TEXT

Top Abstract Text References

Feline immunodeficiency virus (FIV) is a lentivirus that induces a fatal immunodeficiency syndrome in infected cats. Disease is characterized by depressed CD4/CD8 T-cell ratios, wasting and emaciation, and opportunistic infections (see reference 2 and references therein). FIV has demonstrated a tropism in vitro and in vivo for multiple lymphoid subsets, including CD4 T cells, and for macrophages (1, 3, 9, 10, 13, 14). FIV utilizes the chemokine receptor CXCR4, a molecule also identified as a coreceptor for T-cell line tropic isolates of human immunodeficiency virus (HIV), for infection of feline cells (31, 33, 41). CD4 T-cell depletion is the hallmark of immunodeficiency associated with FIV infection in cats and is accompanied by other immunologic abnormalities, including T-cell abnormalities that are also characteristic of HIV infection in humans (2).

Similar to other lentiviruses, FIV gene expression is dictated by virus-encoded factors, including Rev, a posttranscriptional regulator encoded by viral genes orfL and orfH, and a putative viral transactivator encoded by the tat-like gene orfA (also designated orf2) (11, 23, 28, 29, 34, 36-39). Although data from earlier studies, using in vitro transient expression assays, have not strongly supported a role for orfA as a viral transactivator, a recent report provided stronger evidence that the orfA gene product may indirectly up-regulate FIV long terminal repeat (LTR)-directed transcription (11), possibly by interactions with cellular transcription factors. Cellular proteins that interact with OrfA to enhance FIV LTR-directed expression have not yet been identified.

cis-acting response elements in the LTR of the proviral DNA bind cellular transcription factors to mediate lentiviral gene expression. Binding sites for cellular transcriptional factors, such as NF- and SP-1, in the U3 domain of the HIV LTR are critical for efficient viral transcription (6, 7). The U3 domain of the FIV LTR has consensus recognition sequences for the cellular transcription factors AP-1, AP-4, ATF (also known as the cyclic AMP [cAMP] response element, or CRE), NF1, and NF- (see reference 20 and references therein and references 25, 28, 30, and 37). Of these putative target sequences, AP-1, AP-4, ATF, and C/EBP recognition sites have been shown to bind cellular proteins by DNase I footprinting and gel shift assays (17, 37) and appear to be important for basal promoter activity of the FIV LTR in vitro (22, 26, 34, 37). Furthermore, the AP-1 site is required for T-cell activation responses mediated by protein kinase C, as well as activation by c-Fos. The ATF site is required for cAMP-induced responses mediated by protein kinase A (18, 26, 34).

To assess the role of these FIV LTR cis-acting response elements in virus expression and replication, we constructed and characterized in vitro virus expression and production from FIV LTR mutant proviruses encoding U3 deletions of either an AP-1 site, an ATF site, or both sites. In addition, a mutant containing a 72-bp deletion encompassing the AP-1, AP-4, duplicated C/EBP, NF1, and ATF enhancer recognition sequences was constructed and tested. Using both transfection and infection approaches, we found that removal of the ATF site severely reduced virus expression and replication, whereas deletion of the AP-1 site produced only a slight reduction in virus expression. Both an FIV LTR mutant encoding deletions of both the AP-1 and ATF sites and the mutant containing a 72-bp deletion spanning the AP-1 and ATF sites were completely restricted for virus expression. Replacement of the 5' LTR with a chimeric promoter encoding the simian virus 40 (SV40) early promoter enhancer and TATA sequences upstream of the RU5 domain in the LTR mutant provirus restored transient provirus expression to allow production of LTR mutant virus stocks for infection studies.

For construction of FIV AP-1 and ATF deletion mutants, 4 nucleotides (nt) were deleted from the AP-1 response element and 5 nt were deleted from the ATF response element within the U3 domain of the molecularly cloned FIV-PPR provirus (FIV subtype A) (30) (Fig. 1) by using site-directed PCR overlap mutagenesis (16) with PCR primers described in Table 1. An LTR deletion mutant encoding a 72-nucleotide (nt) deletion (nt 92 to 163 of the 5' LTR) encompassing the AP-1, AP-4, duplicated C/EBP, and ATF recognition sequences (Fig. 1) with an insertion of a mutated AscI restriction enzyme site (5' TGCGCGCC 3') was generated by PCR and designated the 4 mutation. To generate AP-1 and ATF mutations within the U3 region of the 3' LTR, a subgenomic fragment of the FIV-pPPR provirus digested with NdeI (nt 8900) and SalI (3' polylinker site) and containing orfH and the entire 3' LTR was cloned into the plasmid pGEM-5Zf+ (Promega Biotech, Madison, Wis.). The resulting construct was named pNS5 and was used as a template for site-directed deletion mutagenesis.

View larger version (13K): [in this window] [in a new window]   FIG. 1.   Mutations in FIV-pPPR cis-acting transcriptional control elements. Mutations are described by nucleotide sequence and location in the 5' U3 domain of the FIV-pPPR provirus (30) and identified by their designated mutant virus. Dots indicate sequence identities, and dashes represent deleted nucleotides.

Terminal primers LTR-A, containing sequence upstream of the 3' LTR (orfH) with an internal NdeI site, and LTR-D, encoding the 3' terminus of the 3' LTR with a flanking SalI site, were used with mutagenic primers LTR-B-AP-1mut and LTR-C-AP-1mut to PCR amplify a 3' LTR with an AP-1 site deletion. Accordingly, an ATF site-deleted 3' LTR was amplified with primers LTR-A and LTR-D and mutagenic primers LTR-B-ATFmut and LTR-C-ATFmut. The amplified AP-1 and ATF mutant LTRs were cloned back into pNS5 to generate plasmids pNSAP-1 and pNSATF, respectively. By the same approach, plasmid pNSAP-1 was used as a template for construction of a 3' LTR encoding both AP-1 and ATF deletions and generation of plasmid pNSAP-1/ATF. For construction of the 4 mutation, a 344-bp fragment encoding upstream sequence flanking the 72-bp deletion site (nt 9205 to 9278) was amplified by PCR from wild-type (WT) FIV-pPPR with primers LTR-A and LTR-B4mut, which introduced 3' flanking SalI and AscI restriction enzyme sites, and cloned into pGEM-5Z f+ to generate plasmid pGEM-F1. Next, a 215-bp fragment (nt 9278 to 9468) encoding the LTR immediately downstream of the deletion site was PCR amplified with primers LTR-C4mut and LTR-D and cloned into pGEM-F1 with AscI and SalI sites to produce plasmid pGEM-LTR4.

To facilitate insertion of mutant LTRs into an FIV provirus, the FIV-PPR provirus was cloned out of FIV-pPPRpUC (30) and into vector pGEM-9Zf() (Promega Biotech) to generate an infectious FIV-PPR provirus construct designated pPPR WT. The NdeI-SalI fragment from plasmids containing a mutated 3' LTR (pNS5 or pGEM-LTR4) was inserted into pPPR WT to generate an FIV-pPPR provirus with a mutated 3' LTR. To prevent generation of WT sequences by recombinatorial events during virus replication, FIV-pPPR mutants encoding deletions within both the 5' and 3' LTRs were constructed and designated type 1 LTR mutants. To complete construction of a type 1 LTR mutant, each mutant 3' LTR was amplified by PCR with primers LTR-Kas1, which spans the 3' terminus of U5 and the primer binding site of the viral genome, and LTR-Spe1 encoding the 5' terminus of the LTR with a flanking SpeI site. Mutant 5' LTRs were cloned into FIV-pPPR proviruses with a mutated 3' LTR to generate FIV-pPPR type 1 LTR mutant proviruses, pPPRAP-1, pPPRATF, pPPRAP-1/ATF, and pPPR4 (Fig. 2A). All mutant LTRs were assessed by dideoxynucleotide sequencing.

View larger version (24K): [in this window] [in a new window]   FIG. 2.   Construction of FIV-pPPR mutant proviruses. Mutations within all LTR mutant proviruses were confirmed by dideoxynucleotide DNA sequencing. (A) Structure and genomes of type 1 FIV-pPPR LTR mutant proviruses. (B) Design of the chimeric SV40pr/RU5 5' LTR. Construction is described in the text. (C) Structure and genomes of type 2 FIV-pPPR LTR mutant proviruses.

To assess virus expression and production following transfection of FIV-pPPR type 1 LTR mutant proviruses, Crandell feline kidney cells (CrFK cells) (a feline adherent cell line) were electroporated with 10 µg of proviral plasmid DNA and cocultivated with primary feline peripheral blood mononuclear cells (PBMCs) harvested from whole blood drawn from specific-pathogen-free (SPF) cats and cultured as described in previous reports (9, 34). Cocultivated feline PBMCs were separated from transfected CrFK cells 24 h later and maintained in culture for 14 to 21 days. Supernatant was collected from all infected cell cultures every 3 to 4 days for up to 3 weeks after cocultivation and assayed for virus production by detection of FIV p24^gag with either an FIV p24^gag antigen-capture enzyme-linked immunosorbent assay (ELISA) previously described (8) or a commercial FIV p24^gag antigen-capture ELISA (Idexx Corp., Westbrook, Maine). Virus expression and production as determined by FIV p24^gag production were severely restricted after transfection of FIV-pPPR proviruses encoding a mutant 5' LTR with a deletion of the ATF response element (Fig. 3A and B). Virus production was not detected in PBMCs cocultivated with CrFK cells after transfection with pPPRAP-1/ATF (Fig. 3A and B) and pPPR4 (Fig. 3B) and was barely detectable after transfection with pPPRATF. In contrast, virus expression and production were detected following transfection of the AP-1 deletion mutant, pPPRAP-1, although virus production was delayed when compared to transfection of pPPR WT. High concentrations of viral antigen were observed in supernatants from PBMCs by 7 to 10 days after transfection with pPPR WT. These observations suggested that the ATF or cAMP response element within the FIV U3 domain is critical for efficient provirus expression and that the AP-1 site is less critical for efficient FIV expression.

View larger version (26K): [in this window] [in a new window]   FIG. 3.   Expression of FIV LTR mutant proviruses posttransfection of CrFK cells and cocultivation with feline PBMCs. CrFK cells were transfected with 10 µg of plasmid DNA and cocultivated with PBMCs 24 h later, as described in the text. Virus production was measured in cell culture supernatants every 3 to 5 days with an FIV p24gag antigen-capture ELISA. Expression of type 1 FIV-pPPR LTR mutant proviruses (A and B) and type 2 mutant proviruses (C and D) after cocultivation is shown. The data shown are representative of three or more transfection experiments.

To generate FIV LTR mutant virus stocks of sufficient titer for infection and replication assays, an FIV provirus driven by a chimeric 5' LTR encoding the SV40 early promoter enhancer and TATA sequences was constructed. The RU5 region, including nt 1 to 16 of the U3 region of the 5' LTR of the FIV-pPPR provirus, was PCR amplified with primers RU5-Bgl2 and RU5-Kas1 and cloned into a transient chloramphenicol acetyltransferase (CAT) expression vector (p22A2s) (34) by using restriction enzyme sites BamHI and PstI to create the construct U3-CAT. Next, the SV40 early promoter, including enhancer sequences and a TATA box, was PCR amplified with primers SV40-SpeI and SV40-BglII and cloned into plasmid U3-CAT with restriction enzyme sites BglII (3' terminus of the SV40 sequence) and SacI (introduced at the 5' terminus by PCR primer) to produce recombinant plasmid pSVRU5-CAT containing the SV40pr/RU5 hybrid promoter (Fig. 2B). The SV40pr/RU5 hybrid promoter was inserted into the FIV-pPPR WT provirus to replace the 5' LTR and produce the provirus construct pSV-pPPR WT (pSVWT) (35) (Fig. 2B). Subsequently, mutated LTRs replaced WT 3' LTRs in pSVWT by using restriction enzyme sites NdeI and SalI to produce type 2 FIV-pPPR LTR mutants (Fig. 2C). The presence of the SV40pr/RU5 hybrid promoter and LTR deletions in all proviral constructs was confirmed by dideoxynucleotide DNA sequencing (U.S. Biochemicals, Cleveland, Ohio).

The chimeric SV40pr/RU5 5' LTR included bp 1 to 16 of the U3 region to preserve the normal spacing between the SV40-derived TATA box and the first base of U5 and to maintain the FIV mRNA cap site (Fig. 2B). The SV40pr/RU5 5' LTR was functional as a promoter when the construct pSVRU5-CAT was tested in transient expression assays (data not shown). Replacement of the WT 5' LTR within pPPR WT with the chimeric SV40pr/RU5 5' LTR produced an FIV provirus (pSVWT) capable of replication and virus production similar to those of pPPR WT (Fig. 3C) after transfection of CrFK cells and cocultivation with feline PBMCs. Levels of virus expression and production, as determined by FIV p24^gag production after transfection with pSVAP-1, were comparable to those observed for pPPR WT and pSVWT, whereas the level of virus production after transfection with pSVATF and pSVAP-1/ATF was reduced compared to that of WT viruses (Fig. 3C). Although virus production was extremely restricted after transfection of pSV4 (Fig. 3D), approximately 50% of the pSV4 transfections yielded low concentrations of virus detectable by FIV p24^gag antigen-capture ELISA. Accordingly, substitution of the mutant 5' LTR in LTR mutants pPPRAP-1, pPPRATF, and pPPRAP-1/ATF with the chimeric SV40pr/RU5 produced proviruses capable of sufficient virus production to generate LTR mutant virus stocks for replication studies.

To assess replication of FIV LTR mutant viruses in both primary feline PBMCs and macrophages, titered virus stocks were prepared by transfection of CrFK cells with either pPPR WT, pSVWT, or the type 2 FIV LTR mutant proviruses and cocultivation with SPF feline PBMCs for short-term passage (21 days or less) as previously described (9, 35). To confirm the retention of specific U3 deletions within FIV LTR mutant virus stocks, genomic DNA was extracted from mutant virus-infected PBMCs with a commercial kit (QIAamp; Qiagen, Chatsworth, Calif.). LTR sequences were amplified by a single round of PCR for 30 cycles (1 min of template denaturation at 94°C, 1 min of primer annealing at 55°C, and 1 min of primer extension at 72°C) in a thermal cycler (Perkin-Elmer Biosystems, Foster City, Calif.) using PCR primer LTR3.19 and either LTR-2 or LTR8.21. PCR products containing LTR sequences were then sequenced directly by dideoxynucleotide DNA sequencing.

For virus infection studies, duplicate wells in a six-well microtiter plate were seeded with 4 × 10⁶ PBMCs and inoculated with 10² 50% tissue culture infective doses (TCID₅₀) of a specific LTR mutant virus stock or with uninfected tissue culture media (mock infection control). For feline macrophage infections, feline monocyte-derived macrophages (MDMs) were prepared from feline PBMCs and cultivated in 24-well tissue culture plates as previously described (9). Duplicate wells (approximately 10⁵ MDMs) of the 24-well plate were inoculated with 10³ TCID₅₀ of a virus stock approximately 7 days after cultivation or were mock infected. Cells (PBMCs and macrophages) were incubated with virus inocula overnight, washed with Hanks buffered salt solution the following day, and fed fresh tissue culture media. Similar to transfection experiments, infected cell cultures were fed fresh media, and supernatant was collected every 3 to 4 days for up to 2 weeks after infection to assay virus production by FIV p24^gag antigen-capture ELISA.

Replication of virus prepared from pSVAP-1 was comparable to that observed for pPPR WT and pSVWT, whereas virus production postinfection of feline PBMCs with pSVATF and pSVAP-1/ATF was delayed and reduced (Fig. 4A). Virus production from feline PBMCs infected with pSV4 was severely restricted, with supernatant concentrations of FIV p24^gag below the level of detection of the assay until 14 days after infection (Fig. 4B). Similar to observations from PBMC infection studies, virus production after infection of primary feline MDMs with pSVATF and pSVAP-1/ATF was delayed and reduced when compared to the WT and pSVAP-1 virus stocks (Fig. 4C). Although the virus production observed from pSVAP-1-infected MDMs was greater than that for either WT FIV virus stock in the experiment illustrated in Fig. 4C, data from other macrophage experiments revealed that pSVAP-1 replication was very similar to that of WT virus (data not shown). Probable causes for the higher virus production from the AP-1 deletion mutant in this experiment are the variability associated with PBMC-derived macrophage preparations from different donor cats and the variation in adherence of monocytes and differentiation to macrophages observed from well to well for the same PBMC preparation. Due to the extremely low titer of the pSV4 virus stock, replication of this mutant was not tested in feline MDMs.

View larger version (18K): [in this window] [in a new window]   FIG. 4.   Replication of FIV LTR mutant viruses in feline PBMCs and MDMs. Feline PBMCs (4 × 106) were infected with 102 TCID50 of either WT FIV-pPPR, or pSVWT, an LTR mutant virus stock generated from type 2 LTR mutant proviruses, or were mock infected, as described in the text (A and B). Feline MDMs were inoculated with 103 TCID50 of a virus stock approximately 7 days after cultivation or were mock infected (C). Inocula were removed 24 h postinfection, and cell culture supernatants were assayed every 3 to 5 days by using an FIV p24gag antigen-capture ELISA. The data shown are representative of two or more infection experiments.

Comparable to the results observed in the transfection studies, data from the PBMC and MDM infection studies indicated that the ATF site was necessary for efficient FIV expression and replication, whereas the AP-1 response element was dispensable for in vitro replication in PBMCs and macrophages. The observation that the AP-1 site was not essential for FIV replication in vitro has been observed previously in a report examining an AP-1 mutant characterized by a 31-bp deletion that removed an AP-1 and an AP-4 site and that was constructed with the FIV molecular clone TM2 (FIV subtype C) (26). The same FIV TM2 AP-1 deletion mutant was, however, slightly attenuated or restricted for virus replication in vivo (19). The dispensability of this transcriptional element does not correlate with observations from previous transient expression assay studies indicating that the AP-1 site was necessary for optimal basal promoter activity and for either phorbol ester- or fos-induced activation of the FIV LTR (26, 34). An AP-1 site encoded by the visna virus LTR was critical for basal activity and for transactivation of the visna virus LTR resulting from interactions of the visna virus Tat protein with cellular transcription factors Fos and Jun (4, 15, 27). Because the structure of OrfA, the putative FIV Tat protein, is very similar to that of visna virus Tat, the AP-1 site may still prove to play a critical role in FIV expression or replication in specific lymphoid subsets or stages of cell differentiation, and studies addressing these issues have not yet been reported.

In contrast, this is the first report describing expression and replication of FIV LTR mutants encoding a deletion of the ATF-binding element and the first study to evaluate replication of LTR mutants in feline macrophages. Deletion of the ATF site markedly reduced viral gene expression and virus production in both primary feline PBMCs and MDMs. Virus expression after transfection with type 1 ATF deletion mutants, including pPPRATF and pPPRAP-1/ATF, was negligible, and preparation of these viruses required transfection of type 2 ATF deletion mutants (pSVATF and pSVAP-1/ATF). Differences were not observed between the expression and replication of mutants encoding both an ATF and AP-1 deletion and those of mutants encoding an ATF deletion only. These findings correlate with those of previous studies that found the ATF response element critical for FIV LTR basal activity, LTR activation by cAMP analogs, and binding of cellular proteins by DNase I footprinting and gel shift assays (17, 18, 34, 37). The importance of the ATF-binding site for virus replication suggests that FIV expression may be at least partially regulated by the ATF/CREB family of transcription factors that bind ATF or cAMP response elements. The LTRs of the retroviruses human T-lymphotropic virus type 1 and bovine leukemia virus also contain sequences homologous to cellular cAMP response elements (12, 24, 40, 42) and are regulated by interactions of their respective viral transactivators (Tax) with transcription factors of the ATF/CREB family. Characterization of cellular proteins that bind either the FIV LTR or the putative FIV Tat protein (OrfA) will be necessary to further define the role of the ATF-binding site in FIV gene expression and replication.

Removal of both the AP-1 and ATF sites coupled with deletion of the intervening transcriptional elements (AP4, C/EBP, and NF1 sites) produced LTR mutants almost completely restricted for expression, including the type 2 mutant pSV4. The reduced expression and replication demonstrated by pSV4-generated virus compared to those of pSVAP-1/ATF indicates that the additional deleted sites contribute to FIV LTR enhancer or promoter activity. Based on previous findings that demonstrated that the duplicated C/EBP sites are critical for FIV LTR promoter activity (22, 37), inclusion of the C/EBP site deletion is the most likely source for the increased attenuation of the 4 LTR mutant. In contrast, the AP-4 site and NF1 sites were not critical for either basal promoter or activation activity of the FIV LTR in earlier studies and may have contributed little to the attenuation of the 4 LTR mutant (11, 34, 37). However, possible effects on surrounding U3 DNA sequence and/or structure resulting from such a large deletion encoded by the 4 mutant may also play a role in the severe attenuation of this mutant, as well as the absence of multiple functional domains. Construction and testing of FIV LTR mutants with a deletion restricted to either the duplicated C/EBP sites, the AP4 site, or the NF1 site would be necessary to confirm the role of these sites in the attenuated replication observed for the 4 LTR mutant.

Generation of infectious FIV and HIV proviruses by replacement of the WT 5' LTR with a chimeric LTR including the human cytomegalovirus immediate-early gene promoter (CMV promoter/RU5) has been described previously, and the resulting chimeric proviruses demonstrated altered cell tropisms, which are especially crucial to development of FIV vectors capable of expression in human cells (5, 21, 32). In this work, we replaced the mutant 5' LTR within the type 1 LTR mutant proviruses with a chimeric SV40pr/RU5 5' LTR to enhance virus expression and facilitate production of virus stocks from proviruses severely restricted for virus expression, including the ATF, AP-1/ATF, and 4 mutants. Experiments to characterize the range of host cells permissive for the expression of pSVWT were not conducted in this study, but would be necessary to determine the utility of this construct for FIV vector development.

Observations from these studies identify cis-acting elements within the FIV U3 domain that contribute to efficient viral replication in vitro in both feline lymphocytes and macrophages and demonstrate that deletion of these elements produces an attenuated virus. These attenuated FIV mutants provide opportunities to characterize the role of these cis-acting elements in virus replication in vivo and establishment of virus load. These mutants also provide opportunities to test LTR mutants as attenuated virus vaccines, including (i) moderately attenuated mutant proviruses that encode small, 4- to 5-bp deletions in their respective target sequences and (ii) a severely attenuated LTR mutant that encodes a 72-bp deletion removing multiple transcription factor-binding sites. Studies testing replication and virus load in cats inoculated with these mutants are currently in progress.