Characterization of a Highly Pathogenic Molecular Clone of Feline Immunodeficiency Virus Clade C

Abstract

We have derived and characterized a highly pathogenic molecular isolate of feline immunodeficiency virus subtype C (FIV-C) CABCpady00C. Clone FIV-C36 was obtained by lambda cloning from cats that developed severe immunodeficiency disease when infected with CABCpady00C (Abbotsford, British Columbia, Canada). Clone FIV-C36 Env is 96% identical to the noninfectious FIV-C isolate sequence deposited in GenBank (FIV-Cgb; GenBank accession number AF474246) (A. Harmache et al.) but is much more divergent in Env when compared to the subgroup A clones Petaluma (34TF10) and FIV-PPR (76 and 78% divergence, respectively). Clone FIV-C36 was able to infect freshly isolated feline peripheral blood mononuclear cells and primary T-cell lines but failed to productively infect CrFK cells, as is typical of FIV field isolates. Two-week-old specific-pathogen-free cats infected with FIV-C36 tissue culture supernatant became PCR positive and developed severe acute immunodeficiency disease similar to that caused by the uncloned CABCpady00C parent. At 4 to 5 weeks postinfection (PI), 3 of 4 animals developed CD4⁺-T-cell depletion, fever, weight loss, diarrhea, and opportunistic infections, including ulcerative stomatitis and tonsillitis associated with abundant bacterial growth, pneumonia, and pyelonephritis, requiring euthanasia. Histopathology confirmed severe thymic and systemic lymphoid depletion. Interestingly, the dam also became infected with a high viral load at 5 weeks PI of the kittens and developed a similar disease syndrome, requiring euthanasia at 11 weeks PI of the kittens. This constitutes the first report of a replication-competent, infectious, and pathogenic molecular clone of FIV-C. Clone FIV-C36 will facilitate dissection of the pathogenic determinants of FIV.

Feline immunodeficiency virus (FIV) causes an AIDS-like syndrome in the domestic cat (12, 17, 31), with marked similarity to AIDS caused by human immunodeficiency virus (HIV) in humans. FIV infection in cats typically manifests itself as a transient acute-phase syndrome, characterized by febrile episodes, lymphadenopathy, neutropenia, and weight loss. This initial phase is followed by a protracted asymptomatic period with progressive loss of CD4⁺ T lymphocytes and a terminal AIDS phase characterized by succumbing to opportunistic infections (31, 33, 55, 56). Considerable variability exists among FIV strains, not only in sequence relatedness but also in pathogenicity in vivo (8, 11, 15, 32, 38, 44, 57).

Following criteria similar to those for HIV subgroup distinction, FIV has been classified into subgroups according to env diversity (42). Sodora et al. (42) defined the original Petaluma strain (30, 46) as a prototype subgroup A and the divergent Japanese TM2 strain (27) as prototype subgroup B. A majority of FIV sequences obtained worldwide were categorized in the A and B subgroups, and these groups vary widely in degrees of pathogenicity in vivo (8, 11, 15, 32, 38, 44, 57). A subtype C was also defined and is the least prevalent FIV subgroup (1, 42). One clade C representative, CABCpady00C, exhibits greater pathogenicity than FIV isolates of subtypes A and B. In initial studies with this isolate, 60% of infected animals developed severe acute immunodeficiency syndrome and opportunistic infection requiring euthanasia by 8 to 12 weeks postinfection (11). Defining the molecular basis for the pathogenicity of clade C isolates is important to the development of the FIV model. Full-length sequences, derived by the Mullins laboratory employing PCR technology, were used to define the C subtype (GenBank accession no. AF474246; termed FIV-Cgb herein) and served as a template for the development of FIV clade C (FIV-C)-specific probes and primers. However, we were unable to generate infectious virus from this clone and, therefore, resorted to lambda cloning strategies to obtain a full-length infectious clone of FIV-C. Here we report the molecular cloning, sequencing, and in vivo analysis of an infectious, pathogenic FIV-C molecular clone, FIV-C36 (GenBank accession number AY600517).

MATERIALS AND METHODS

Animals and in vivo infections.

Two-week-old specific-pathogen-free (SPF) cats supplied by the Andrea D. Lauerman SPF cat colony at Colorado State University (CSU) were housed and infected with FIV-C36 containing cell supernatant (1 ml intravenously and 0.75 ml orally at 10^3.4 50% tissue culture infective doses/ml) at CSU in accord with procedures approved by the CSU Animal Care and Use Committee and U.S. Department of Agriculture protocol.

Virus source and in vitro infections.

Plasma (0.1 ml) from a clinically symptomatic cat (no. 185) infected with the CABCpady00C virus (11) was used to infect ∼5 × 10⁶ 104-C1 cells. Total genomic DNA was extracted from the 104-C1 cells 7 days postinoculation (PI) for subsequent use in making the phage library (see below).

Genomic library screen.

FIV-C genomic DNA was isolated as described above, digested with BamHI, and then ligated into BamHI of the bacteriophage λ vector, EMBL-3 (Stratagene). Packaging was performed according to manufacturer's instructions and as previously described for cloning of clade A FIVs (35, 46). The library was amplified, titers were determined, and plaque lifts were screened by Southern blotting with two different 5′ FIV-C probes. The first probe was a 1.4-kb SacI-XhoI fragment spanning from the 5′ long terminal repeat (LTR) to the middle of the gag gene. The second probe was a 700-bp XhoI-EcoRI fragment spanning the 3′ portion of the gag gene. Positive clones were picked and screen purified five subsequent times with either probe to get a 100% positive screen.

Cell lines, transfection, and coculture.

Crandell feline kidney (CrFK) cells and primary activated peripheral blood mononuclear cells (PBMC) were maintained as previously reported (20). CrFK cells (2 × 10⁵) were plated in six-well plates and allowed to attach overnight. Transfections were conducted with 100 to 400 ng of full-length FIV-C36 clone DNA or 5 μg of pUC18 or FIV-34TF10, a molecular FIV clone obtained from the tissue culture-adapted Petaluma isolate (46). Fugene6 (Roche) was used according to the manufacturer's instructions. Transfection was allowed to proceed overnight. At ∼18 h posttransfection, CrFK medium was removed and 10⁶ primary activated PBMC (5 to 7 days in culture) were seeded over the transfected CrFK cells. Coculturing was allowed to occur for ∼24 h, at which time the PBMC were removed and seeded into T25 flasks while the CrFK were fed with fresh Dulbecco's modified Eagle's medium-10% fetal bovine serum.

RT assay.

Cell culture supernatant from virus-infected PBMC was harvested every 4 days. Fifty microliters of supernatant was used for a micro reverse transcriptase (RT) assay with 96-well plates. Viruses were disrupted in 20 mM dithiothreitol, 0.5% Triton X-100, and 750 μM KCl. Viral lysates were then incubated in a reaction mixture containing 1 M Tris-HCl (pH 8.1), 100 mM MgCl₂, 0.025 U of poly(rA-dT)_12-18, and 0.1 mCi of [³H]dTTP at 37°C for 2 h. The reaction was briefly stopped by incubation on ice, and total samples were transferred to DE81 Whatman filter circles. Filters were fixed in 0.1 M sodium pyrophosphate for 5 to 10 min, followed by three washes in 0.3 M ammonium formate (pH 7.8) and one 95% ethanol wash. Filters were dried under a lamp, and radioactivity was measured with a liquid scintillation counter.

PCR amplification.

PCR analysis was conducted by using Promega Taq polymerase and various FIV-C primer sets spanning the expected ∼9.4 kb of cloned FIV-C36. DNA was denatured at 94°C for 5 min, followed by 30 cycles of 94°C for 15 s, 58°C for 15 s, and 72°C for 1 min/kb, terminated with a 7-min extension at 72°C, and cooled to 4°C.

Viral RNA extraction and reverse transcription.

Viral RNA was extracted from infected cat plasma harvested at various time points throughout the experiment with the QiaAmp Viral RNA kit (QIAGEN). For extraction efficiency comparisons, 10⁹ copies of kanamycin (KAN) RNA (Promega) were spiked into each of the plasma samples prior to RNA extraction. To remove residual cellular DNA, a 10-min on-column DNase treatment (QIAGEN) was added to the manufacturer's protocol. In vitro transcription was performed on extracted FIV-C36 and KAN RNA by using Stratascript RT (Stratagene) according to the manufacturer's instructions. cDNA was generated at 42°C for 1.5 h with primers for the FIV-C RT gene d(GGGGGTTCTTCCTGTAATTTATC) and the KAN gene d(AATGGCTGGCCTGTTGAACAA).

Quantitative RT-PCR.

PCR products were generated from the FIV-C36 and KAN cDNAs by using an ABI PRISM 7700 thermal cycler (PE Applied Biosystems, Foster City, Calif.) and LUX primer technology (Invitrogen). Primer pairs were designed by using the LUX Designer software (Invitrogen) (29). A 5′ region of the FIV-C36 RT gene (96 bp) and an 80-bp KAN gene sequence were amplified per sample in separate wells. The following primer pairs, one of which was 3′ labeled with the reporter dye 6-carboxy-fluorescein, were used: FIV-C36 RT forward primer, d(GGACTACCTCACCCTGCTGGA); reverse primer, d(CTACTTTGGATCGAGGGGAATGGTAAAGTAG); KAN forward primer, d(CAAACCCAATCACGAATGAATAACGGTTTG); reverse primer, d(AATGGCTGGCCTGTTGAACAA). PCRs (25 μl) were set up in duplicate per sample per gene (RT or KAN) with 5 μl of cDNA, the ready-to-use Platinum Quantitative PCR SuperMix-UDG (Invitrogen) mixture, and either the primer pair (10 μM each) for FIV-C36 RT or the primer pair for KAN. Reaction mixtures were first heated at 50°C for 2 min (UDG treatment) and denatured at 95°C for 2 min. A two-step cycling protocol was used for 40 cycles of 95°C for 15 s and 60°C for 45 s.

Standard curves were generated in duplicate for each quantitative RT-PCR (QRT-PCR) run by using 1 log serial dilutions (10⁷ to 10¹) of either a plasmid containing the 5′ 4-kb sequence of FIV-C (includes the RT gene) or the pET28a(+)vector plasmid (Novagen) containing the KAN gene. Standard dilutions were aliquoted and maintained at −20°C for less than 2 months to ensure a single freeze-thaw cycle and DNA stability. The QRT-PCR standard curve slope was −3.48, indicating a PCR efficiency [E = 10(−1/slope)] of 94%. The correlation coefficient (R²) was 0.99, indicating good linearity of the standard curve. Accumulated PCR products resulting in fluorescence signals were analyzed by using a Sequence Detector Software program (SDS, version 1.9.1). We obtained nearly identical values for KAN quantitation in every KAN sample well, indicating minimal loss due to experimental manipulation. Background values (baseline plasma samples) were subtracted per sample. Detection limits were in the range of 100 RNA copies/ml of plasma. Data were analyzed with a Microsoft Excel program.

Cloning and sequencing.

The lambda phage clone containing the FIV-C36 sequence was subcloned by EcoRI into pBluescript SK±, and three fragments were isolated. One fragment contained the 5′ portion of FIV-C36 including ∼700 bp upstream of the 5′ LTR, a 2.2-kb EcoRI fragment, which contained most of the pol gene of FIV-C36, and a 3′ portion that contained an additional ∼280 bp of sequence downstream of the 3′ LTR. Using FIV-C primers, the entire FIV-C36 genome was sequenced by using an Applied Biosystems Abi 373 machine.

Western hybridization and immunoprecipitation.

For viral Western blotting, 10⁶ infected PBMC were washed and resuspended in 100 μl of phosphate-buffered saline before undergoing three cycles of freeze-thawing. Cell debris was spun down as a pellet, and 20 μl of 8 M urea and 20 μl of Laemmli buffer (18) were added to the viral supernatant. The mixture was passed through a 27-gauge needle 20 times and heated to 95°C, and 30 μl was loaded onto a 10 to 20% Tris-Tricine gel. The gel was then transferred to a nitrocellulose membrane and incubated overnight in primary polyclonal antibody against capsid p24 protein (produced by our laboratory). After washing, the membrane was incubated in secondary antibody (goat anti-rabbit immunoglobulin G [heavy plus light chains] horseradish peroxidase human and mouse adsorbed; Southern Biotechnology Associated, Inc., Birmingham, Ala.). Capsid protein was made visible by enhanced chemiluminescence (Pierce).

Alignments.

Multiple alignments were conducted by using the ClustalW algorithm in MacVector, version 6.5, software (Oxford Molecular, Madison, Wis.) for calculating percent identities of various gene products of different FIV strains.

RESULTS

FIV-C36 is significantly divergent from FIV-A and -B isolates.

FIV-C isolate CABCpady00C produced severe pathogenic effects in 60% of infected cats, including marked reduction in CD4⁺ T cells, thymic atrophy, systemic lymphoid depletion, and severe opportunistic infection (11). To isolate a potentially infectious clone of FIV-C, the Lambda EMBL 3/BamHI system (Stratagene) was used to create a genomic library by using DNA from CABCpady00C (11)-infected 104-C1 T cells. The genomic BamHI DNA library was screened by Southern blotting for FIV-C clones with 5′ FIV-C-specific probes. The molecular clone FIV-C36 was isolated, and Southern analysis revealed an ∼18- to 20-kb fragment flanked by EMBL 3 right and left arms (data not shown). Sequencing of FIV-C36, with FIV-C primers, revealed an overall length of 9,491 nucleotides, which is consistent with other FIV genomes (30, 46).

The genomic organization of FIV-C36 is similar to known subtype A FIVs, such as FIV-PPR (35), FIV-34TF10 (46), and FIV-14 (30, 46), as well as the subtype B isolate FIV-TM2 (27). The LTRs of FIV-C36 show conservation in the 2-base inverted repeats at their 5′ and 3′ ends, as well as the TATA promoter box and poly(A) signal (Fig. 1). The U3 region of the LTR contains consensus DNA binding elements for the transcriptional enhancer-promoter proteins AP-1 and AP-4 and the cyclic AMP response element or ATF binding element (21). However, unlike the other FIVs, a second consensus AP-4 site and a CCAAT promoter element consensus sequence found in FIV-PPR are lacking from the FIV-C36 LTR. Some imperfect direct repeats, often identified as enhancer sequences, are also found in the FIV-C36 LTR. A consensus sequence for LBP-1 (leader binding protein 1) is present in FIV-34TF10 and FIV-TM2 but not in FIV-PPR, FIV-Cgb, or FIV-C36.

FIG. 1.

Schematic comparison of FIV LTR organization. The limits and locations of the U3, R, and U5 regions and the relative positions of the various protein binding elements are indicated for subtype A (PPR and 34TF10) and subtype C (FIV-Cgb and FIV-C36). IDR, imperfect direct repeat.

The predicted amino acid sequences of the different gene products of FIV-C36 were compared to the existing FIV sequences as a first step in understanding the characteristics of this infectious clone. The Env protein of lentiviruses, consisting of a surface protein and a transmembrane region (Fig. 2), plays a major role in infectivity and cell specificity due to its involvement in cell receptor docking. Env sequences are, therefore, used to phylogenetically categorize HIV type 1 (HIV-1) (28) as well as FIV (37, 42). We compared FIV-C36 Env to the translated Env sequences of four known FIVs (Fig. 2 and 3). FIV-Cgb and FIV-C36 Env proteins have an overall 96% identity at the amino acid level (Fig. 3). FIV-C36 Env has only 78% overall amino acid identity to FIV-PPR Env and 76% identity to FIV-34TF10 and FIV-TM2 Env (Fig. 3). The Env proteins of the two clade A viruses, FIV-PPR and FIV-34TF10, have 84% amino acid sequence identity. A high degree of sequence diversity is evident in the hypervariable regions (Fig. 3) compared to the other FIV subtypes. Whether any particular variance has an effect on Env protein folding and function remains to be analyzed. The notable point is the large number of divergent residues within the surface protein. As previously reported, any one of the changes may be responsible for tropism determination, cell receptor docking, and infectivity of FIV-C36 (19, 51). Furthermore, we noticed a large cluster of divergence within hypervariable regions 1 and 2 (V1 and V2), which span residues 49 to 75 and 95 to 175, respectively (Fig. 3). V1 and V2 reside in the putative leader region of FIV env, coding for the 5′ exon of Rev.

FIG. 2.

Bar diagram of the genomic organization of FIV-C36. (a) LTRs flank the genome comprised of viral genes encoding MA, CA, nucleocapsid (NC), P2, PR, RT, deoxyuridine triphosphatase (DU), integrase (IN), virion infectivity factor (Vif), ORF A, Rev, and envelope (Env). Env is enlarged to show the relative positions of the leader sequence (L), surface protein (SU), and transmembrane sequence (TM). The relative location of the hypervariable regions of env (V1 to V9) are marked by thick black lines and numbered accordingly. env codes for ∼860 amino acids marked in increments of 100. Some of the specific residue divergences between FIV-Cgb and FIV-C36 and their relative positions are depicted. (b) A large number of divergent residues are found within V2.

FIG. 3.

ClustalW alignments of FIV Env residues. (a) Alignment of Env residues of FIV-C36, FIV-Cgb, FIV-PPR, FIV-34TF10, and FIV-TM2. (b) Percent differences at the amino acid level of Env among the various FIV species.

Sequence diversity was markedly reduced among other viral proteins, with only minor divergence in the Gag-Pol protein matrix (MA), capsid (CA), nucleocapsid, P2, protease (PR), RT, deoxyuridine triphosphatase, and integrase (Fig. 4). Interestingly, the sequence comparison of the virion infectivity factor, Vif, between FIV-PPR and FIV-34TF10 revealed 90% sequence identity at the amino acid level. However, FIV-PPR and FIV-34TF10 Vif proteins share only 83 and 86% identity with FIV-C36, respectively (Fig. 4). The clade B strain FIV-TM2 Vif protein has only 80% identity to FIV-C36.

FIG. 4.

Comparison of the various gene products between different FIVs. Viral gene products MA, CA, PR, RT, virion infectivity factor (Vif), and ORF A were compared at the amino acid level between FIV-PPR (PPR), FIV-34TF10 (34TF10), FIV-TM2 (TM2), FIV-Cgb, and infectious clone FIV-C36. For comparison purposes, the stop codon TGA of ORF A in FIV-34TF10 was repaired to produce tryptophan (TGG) (*).

When comparing the open reading frame (ORF) A protein sequences among the clades, the representative FIVs in clades A and B show a significantly increased divergence compared to FIV-Cgb and FIV-C36, ranging from 30 to 40% variance (Fig. 4). Many of the residue differences, however, are conserved, and it remains to be determined whether significant differences exist in the relative activities and/or specificities for the ORF A proteins of different clades.

FIV-C36 exhibits lymphoid tropism like primary field isolates.

To analyze FIV-C36 tropism, transfection and coculture experiments were conducted in vitro. The FIV-C36 productively infected activated primary feline PBMC but did not infect the CXCR4-bearing nonlymphoid feline cell line CrFK. The latter cell type was, however, readily transfectable by FIV-C36 and generated virus particles. Thus, various amounts of FIV-C36 DNA were transfected into CrFK and transiently cocultured with 10⁶ fresh feline PBMC at 18 h posttransfection. The cocultured PBMC were removed after 24 h and cultured independently from the transfected CrFK cells. The infectivity of FIV-C36 was monitored by an increase in RT activity in the cultured primary feline PBMC. PBMC infected with FIV-C36 generated a Mg²⁺-dependent RT activity at ∼10 days postcoculture (Fig. 5b). By contrast, no significant RT activity was detectable in supernatants of the transfected CrFK cells when cultured independently, consistent with an inability of the virus to amplify in the adherent cell line (Fig. 5a). FIV-34TF10 was used as a concurrent control for transfection studies because this tissue culture-adapted clone readily infects and replicates in CrFK cells (Fig. 5a). However, FIV-34TF10 does not productively infect feline PBMC due to a stop codon in the ORF A gene (34, 46).

FIG. 5.

FIV-C36 infects primary PBMC as measured by RT. CrFK cells were transfected with 100, 300, or 400 ng of FIV-C36. At 18 h posttransfection, 10⁶ primary PBMC were cocultured with the transfected CrFK. At 24 h postcoculture, primary PBMC were removed from CrFK and both cell types were maintained in appropriate media. Supernatant was harvested from all cultures every 4 days for RT analysis. (a) CrFK, transfected CrFK cells; TF1 to TF3, CrFK transfected with 100, 300, or 400 ng of FIV-C36 DNA, respectively; TFpuC18, CrFK transfected with 5 μg of plasmid DNA as a negative control; TF34TF10, CrFK transfected with 5 μg of 34TF10 DNA as a positive control (34TF10 infects CrFK cells). (b) Primary PBMC cocultured with FIV-C36 supernatant show an increase in RT activity over time. The arrow indicates the point at which fresh primary PBMC were added to the culture to maintain viral infectivity (34TF10 does not infect primary PBMC). (c) Western analysis of FIV-C36-infected PBMC supernatant reveals the presence of capsid p24 protein not seen in noninfected PBMC.

To further document replication of FIV-C36, we analyzed expressed viral proteins by Western blotting (Fig. 5c). Infected PBMC and their respective supernatants were harvested, and lysates were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Capsid p24 protein was easily demonstrated in infected cells, consistent with the appearance of RT activity in cells (Fig. 5b).

In PBMC, FIV-C36 infection was accompanied by changes in cell morphology, including syncytium formation and cell death. Infected PBMC began to form multinucleated cells and exhibited balloon degeneration (data not shown). Due to the progressive loss of infectible PBMC, the RT activity decreased over time (Fig. 5b). Upon addition of fresh PBMC (Fig. 5b), a renewed increase in RT activity was observed, as infectious virions infect the freshly added cells. By contrast, we observed no change in the phenotype of the FIV-C36-transfected CrFK, whereas FIV-34TF10-infected CrFK formed large syncytia (35, 36) and underwent cell death as previously observed (data not shown).

In vivo pathogenicity of FIV-C36.

To examine the in vivo activity of FIV-C36, preliminary experiments were performed with tissue culture supernatant from infected feline PBMC of varied passage lengths to inoculate 8- to 10-week-old SPF cats. All inoculated cats became positive for FIV-C36 sequences and antiviral antibodies (data not shown) by 2 weeks PI. Moreover, marked declines in CD4⁺ T cells occurred by 8 weeks PI, ranging from 8 to 52% of control levels, reminiscent of the uncloned CABCpady00C (11). In one animal inoculated with FIV-C36 phage DNA, CD4⁺-T-cell depletion was also noted, but to a lesser extent (65% of control levels) versus animals infected with the short-term-infected PBMC inocula (8% of control levels). Attempts were made to increase the viral titer by longer passage in cultured PBMC. While higher RT values were achieved after 9 weeks in culture, inoculated animals demonstrated no diminution in CD4⁺ T cells or clinical illness (data not shown), consistent with attenuation during tissue culture passage.

Based on the above preliminary findings, we inoculated one litter of 4 2-week-old SPF kittens (designated cat C, S, Bw, and T) with tissue culture supernatant (1 ml intravenously and 0.75 ml orally at 10^3.4 50% tissue culture infectious doses/ml) from short-term culture (12 to 18 days posttransfection) of FIV-C36 in PBMC. At 4 weeks PI, one of the 4 inoculated kittens (cat C) was found dead and 2 of the remaining 3 kittens had developed an acute severe clinical illness characterized by weight loss, fever or subnormal body temperature, dehydration, malaise, anorexia, and readily detectable oral and pharyngeal ulceration. The response to supportive therapy (antibiotic treatment and fluid replacement) was not sufficient in these animals, mandating euthanasia by week 5. Necropsy revealed oral and pharyngeal ulcerations in 2 of 3 animals, severe thymic atrophy in all 3 cats, and gross evidence of acute pneumonia in the lymph nodes and spleens of all animals (Fig. 6). PCR and cultures for secondary feline viral pathogens were negative. Bacterial culture of oral lesions and lungs yielded mixed populations including beta-hemolytic streptococci. The oral and pulmonary lesions were thus considered opportunistic. The clinical syndrome and lesions were similar to those observed in the original studies with the uncloned CABCpady00C virus isolate (11). All 3 infected, euthanized animals experienced marked declines in CD4⁺ and CD8⁺ T cells (Fig. 7), consistent with past observations (11, 15, 44, 57). In the one surviving inoculated cat, Bw, a consistent increase in CD8⁺ T cells and a relatively stable CD4⁺-T-cell number contrasted with its moribund littermates (Fig. 7e). QRT-PCR revealed viral RNA loads in plasma to be very high in all 3 affected, euthanized cats (Fig. 8). Viral RNA levels continued to rise over the first 4 weeks for cat C (found dead during week 4) and cat S and remained high for cat T (both euthanized during week 5). By contrast, the one surviving cat, Bw, had decreasing viral RNA levels concomitant with an increase in CD8⁺ T cells (Fig. 7e), a pattern typical of FIV-C survivors (10).

FIG. 6.

Severe lymphoid depletion in an SPF cat euthanized due to clinical immunodeficiency syndrome 5 weeks after infection with the FIV-C36 molecular clone. (A) Control thymus, note the cortex (C), containing abundant lymphocytes, and normal organization and distinction of cortex and medulla (M); (B) markedly atrophied thymus from an FIV-C36-infected cat illustrating severe depletion of cortical (C) lymphocytes and obliteration of cortical-medullary (M) junction and overall organization; (C) control lymph node demonstrating abundant normal lymphoid follicles (arrows); (D) lymph node from FIV-C36-infected cat demonstrating marked lymphocyte depletion with loss of lymphoid follicles (arrows). Magnifications are given on the figure.

FIG. 7.

CD4⁺ and CD8⁺ T lymphocytes in infected cats. (a to e) CD4⁺- and CD8⁺-T-cell counts of FIV-C36-infected and age-matched control cats per week. The surviving cat, Bw (e), shows a significant increase in CD4⁺ and CD8⁺ T cells as of week 2.

FIG. 8.

QRT-PCR viral RNA load measurements. (a to d) Viral loads in FIV-C36-infected cat plasma for each week. Week 1 samples and the week 4 sample for cat T (*) were not available. (e) Viral load in dam plasma for weeks 0, 5, and 6. Week 0 viral loads were below the detection limits (10² RNA copies/ml) of QRT-PCR.

Of particular interest was that the mother cat (dam) nursing the 4 infected kittens started to show symptoms of disease similar to those in the 3 affected, euthanized kittens. Analysis of plasma samples for viral load revealed that she had become QRT-PCR positive by 5 weeks PI of the kittens (Fig. 8e). By 11 weeks PI of the kittens, she was euthanized, exhibiting oral ulcers, decline in CD4⁺ T cells (from 1,034 at the baseline to 29 at 11 weeks PI of the kittens), anorexia, and weight loss. Necropsy revealed marked thymic atrophy and lymphoid depletion. To verify the dam's infective viral origin, we cloned and sequenced the 5′ env region of the virus isolated from dam plasma. The dam env viral sequence exhibited 99.9% sequence identity to FIV-C36 (data not shown). To our knowledge, this is the first report of offspring-to-parent transmission of FIV infection and attests to the pathogenic nature of the FIV-C36 clone.

A second in vivo experiment was performed wherein 2 2-week-old newborn animals were infected with FIV-C36 tissue culture supernatant, as detailed above for the first experiment. Similar results were obtained, with severe CD4⁺-T-cell depletion and appearance of oral lesions and opportunistic infections, necessitating euthanasia at 7 weeks postinfection. As in the first experiment, the dam became PCR positive and exhibited CD4⁺-T-cell depletion, but she still survives at this writing, approximately 12 weeks postinfection of the kittens (data not shown).

DISCUSSION

The results of the present study identify several interesting features of the molecular structure of the newly isolated molecular clone, FIV-C36. FIV-C36 demonstrated an overall similarity to known FIV genomic structure and gene organization. The LTR contains most of the typical enhancer protein binding elements present in other FIVs. Missing from the FIV-C36 LTR is the canonical CCAAT promoter element, a second AP-4 binding site, and the LBP-1 element. Sparger et al. (45) showed that mutation of the second AP-4 element caused no loss of transcriptional power of the viral LTR and, therefore, seems redundant. LBP-1 has been shown to repress HIV-1 transactivation (6, 22) and may, therefore, be disadvantageous to the virus. An ATF binding element is retained in FIV-C36, previously shown to be a crucial protein necessary for viral replication (4, 9).

Comparison of the various structural, regulatory, and accessory gene products of FIV-C36 to those of other FIVs revealed some interesting differences as well as similarities. No major differences were noted when comparing the structural proteins MA and CA or the enzymatic proteins PR and RT. However, increased divergence in the accessory protein Vif was detected between FIV-C36 and FIV-PPR, FIV-34TF10, and FIV-TM2 (83, 86, and 80% homology, respectively) (Fig. 4). This includes various nonconserved residue changes as well as potentially phosphorylated amino acids. The Vif protein plays a crucial role in early infection in vivo, and FIV lacking Vif results in significantly reduced viral loads (16). Vif is necessary for viral replication (43, 53) and productive infection of feline PBMC, feline cell lines, and macrophage cells (39, 48). Furthermore, HIV-1 Vif phosphorylation significantly influences HIV-1 replication and infectivity (58) and works to facilitate infection by preventing cytosine deamination by APOBEC3G/CEM15 (25, 26, 59). Although the latter function has yet to be specifically ascribed to FIV Vif, sequence divergence in Vif may have an impact on host cell range and disease phenotype. Posttranslational modification studies of HIV-1 have shown the Vif protein to be a highly phosphorylated protein and the point mutation of serine¹⁴⁴ to alanine completely eliminated Vif activity (58). Whether a change in the phosphorylation pattern of FIV Vif can influence infectivity has, to our knowledge, not been examined. It will be of interest to determine whether a similar mechanism of Vif function exists in FIV.

The accessory protein, ORF A, facilitates transcriptional transactivation of the viral LTR (9) but by a mechanism distinct from those of other lentiviruses (4). FIVs lacking a functional ORF A, such as the laboratory-adapted molecular clone FIV-34TF10, can productively infect CrFK cells but cannot productively infect primary PBMC or feline T-cell lines (35, 47, 52). FIV-34TF10 fails to productively replicate in PBMC, since its ORF A is eliminated due to a premature stop codon. Repair of the premature stop codon allows FIV-34TF10 to replicate in PBMC (52). The accessory protein ORF A also exhibits extensive amino acid divergence between FIV-C36 and ORF A's of clade A and B FIVs, in the range of 30 to 40% variability (Fig. 4). There were very few nonconserved amino acid changes, suggesting that the overall function of the protein may not be significantly affected. However, single residue changes, such as Ser¹⁴⁴ to alanine in Vif (58), can affect the function of a protein, and therefore, we are testing ORF A of FIV-C36 to understand its relevance in viral infectivity.

The Env protein is a major determinant of host cell tropism and receptor attachment (41, 50, 51, 54). The largest differences in Env were noted when comparing the FIV-C36 Env protein to Env proteins from the other clade representatives chosen for this study. FIV-C36 Env has an overall identity of 96% compared to the PCR-derived FIV-Cgb sequence (Fig. 3b). However, when compared to clade A and B Env proteins, the identity is only 78 and 76%, respectively. Predictably, the biggest diversity is located in regions V3 to V5 (2, 3, 40). The V3-V4 region was shown to contain macrophage-tropic determinants, whereas a loss in positively charged residues in V3 may allow a T-cell-tropic phenotype (49). An enlarged target cell repertoire, allowing for successful infection of a widened array of cells, can determine increased pathogenicity of a viral isolate. Chimera construction and point mutation analysis are being conducted to identify the determinants of increased pathogenicity of FIV-C36.

Interestingly, V1 and V2 also show large variability. V1 and V2 are located in the putative Env leader which also codes for the N-terminal half of Rev (Fig. 2a), whereas the C-terminal Rev is spliced from an exon located 3′ of env (34). Rev in HIV-1 (23) and FIV (34) binds to mRNAs containing Rev response elements, facilitating nuclear export of singly spliced and unspliced viral mRNA. Successful shuttling of mRNA into the cytoplasm is a crucial part of virion formation. The functional regions of FIV Rev map to the C-terminal portion of the molecule (24), and partial Rev activity has been demonstrated for the C-terminal portion of Rev alone (9). Yet, the latter study also demonstrated that the N-terminal domain was required for full activity, so it remains to be determined how the V1 and V2 changes impact virion phenotype. Whether the significant divergence in the V1 and V2 sequences has an effect on the biological activity of Rev function is not certain.

FIV-C36 was demonstrated to be infectious by conducting transfection and infection studies in vitro and in vivo. CrFK cells are readily transfectable with the FIV-C36 clone and produce viable virions within 24 h. Coculturing primary feline PBMC with the transfected CrFK allowed infection of PBMC and subsequent analysis of RT activity, a sign of infectivity of the virus. One of the first signs of lentivirus infection is the formation of syncytia in cultured PBMC. FIV-C36-infected PBMC undergo characteristic morphological changes. Individual cells begin to fuse, forming giant multinucleated cells and ballooning, which leads to cytopathic cell death. Shortly following the appearance of ballooning cells, we observed an increase in RT activity at ∼10 days postcoculture of the PBMC but not so in the transfected CrFK, indicating host cell specificity typical of field strains of FIV.

In our preliminary series of in vivo studies with FIV-C36 viral supernatants or phage DNA containing FIV-C36, all inoculated animals became PCR positive and developed modest decreases in circulating CD4⁺ T cells, suggesting that FIV-C36 successfully infected T cells but that this infection failed to rapidly amplify. We suspect the long-term (9 weeks) in vitro passage of FIV-C36 generated higher viral titers yet resulted in virus attenuation. In the one animal that received FIV-C36 DNA in the context of lambda phage, we observed CD4⁺-T-cell depletion. Therefore, in the next experiment, a more short-term PBMC-propagated virus pool was generated for use in the second preliminary pathogenicity screen. Substantially greater virulence and replication was observed, more closely approximating the original uncloned CABCpady00C virus, which was lethal in 60% of intravenously inoculated animals (11). This finding underscores the need to limit or avoid in vitro passage, and it is likely that DNA inoculation is the safest protocol to avoid drift in genotype. The only kitten that survived the FIV-C36 inoculation study had a robust increase in CD8⁺ T cells and a concomitant lower virus burden than the rapid-progressor littermates. Although cytotoxic T lymphocyte responses were not measured in these studies, earlier reports imply that cell-mediated responses may be responsible for the observed survival of this one offspring (5, 7, 13, 14).

The finding that the dams became infected in both experiments where they nursed infected offspring is surprising and of great interest. To our knowledge, this is the first report of lentivirus transfer from offspring to parent. Although there may have been some potential for blood contact, we have no evidence that this occurred. Rather, it appears that either the act of suckling or, more likely, grooming by the mother resulted in sufficient mucosal exposure to transfer the virus from the offspring to the dam. Importantly, the result indicates that contact more casual than previously thought can result in infection of naive animals. Furthermore, the finding that the dam in the first experiment went on to develop the full acute-phase disease indicates that the severe immunodepletive pathogenicity is neither limited to young animals nor to intravenous inoculation. Experiments are planned to further evaluate the route of infection and the importance of host age and familial innate resistance and immunity in the acute pathogenic response.

The identification of this replicative clone of FIV-C isolate CABCpady00C allows numerous future studies employing chimeric virus constructs to identify the viral genetic determinants of the acutely pathogenic phenotype. Env residue changes and their influence on folding and charge represent an important field of study, since changes in Env have been shown to alter cell tropism. Furthermore, the extent of involvement of other FIV genes in determining infectivity and pathogenic properties of FIV remain to be understood. Further studies with specific chimeric FIVs will allow the elucidation of molecular mechanisms controlling host cell range and pathogenesis.