed 24 days (or later) after transfection with pSVorfA-trp43stop. These findings demonstrate the presence of selective pressure to maintain an intact full-length orf-A gene, and they also indicate that the N-terminal acidic or hydrophobic region of Orf-A alone is not sufficient for efficient virus replication in feline PBMC.

Furthermore, replication studies of mutant viruses encoding point mutations in putative domains shared by FIV Orf-A and ungulate lentivirus Tat proteins revealed that the central leucine-rich and C-terminal cysteine-rich domains were necessary for efficient viral replication in feline PBMC and T-cell lines (Fig. 2; Table 3). In addition, our studies evaluated the previously unrecognized importance of tryptophan residues spaced 22 amino acids apart, at residues 43 and 66, for Orf-A-mediated viral replication. These conserved tryptophans are reminiscent of a WW domain, that is, a motif of 38 to 40 amino acids, distinguished by two highly conserved tryptophan residues spaced 20 to 22 residues apart, with a proline residue spaced 2 residues downstream of the first W residue (42). WW domains have been shown to be important for protein-protein interactions and have been identified in a variety of proteins including the signal-transducing factor YAP65 (Yes-associated protein), cytoskeleton components, ubiquitin-protein ligases (NEDD-4), nuclear proteins, and the putative transmembrane protein CD45-AP (33). Interestingly, our data indicate that W43 is absolutely critical for efficient virus replication in PBMC and MCH5-4 cells. In contrast, W66 is dispensable for replication, and the restriction for virus replication (at multiple steps as discussed below) observed for the orfA-WW-AA mutant most likely results from the W43 modification only. These findings suggest that W43 may be an important structural determinant rather than a component of a WW domain. However, additional studies will be necessary to characterize the function of this specific residue in Orf-A.

Mapping of orf-A sequences critical for efficient replication in feline PBMC and MCH5-4 cells was further defined by replication studies of the orf-A deletion mutant ΔorfAmid. This particular mutant encodes a 13-amino-acid deletion and a point mutation that includes the 4 terminal residues of the N-terminal acidic or hydrophobic domain, the first 3 residues of the leucine-rich region, and the 7 residues positioned between these two domains. The consistent failure to detect replication for the orf-A mutant ΔorfAmid was surprising and suggests that specific amino acid residues or a domain included within this deletion site is critical for either the function or the conformation of the protein. Hydrophobic alpha-helices and beta-sheets located within the N-terminal acidic or hydrophobic and leucine-rich domains, based on amino acid analysis (GeneWorks) (data not shown), would be disrupted by this 13-residue deletion. Alteration of this predicted secondary structure might explain the completely restricted virus replication observed for this mutant. Confirmation of stable protein expression of a GFP-ΔorfAmid fusion protein by cells transfected with the pEGFP-ΔorfAmid expression plasmid (data not shown) argues against complete abolishment of Orf-A expression as a possible cause for the restricted replication observed for this mutant. Point mutations within this deletion site will be necessary to further define the critical replication determinants deleted within this mutant.

We also investigated specific steps in the viral replication cycle that may be affected by orf-A mutations. Because Orf-A has been reported to be a weak transactivator of the FIV LTR (7, 12), we examined the effect of Orf-A on FIV-LTR-directed proviral gene expression. A relatively small three- to ninefold reduction in the cell-associated FIV mRNA copy number was found for cells transfected with the mutant FIV-pPPRΔorfAmid provirus relative to the copy number in cells transfected with WT FIV-pPPR (Fig. 3A). A reduction of twofold or less in viral RNA expression was observed with transfection of the orfA-WW/AA, orfA-L-A, and orfA-C-A mutant proviruses (Fig. 3B to D; Table 3). These findings demonstrate that Orf-A only weakly upregulates viral RNA expression (Fig. 3), and they are in agreement with previous reports describing a very weak transactivating activity by Orf-A on the FIV LTR (22, 41, 44, 50). This small effect on FIV LTR-directed proviral gene expression would not likely account for the severe restriction of virus replication observed after transfection with these mutant proviruses. Instead, these data suggested that Orf-A might express alternative functions for FIV replication and pathogenesis.

Virus particle formation from cells transfected with orf-A mutant proviruses was examined. With the exception of the orfA-L-A mutant, all orf-A mutant proviruses that were tested were moderately or severely restricted in the ability to produce cell-free virion-associated viral RNA (Fig. 4). Although viral mRNA expression from the ΔorfAmid mutant provirus was only slightly reduced relative to that from the WT, cell-free virus-associated RNA was undetectable at 20 h after transfection with this mutant (Fig. 4A and E). Similarly, production of cell-free viral RNA was either undetectable (orfA-WW/AA) or moderately reduced (orfA-C-A) for FIV mutant proviruses (Fig. 4C and D; Table 3). However, although cell-free viral RNA was not detected in supernatants after transfection with the orfA-WW/AA mutant provirus, replication studies of this mutant in feline PBMC revealed very low but detectable virus production over time in culture (Fig. 2). These observations suggest that virus particle formation is severely, but not completely, restricted for orfA-WW/AA. Interestingly, the orfA-L-A mutant provirus consistently produced detectable quantities of cell-free viral RNA comparable to those measured for the WT FIV provirus (Fig. 4B). Observations from Western blot analysis (Fig. 6) of viral protein expression in transfected cells revealed that mutant orf-A provirus constructs were capable of expressing viral proteins, i.e., FIV p24^gag. These findings ruled out the likelihood that the defect in virus particle formation associated with orf-A mutation was due to a block in viral protein expression, and they suggested that either viral RNA encapsidation or virus particle assembly, release, or maturation was affected by Orf-A mutation. Collectively, these data suggest that virus particle formation is a major function for Orf-A and that this activity is affected by multiple sequences within Orf-A, including amino acid residues 23 to 36, residue W43, and the cysteine-rich domain. Rescue of virus particle formation by cotransfection of the ΔorfAmid deletion mutant with a mammalian expression plasmid for Orf-A (Fig. 4E) further demonstrates that virus particle formation or release is an important function of Orf-A in FIV replication and confirms functional activity for Orf-A transiently expressed as a GFP-Orf-A fusion protein.

Data from the present study also revealed that virus particle infectivity was affected by mutations within the FIV orf-A gene. Virus preparations of orfA-L-A and orfA-WW/AA mutants showed severely restricted abilities to infect feline PBMC (Fig. 5A and C; Table 3). The extreme reduction in virus infectivity observed for the orfA-L-A mutant virus most likely accounted for its severely restricted virus replication, since no defects in expression of cell-associated FIV RNA, viral proteins, or cell-free virions were observed for this mutant. Based on these observations, virus particle infectivity is another function expressed by Orf-A, and leucine residues within the leucine-rich domain, as well as the W43 residue, are critical for this function. The orfA-C-A mutant virus showed a moderate restriction of virus infectivity (Fig. 5B), just as it exhibited a moderate reduction in virion-associated viral RNA release. Taken together, the findings for the orfA-C-A mutant suggest that mutation of the cysteine residues may produce a structural or conformational modification that moderately affects multiple functional domains of Orf-A associated with virus particle formation and infectivity. Virus infectivity could not be tested for the ΔorfAmid mutant because transfection of this mutant provirus never produced detectable virus even with prolonged passage of either PBMC or MCH5-4 cells after transfection and cocultivation.

Taken together, our observations regarding virus particle formation and release and virion infectivity for this particular panel of FIV orf-A mutants suggest that multiple components of the orf-A gene product, including residues 23 to 36 and leucine-rich and cysteine-rich domains, represent either critical functional or critical conformational domains. Interestingly, residue W43 also proved to be critical for multiple functions of Orf-A, whereas W66 was dispensable. Reversion of the orfA-trp43stop mutant to WT Orf-A after 3 weeks of passage suggests that some degree of replication of this mutant virus occurs, and stable expression of this truncated Orf-A variant as a GFP-Orf-Atrp43stop fusion protein has been verified (Gemeniano et al., submitted). However, expression of the N-terminal acidic or hydrophobic region alone as encoded by the orfA-trp43stop mutant is not sufficient for efficient replication. Additional mutagenesis studies will be necessary to further define specific functional domains of FIV Orf-A. Studies to identify cellular proteins that interact with Orf-A will also be necessary to determine the mechanism by which these Orf-A domains function.

In addition, our observations also demonstrate that this accessory gene affects later steps of virus replication such as virus particle formation or particle release, as well as early steps involving either virus binding, entry, or RT. These specific functions have not been reported previously for this FIV-encoded gene and represent novel findings for this animal model of HIV infection. These findings plus the absence of a significant disruption in expression of viral transcripts or viral proteins with our orf-A mutant proviruses do not support the classification of this gene as an analog to the primate lentivirus tat genes. Instead, our observations suggest that Orf-A may express functions representative of multiple HIV-encoded accessory proteins. Specifically, the effect of orf-A mutation on virus particle formation implies similarities to HIV-1 Vpr, Vpu, and Nef, whereas the importance of Orf-A for virus infectivity suggests similarities to HIV-1 Vpr and Nef. Studies characterizing specific steps involved in virion formation (RNA encapsidation, assembly, budding, release, and maturation) and infection (entry, uncoating, and RT) that are affected by orf-A mutation will further elucidate the function of this gene that is necessary for efficient FIV replication. These future studies should also more clearly characterize the similarity of orf-A with specific accessory genes carried by primate lentiviruses.