Biologic Studies of Chimeras of Highly and Moderately Virulent Molecular Clones of Simian Immunodeficiency Virus SIVsmPBj Suggest a Critical Role for Envelope in Acute AIDS Virus Pathogenesis

Abstract

Previous studies identified three molecular clones of the acutely pathogenic SIVsmPBj strain that varied in terms of relative in vivo pathogenicity. One clone, SIVsmPBj6.6, reproducibly induced a rapidly fatal disease in pigtailed macaques. In contrast, a highly related clone (SIVsmPBj6.9) was only minimally pathogenic in macaques. PBj6.6 and PBj6.9 shared a tyrosine substitution at position 17 in the Nef protein that is a major determinant of virulence but differed at one residue in Vpx (C89R), three residues within the envelope (D119G, R871G, G872R), and a single residue in Nef (F252L). SIVsmPBj6.9 was less efficient in inducing proliferation of resting macaque peripheral blood mononuclear cells in vitro than SIVsmPBj6.6 and exhibited a marked reduction in infectivity relative to SIVsmPBj6.6. Chimeric viruses for each of these variable residues were constructed, and their biologic properties were compared to those of the parental strains. Differences in Vpx and Nef did not alter the basic biologic phenotype of the chimeras. However, the D119G substitution in the envelope of SIVsmPBj6.9 was associated with a marked reduction in the infectivity of this virus relative to SIVsmPBj6.6. An associated processing defect in gp160 of SIVsmPBj6.9 and chimeras expressing the D119G substitution suggests that a reduction in virion envelope incorporation is the mechanistic basis for reduced virion infectivity. In vivo studies revealed that substitution of the PBj6.9 amino acid into PBj6.6 (D119) abrogated the pathogenicity of this previously pathogenic virus. Introduction of the PBj6.9 G119, however, did not confer full virulence to the parental PBj6.9 virus, implicating one or all of the other four substitutions in the virulence of SIVsmPBj6.6.

The infection of macaque monkeys with simian immunodeficiency virus (SIV) is a useful animal model to investigate the pathogenesis of human immunodeficiency virus type 1 (HIV-1). SIV-induced disease is similar to human AIDS, with the development of high virus loads, progressive depletion of CD4⁺ T cells, opportunistic infections, and death of infected animals within a few months to years (4). In contrast to the majority of SIV isolates, a virus isolated from a pigtailed macaque (PBj) infected with the AIDS-inducing SIVsmm9 strain evolved a variant pathogenesis (5, 12–15). This virus, designated SIVsmPBj14 (for the macaque of origin and the month post-SIV inoculation), induced an acute and lethal illness within 14 days of inoculation characterized by profuse diarrhea, dehydration, severe lymphopenia, and an extensive cutaneous rash. Pathologic features included major gastrointestinal cytopathology with villus blunting (15), massive mononuclear cell infiltration within the gastrointestinal tract, high levels of virus replication in the gastrointestinal-associated lymphoid tissue, and immune system hyperactivation (14, 15). Elevated levels of cytokines such as tumor necrosis factor alpha (14, 21, 39) and interleukin-6 (2) produced within the sites of the lesions (5) suggest that the pathogenesis of this novel disease syndrome is cytokine mediated (22, 48). Evidence of increased apoptosis within gastrointestinal lesions and lymphoid tissues (18) also suggests that apoptotic mechanisms may contribute to pathogenesis. The ability of PBj14 viruses to activate and replicate in resting macaque peripheral blood mononuclear cells (PBMC) (13) is predictive of pathogenesis in vivo (33).

Several representative molecular clones have been derived from the original PBj14 biological clone (SIVsmPBj14-bcl2). Despite their common origin, these various clones vary considerably in terms of in vivo virulence. At least two (PBj6.6 and PBj4.19) fully reproduce the virulence of the biologically cloned virus isolate. Two others induce moderate symptoms (PBj6.9 and PBj1.9), and some do not appear to induce acute disease (PBj6.12) (5, 6, 26, 33, 34). As with the uncloned viruses, the ability of these viruses to induce proliferation of resting PBMC appeared to be an accurate predictive marker for in vivo pathogenicity. Sequence comparison between the parental SIVsmm9 and SIVsmmPBj14 viruses identified 36 amino acid changes throughout the genome which might be responsible for the novel pathogenesis, as well as a duplication of the NF-κB site and an insertion in the V1 region of Env (3, 5, 6). Several regions of the genome of SIVsmPBj that may be important for pathogenesis have been identified. The principal pathogenic determinant identified is a mutation (17RQ to 17YE) that introduces an immunoreceptor tyrosine-based activation motif in Nef (8). However, other unique features, such as duplication of the NF-κB site in the long terminal repeats (LTR) (3, 5, 32, 33), the U3 LTR promoter region (7), the viral envelope (33, 34), and the nef (8, 9, 37) and vpx genes (20), play a minor role in pathogenesis.

Although the pathogenesis of the various molecular clones of SIVsmPBj varies significantly, these viruses are remarkably similar in terms of sequence identity. The highly pathogenic PBj6.6 and the less pathogenic PBj6.9 viruses differ by only five amino acids distributed in three genes of the 3′ half of the genome (33). They differ at one position within Vpx (C89R), three positions, within Env (D119G, R871G, and G872R), and a single position within Nef (F252L). Interestingly the Nef tyrosine mutation is present in both the PBj6.6 and PBj6.9 viruses. The purpose of the present study was to construct chimeras between the highly pathogenic SIVsmPBj6.6 and the less pathogenic SIVsmPBj6.9 in order to map the substitutions responsible for the differences in their pathogenesis in vivo.

MATERIALS AND METHODS

Generation of chimeric PBj molecular clones.

The infectious molecular clones PBj6.6 and PBj6.9 have been previously described (33) and consist of the entire proviral DNA cloned into pGEM3Zf(+). nef, vpx, and env exchange viruses, along with viruses containing point mutations, were generated by standard techniques of restriction digestion (30). Chimeric viruses resulting from the exchange of the nef, vpx, and env genes between the PBj6.6 and PBj6.9 parental viruses are numbered, with the lower-numbered virus of a pair corresponding to the PBj6.6 virus background. Similarly, the higher-numbered virus of each pair corresponds to the PBj6.9 virus background containing the PBj6.6 amino acid changes. For example, virus 018 was the PBj6.6 virus now containing the PBj6.9 Nef gene amino acid changes and vice versa for virus 020 (PBj6.9 virus with PBj6.6 Nef). Plasmid DNA isolations from Escherichia coli were performed using Midi and Mini prep kits (Qiagen Inc., Santa Clarita, Calif.).

(i) nef exchange.

To generate chimeras exchanging nef genes (018 and 020), a one-step gene replacement procedure was used (24). The mutagenesis vector pALTER Ex-I (Altered Sites II in vitro mutagenesis system; Promega, Madison, Wis.) was modified by the addition of a ClaI-SpeI-XhoI linker into the NotI- and NdeI-digested plasmid. ClaI-XhoI fragments containing the nef gene from PBj6.6 and PBj6.9 proviral DNA were inserted into the modified pALTER Ex-1 vector, producing pMH97.09 and -97.11, respectively. PCR products corresponding to the PBj6.6 and PBj6.9 nef genes were obtained using primers MHA6 (5′-ATG GGT GGC GTT ACC TCC AAG AAG-3′ [nucleotides 9061 to 9084]) and MHA7 (5′-TTA GCT TGT TTT CTT CTT GTC AGC C-3′ [nucleotides 9846 to 9822]). Vent DNA polymerase (New England Biolabs, Inc.) was used for the amplification with the reaction buffer supplied [10 mM KCl, 20 mM Tris-HCl (pH 8.8), 10 mM (NH₄)₂SO₄, 0.1% Triton X-100], 2.5 mM (each) deoxynucleoside triphosphates (Pharmacia), 25 μM (each) primers, 1 ng of plasmid DNA template, and 1 U of enzyme activity. Reactions were performed in a DNA thermal cycler (Perkin-Elmer) for 15 cycles (94°C for 1 min, 60°C for 1 min, 72°C for 1 min, followed by a 5-min extension at 72°C). The gel-purified PBj6.6 nef PCR product (1.25 pmol) was annealed to 0.05 pmol of the denatured pMH97.11 plasmid, along with 0.25 pmol of an oligonucleotide, amp^r, (Promega) that restored ampicillin resistance to the vector, allowing for selection of the mutated constructs. DNA synthesis was achieved using T4 DNA polymerase and DNA ligase as recommended (Promega). Following transformations into ES1301 mutS E. coli and JM109 E. coli cells, nef gene exchanges were identified by selection of colonies on Luria-Bertani agar plates containing 125 μg of ampicillin per ml and were confirmed by sequencing. The ClaI-XhoI fragment containing the exchanged nef gene was built back into the parental ClaI-XhoI-digested molecular clones.

(ii) vpx exchange.

To construct viruses 034 and 035, pGEM3Zf(+) was first digested with SmaI and HindIII to remove the BamHI to SphI sites from the multiple cloning site. The HindIII end was filled with Klenow DNA polymerase, and the blunted vector, pGEM3Zf(+) BS, was religated. EcoRI fragments from PBj6.6 and PBj6.9 molecular clones were inserted into the EcoRI-digested pGEM3Zf(+) BS vector. Digestion of this construct with NsiI and XbaI releases a fragment encoding the single amino acid change that distinguishes Vpx of the PBj6.6 virus from that of the PBj6.9 virus. The vpx exchange was made by cloning the NsiI-XbaI fragment from the PBj6.6 virus into the PBj6.9 NsiI-XbaI-digested EcoRI subclone and vice versa. After confirmation by sequencing, the EcoRI fragment containing the vpx exchange into EcoRI-digested PBj6.6 and -6.9 DNA was inserted into the remainder of the clone. The correct orientation of this insert was determined by restriction digestion with the ClaI and SpeI enzymes.

(iii) gp120 exchange.

The EcoRI clones of PBj6.6 and PBj6.9 in pGEM3Zf(+) BS (described above) were digested with XbaI and NcoI. This fragment encodes the single amino acid change in gp120 that distinguishes PBj6.6 from PBj6.9. Cloning the PBj6.6 XbaI-NcoI fragment into the XbaI-NcoI-digested PBj6.9 pGEM3Zf(+) BS and vice versa achieves the exchange. After confirmation by sequencing, the EcoRI fragment now containing the gp120 V1 amino acid change (position 119 of the Env protein) was built back into the PBj6.6 and PBj6.9 parental molecular clone DNA. For both the vpx (NsiI-XbaI) and gp120 (XbaI-NcoI) fragment exchanges between the PBj6.6 and PBj6.9 viruses, only the desired coding changes were introduced, as the nucleotide sequences of the PBj6.6 and PBj6.9 viruses are otherwise identical in these regions.

(iv) Viruses 120, 127, and 137.

Mutagenesis of aspartate 119 in the V1 region of the PBj envelope to arginine (R119), producing virus 120, and to glutamate (E119), producing virus 127, was performed. Plasmid pMH98.132 (XbaI-NcoI fragment of the V1 env region cloned into pMH97.08, a modified form of the pALTER mutagenesis vector) was used with the oligonucleotides MHA34 (5′ CC TGT TAA ACC CCA TCT TCT TGT CTC ACT TTT ATT AC 3′ [nucleotides 6931 to 6895]) and MHA39 (5′ GGT GTT CCT GTT AAA CCC CAT CTC TCT GTC TCA CTT TTA TTA CAC CTC 3′ [nucleotides 6937 to 6890]) to generate viruses 120 and 127, respectively. Mutagenesis of the third glycosylation site, g3, in the PBj envelope was achieved by changing asparagine 153 to glutamine (N153Q) using oligonucleotide MHA43 (5′-AAT TTT TAT ACA AGG ATC ACT GTC CTG TAC AAC CTT TGC TGT TAT TGG 3′ [nucleotides 7040 to 6993]), producing virus 137. In each case the desired mutations were confirmed by sequencing and by the mutated restriction fragments and were reinserted into the PBj6.6 molecular clone.

Production and infectivity of chimeric PBj viruses.

Viruses were produced by transfection of 293 cells (17) by the calcium phosphate method (CellPhect Kit; Pharmacia, Piscataway, N.J.). 293 cells (5 × 10⁶) were plated 24 h prior to transfection and then transfected with 10 μg of plasmid DNA. The cells were maintained in Eagle's minimum essential medium (Biofluids) supplemented with 10% fetal bovine serum, penicillin (100 U/ml), streptomycin (100 μg/ml), glutamine (2 mM), and HEPES (10 mM). Culture supernatants collected at 48 h posttransfection were spun for 10 min to remove cell debris, filtered through a 0.45-μm-pore-size filter unit (Millipore), aliquoted, and stored at −80°C. Viral stocks were quantitated by an SIV p27 Gag antigen capture enzyme-linked immunosorbent assay (ELISA) (Retrotek) and by reverse transcriptase (RT) activity (35). The 50% tissue culture infective dose (TCID₅₀) per milliliter for each virus stock was determined on CEM 174 cells or phytohemagglutinin (PHA)-activated macaque PBMC as described previously (29, 35). To analyze the single-cycle replication of the various PBj viruses, the MAGI-CCR5 assay was performed (45), with minor modifications. The infection of 3 × 10⁴ MAGI-CCR5 cells was performed with infection medium containing 16 μg of DEAE-dextran/ml. The number of infection centers was determined by the average of three replicate wells.

Pigtailed macaque PBMC were isolated from heparinized blood on lymphocyte separation medium (ICN Biomedicals Inc., Aurora, Ohio) according to the manufacturer's instructions. PBMC were activated by the addition of PHA (Sigma, St. Louis, Mo.) to a final concentration of 0.5 μg/ml and incubated for 72 h at 37°C. The PBMC were maintained in RPMI 1640 culture medium (Biofluids) supplemented with 10% fetal bovine serum, 10% interleukin-2 (Advanced Biotechnologies Inc.), penicillin (100 U/ml), streptomycin (100 μg/ml), glutamine (2 mM), and HEPES (10 mM). Unstimulated or PHA-activated pigtailed macaque PBMC were resuspended at 10⁶ per ml in RPMI 1640 culture medium. Virus normalized for the TCID₅₀ titer was added to the cells in a 1- to 2-ml volume, and the tubes were incubated for 2 h at 37°C, with mixing every 15 min. The cells were washed twice to remove unattached virus and finally resuspended at a density of 10⁶ per ml. Cultures were maintained by the addition of fresh medium as appropriate. Samples of culture supernatant were removed at various time points to quantitate virus production by measuring RT activity.

Proliferation assay in macaque PBMC.

Lymphocyte proliferation was measured by the incorporation of [6-³H]thymidine (Amersham) into the DNA of proliferating PBMC as a consequence of activation and infection by the PBj chimeric viruses. PBMC were isolated from blood and resuspended at 10⁶ per ml in RPMI 1640 medium containing 10% human AB serum (Sigma), penicillin (100 U/ml), streptomycin (100 μg/ml), glutamine (2 mM), and HEPES (10 mM). One hundred microliters of PBMC (10⁵ cells) was placed into each well of a flat-bottomed 96-well microtiter Costar plate (Corning Inc., Corning, N.Y.). Virus normalized by either p27 antigen content, RT activity, or TCID₅₀ was added to the PBMC in a 100-μl volume. Cultures were incubated undisturbed for 5 days at 37°C. One microcurie of [6-³H]thymidine (Amersham) was added to each well, and the cultures were incubated for an additional 18 h. Cellular DNA was isolated using a Tomtec cell harvester (Wallac, Perkin-Elmer Life Sciences), and the tritiated thymidine incorporated was quantitated using a 1205 Betaplate Reader scintillation counter (Wallac, Perkin-Elmer Life Sciences). Results are expressed as a stimulation index (SI), where the average counts per minute of five replicate wells for each test sample was divided by the counts per minute for uninfected PBMC (negative control).

Cell labeling and radioimmunoprecipitation.

Transfected 293 cells were labeled at 24 h posttransfection by the addition of 150 to 200 μCi of [³⁵S]cysteine and [³⁵S]methionine (Amersham) in Dulbecco's modified Eagle's minimum essential medium (DMEM) (Biofluids) lacking cysteine and methionine. The cultures were incubated overnight at 37°C and then lysed, and viral proteins were immunoprecipitated as described below. For pulse-chase analyses, the transfected 293 cells were removed from the tissue culture flasks by scraping and were starved by incubation at 37°C for 30 min in DMEM lacking both cysteine and methionine. Five hundred microcuries of [³⁵S]-labeled cysteine and methionine was added, and the cells were incubated for an additional 30 min at 37°C. Finally, the labeled cultures were washed once and resuspended in complete medium and then incubated for various time points at 37°C.

Immunoprecipitation of radiolabeled proteins was performed either on cellular or viral lysates in 1× radioimmunoprecipitation assay (RIPA) buffer (20 mM Tris-HCl [pH 7.5], 5 mM EDTA, 30 mM NaCl, 12 mM sodium deoxycholate, and 2% NP-40). Cell lysates were precleared by the addition of 50 μl of protein A-agarose beads (Gibco BRL) and 80 μl of a 1% bovine serum albumin solution (in phosphate-buffered saline [PBS]) by incubation at 4°C for 60 min with agitation. The precleared lysate was obtained by centrifugation of the samples in a microcentrifuge. Polyclonal serum from an SIVsm-infected pigtailed macaque or a macaque monoclonal antibody specific for SIV gp120 (immunoglobulin G [IgG] 201; see reference 16) was attached to 50 μl of protein A-agarose beads by incubation in PBS for 60 min at 4°C. After one wash with PBS, the agarose-antibody beads were incubated with 0.5 to 1 ml of cell or viral lysate at 4°C for 1 to 2 h with agitation. The cellular samples were then washed five times in 1× RIPA buffer, and the viral lysates were washed three times. The immunoprecipitated proteins were prepared for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis by the addition of an equal volume of 2× Laemmli loading dye. The samples were boiled for 4 min, rapidly placed into ice, centrifuged in a microcentrifuge for 2 min, and loaded onto the gel. After electrophoresis, the 10% polyacrylamide gels were fixed by immersion in a 30% methanol–10% acetic acid solution for 2 h, washed briefly in water, and enhanced by a 25-min incubation in ENLIGHTNING solution (NEN Life Science Products). The gels were dried under a vacuum at 60°C for 2 to 3 h, and protein bands were visualized by autoradiography.

Spontaneous shedding of gp120 from virions.

293 cells transfected with the PBj molecular clones were labeled at 24 h posttransfection with 200 Ci of [³⁵S]cysteine and [³⁵S]methionine overnight. The culture supernatant was removed and filtered through a 0.45-μm-pore-size filter unit (Millipore) and then centrifuged at 160,000 × g for 50 min at 4°C using a Beckman TL55S rotor. The resulting virus pellet was resuspended in 1 ml of fresh Eagle's minimum essential medium. Half of this sample was repelleted immediately, and the virus pellet and supernatant fractions were immunoprecipitated. The other half of the sample was incubated for 48 h at 37°C and centrifuged, and the pellet and supernatant fractions were similarly analyzed. The release of gp120 spontaneously shed into the supernatant from the virion was quantitated as the percentage of the total gp120 present at that time point [(free gp120)/(gp120 on the virion) + free gp120].

Animal studies.

Four groups of three juvenile pigtailed macaques (Macaca nemestrina) were inoculated intravenously with 2.15 × 10³ TCID₅₀ of PBj6.6, PBj6.9, or the gp120 exchange viruses PBj060 and PBj063 per ml. The animals were monitored closely for the clinical signs of PBj-induced disease. Monkeys were maintained in accordance with the guidelines of the Animal Care and Use Committee of the National Institutes of Health and the Guide for the Care and Use of Laboratory Animals. Sequential plasma samples were assayed for p27 antigenemia (p27 antigen ELISA; Retrotek), and viral RNA levels (25) and lymphocyte subsets were assessed daily by flow cytometric analysis (Fast Systems, Rockville, Md.). Lymph node biopsies were obtained at day 5 postinfection. Half of the lymph node was formalin fixed and embedded in paraffin for in situ hybridization. The remaining tissue was separated into a single-cell suspension by passage through a Costar cell strainer (Corning Inc.). Serial 10-fold dilutions of the lymph node mononuclear cells (LNMCs) were cocultured with uninfected CEM 174 cells to estimate the virus load in these tissues. Virus isolation was performed weekly from 7 ml of blood by cocultivation of PHA-activated PBMC with uninfected CEM 174 cells using the supernatant RT activity as a readout for positive virus isolation. Animals were euthanatized and necropsied if they developed signs of acute disease, such as diarrhea, dehydration, lethargy, and anorexia. Histopathologic analyses were performed on lymphoid tissues and the gastrointestinal tract, and SIV-specific in situ hybridization was performed on these tissues as previously described (19, 20). Animals that survived the acute phase of the PBj challenge were euthanatized 3 to 4 weeks after inoculation, and complete necropsies were performed.

RESULTS

The molecularly cloned SIVsmPBj6.6 virus has been well characterized as a highly pathogenic virus in pigtailed macaques (20, 33). In contrast, the SIVsmPBj6.9 virus cloned from the same biological isolate of SIVsmPBj appears to be considerably less pathogenic (33). Both viruses replicate in activated macaque PBMC and monocyte-derived macrophages, but they differ significantly in the ability to induce proliferation of resting macaque PBMC in vitro (33). The difference in virulence of these two cloned viruses is intriguing since sequence comparison reveals only five amino acid differences between these viruses (Fig. 1B). Interestingly, the viruses share a tyrosine substitution at position 17 of the Nef protein which has been implicated in the unique pathogenesis of SIVsmPBj. Three of the substitutions involved changes in a single gene product (C89R in Vpx, D119G in Env, and F252L in Nef), whereas two adjacent changes mapped to the region of overlap between the coding regions of Env and Nef. To determine the amino acid substitutions responsible for the difference in pathogenesis between PBj6.6 and PBj6.9, chimeric viruses were constructed as shown in Fig. 1. A total of eight chimeras were generated that exchanged nef, vpx, the gp120 portion of env, and the nef-gp41 substitutions. Virus stocks of chimeric and parental strains were generated by calcium phosphate-mediated transfection of 293 cells and were normalized for viral input by SIV p27 antigen content as well as TCID₅₀ for subsequent studies of their in vitro properties.

FIG. 1.

(A) Schematic of the SIVsmPBj virus genome showing the genomic organization. (B) The 3′ half of the PBj genome and locations of the five amino acid differences that distinguish PBj6.6 and PBj6.9 viruses, along with the locations of the restriction sites used to make the amino acid exchanges. (C) Schematic representation of the chimeric and point-mutated PBj viruses. PBj6.6 virus is represented by a white box, and PBj6.9 is represented by a black box. The nef, vpx, and gp120 exchange viruses are numbered and shown schematically, reflecting the exchange of amino acids between the two parental clones.

Increased infectivity of SIVsmPBj6.6 relative to SIVsmPBj6.9.

Each of the virus stocks was characterized for RT activity, SIV p27 antigen content, infectivity in MAJI-CCR5 cells, and TCID₅₀ in CEM 174 cells. As shown in Table 1, PBj6.6 produced an average of sixfold more SIV-expressing cells in a single cycle of infection in the MAGI-CCR5 cell line (45) than were observed for a similar RT input of the less pathogenic PBj6.9 virus. Addionally, the specific infectivity of the PBj6.6 and PBj6.9 virus preparations was determined by calculating the infectivity to particle ratio (TCID₅₀/p27 ratio) of each virus. Interestingly, we observed that the ratio between the infectivity of the virus and viral antigen content differed significantly between PBj6.6 and PBj6.9. Specifically, the infectivity of the highly pathogenic PBj6.6 was approximately 10-fold higher than that measured for PBj6.9 (19.2 versus 1.9).

TABLE 1.

Relative infectivity of SIVsmPBj6.6, -6.9, and chimeric viruses 060 and 063 in MAGI-CCR5 cells^a

Virus	Infectivity (avg no. [± SD] of blue cells)b
Medium	5 ± 1
PBj6.6	140 ± 14
PBj6.9	23 ± 3
060	35 ± 6
063	150 ± 14

^aMAGI-CCR5 cells (3 × 10⁴) were infected with 2 × 10⁵ cpm of RT activity for each virus or with medium only. 

^bThe average number of blue-stained infection centers for three replicate wells for each virus is shown. The negative control was medium only. 

The effect of different exchanges upon the infectivity of each of the chimeric viruses was also measured. For the majority of the chimeras, the parental backbone of the virus (PBj6.6 or PBj6.9) appeared to confer the same relative difference in infectivity between chimeric pairs. Therefore, if the majority of the virus consisted of PBj6.9, the chimera had reduced infectivity relative to the reciprocal PBj6.6 chimera. All of the viruses with the D119G gp120 substitution proved to be exceptions to this rule. The specific infectivity for virus 060 (PBj6.6 virus with PBj6.9 gp120) was reduced to PBj6.9 levels (TCID₅₀/p27 ratio of 1.4), and for virus 063 (PBj6.9 virus with PBj6.6 gp120) the ratio was increased to 16.6. The difference in this ratio between reciprocal gene exchanges was unaffected by the exchange of the nef or vpx genes (data not shown). These data suggest that the D119G substitution in the PBj6.9 envelope is responsible for reducing the specific infectivity of this virus relative to the virulent PBj6.6.

Proliferation of PBMC in vitro.

The ability of SIVsmPBj viruses to induce activation and proliferation of resting PBMC in vitro has previously been used as an indicator of pathogenesis in vivo (33, 40). The ability of the parental viruses and the chimeric viruses to induce proliferation of resting PBMC was assessed by thymidine incorporation following 5 days of incubation with virus that had been normalized for viral input by either viral antigen (10 ng of SIV p27) or TCID₅₀. Both methods for normalizing virus input were used, since the specific infectivity of the virus stocks differed widely, as described above. Using a constant input of infectious virus, the SIs for the parental viruses and eight chimeras were indistinguishable (data not shown).

However, using antigen levels of input virus for normalization resulted in significant differences among the parental strains and the chimeras, as shown graphically in Fig. 2. As previously reported (33), the PBj6.6 virus was clearly more effective in inducing PBMC proliferation than PBj6.9. This differential between the SI profiles of reciprocal chimeras remained constant with the nef chimeras (018 and 020) and the vpx chimeras (034 and 035). In both cases, greater proliferation was observed with the viruses having the SIVsmPBj6.6 background. However, there was a marked change in the profile observed for the chimeras exchanging the gp120 region (060 and 063). The SI for virus 060 (PBj6.6 virus containing PBj6.9 gp120) was considerably lower than that for the wild-type PBj6.6 virus (Fig. 2). Accordingly, the SI for virus 063 (PBj6.9 virus containing PBj6.6 gp120) now showed an increase compared to the SI typical for the PBj6.9 virus. An analysis of double exchange viruses 077 (PBj6.6 virus with PBj6.9 gp120 and PBj6.9 nef) and 088 (PBj6.9 virus with PBj6.6 gp120 and PBj6.6 nef) further suggested that the change in PBMC proliferation profile was dependent on the gp120 exchange only. This result suggested that the extent of PBMC proliferation was determined by the substitutions at position 119 in the envelope gene of these viruses.

FIG. 2.

Proliferation profile of macaque PBMC following infection with the PBj viruses and the indicated chimeras. SIs for each of the viruses above background levels obtained with medium alone are shown in a bar graph format, with standard deviations indicated. PBMC were mixed with medium alone (m), 10 ng of each virus, or PHA. The cultures were incubated at 37°C for 5 days, and on day 6 the uptake of 1 μCi of tritiated thymidine into the proliferating cells was determined for five replicate wells by cell harvesting.

Kinetics of infection in activated and resting PBMC.

To identify differences in infectivity, as suggested from the relative abilities of PBj6.6 and PBj6.9 viruses to induce PBMC proliferation, the growth kinetics of each virus were examined in unstimulated or PHA-stimulated pigtailed macaque PBMC. All viruses replicated equivalently in PHA-stimulated PBMC (data not shown). All of the viruses also replicated in unstimulated PBMC; however, the replication of PBj6.6 was more rapid and peak virus production was higher than for PBj6.9 (Fig. 3). Both of the envelope exchange viruses replicated at an intermediate level when compared to the PBj6.6 and PBj6.9 viruses. The replication of virus 060 (PBj6.6 virus containing PBj6.9 gp120) was reduced compared to wild-type PBj6.6 virus. Correspondingly, the replication of virus 063 (PBj6.9 virus containing PBj6.6 gp120) was enhanced compared to the parental PBj6.9 virus. As the viral replication analysis was performed on unstimulated PBMC, the resultant profiles are representations of the ability of the viruses to both replicate in and induce the proliferation of PBMC. This analysis indicated that for the replication of the PBj6.6 and PBj6.9 viruses in PBMC in vitro, the identity of the amino acid present at position 119 of the gp120 protein was an important factor in determining the growth kinetics of each virus.

FIG. 3.

Replication kinetics of the PBj viruses and the chimeras in unstimulated macaque PBMC. Production of supernatant RT activity following infection of unstimulated macaque PBMC with PBj6.6, PBj6.9, 060, and 063 viruses is plotted sequentially. A total of 1,200 TCID₅₀ of each virus was used to infect 0.5 × 10⁶ unstimulated macaque PBMC. The culture supernatant was sampled every two days, and the RT activity was determined for the PBj6.6 virus.

The D119G substitution is associated with a processing defect in Env.

Since the aspartate (D119) in the V1 region of the PBj envelope protein appeared to be involved in PBMC proliferation, viral replication, and infectivity, we evaluated its effect on gp160 processing and the interaction between gp120 and gp41. 293 cells transfected with various molecular clones were labeled with 100 to 150 μCi of [³⁵S]methionine and [³⁵S]cysteine, and viral proteins were radioimmunoprecipitated from cell and viral lysates. Immunoprecipitated proteins were resolved by gel electrophoresis to visualize gp160 and gp120 envelope proteins associated with the cells, gp120 incorporated into virion particles, and the non-virion-associated, or free, gp120.

A comparison of the gp160 processing pattern for PBj6.6 and PBj6.9 viruses (Fig. 4, lanes 1 to 4) demonstrated a difference in the ratio between gp120 and gp160 protein in cell lysates. gp160 processing of PBj6.6 virus appeared to be efficient, as evidenced by the presence of significant amounts of gp120. In contrast, significantly less gp120 than gp160 protein was observed in cell lysates of the PBj6.9-producing cells. This suggested that there is a defect in gp160 processing of PBj6.9 or a reduced ability of the PBj6.9 gp120 protein to remain associated with its gp41 protein at the plasma membrane. The pattern of gp160 processing of the chimeric virus that expressed the PBj6.9-specific glycine at position 119 in Env (lanes 5 and 6) was similar to that of the PBj6.9 virus (lanes 3 and 4). Correspondingly, replacement of the glycine with aspartate at amino acid 119 in PBj6.6 (063) produced a processing pattern more like that of PBj6.6 than the parental PBj6.9 (compare lanes 7 and 8 with lanes 1 and 2). Therefore, the identity of the amino acid at position 119 of the envelope protein appeared to determine the pattern of gp160 processing for each virus. Viruses with nef exchanges had unaltered gp160 processing profiles. Thus, virus 018 (PBj6.6 with PBj6.9 nef) showed similar processing to that of the PBj6.6 virus. In a like manner, virus 020 (PBj6.9 virus with PBj6.6 nef) showed a gp160 processing efficiency similar to that of the parental PBj6.9. The env gene of SIV overlaps the nef reading frame; therefore, exchanging the entire nef gene also exchanged the codons for amino acids 871 and 872 of the envelope for each virus. This exchange had not affected the gp160 processing profile, further suggesting that it was the character of the amino acid at position 119 of the gp120 protein that influenced envelope polyprotein processing.

FIG. 4.

gp160 env precursor processing in 293 cells transfected with the PBj molecular clones. ³⁵S-labeled cell lysates were obtained at 48 h posttransfection and analyzed by RIPA with a polyclonal serum (PCS; odd lanes) or Fab 201 IgG to SIVsm env (even lanes) (16). Electrophoresis was performed on a 10% polyacrylamide–SDS gel. Lanes 1 and 2, PBj6.6; lanes 3 and 4, PBj6.9; lanes 5 and 6, virus 060; lanes 7 and 8, virus 063; lanes 9 and 10, virus 018; lanes 11 and 12, virus 020. The locations of gp160 and gp120 are indicated, along with the molecular mass standards (M/kD) shown (Rainbow Markers; Promega).

Pulse-chase analysis of gp160 polyprotein processing.

To examine the kinetics of gp160 processing, pulse-chase analysis was performed on 293 cells at 48 h posttransfection with the various PBj molecular clones. The transfected cells were pulse-labeled for 30 min with 500 μCi of ³⁵S-labeled methionine and cysteine and chased by the addition of complete medium for various time periods. Samples were obtained every 2 h to analyze the viral envelope proteins associated with the cellular material, and the relative amounts of free gp120 in the culture supernatant are shown in Fig. 5.

FIG. 5.

Pulse-chase analysis of gp160 processing for 293 cells transfected with the PBj and gp120 exchange viruses. Samples were obtained at 0, 2, 4, 6, and 8 h and analyzed by RIPA. (A) cell-associated gp160 processing. SIVsm Fab 201 IgG was used to precipitate the viral envelope proteins. Lanes 1 to 5, PBj6.6 virus; lanes 6 to 10, PBj6.9 virus; lanes 11 to 15, virus 060; lanes 16 to 20, virus 063. (B) Immunoprecipitation using a polyclonal serum of the free gp120 in the supernatant at each time point of the pulse-chase analysis. Lanes 1 to 5, PBj6.6 virus; lanes 6 to 10, PBj6.9 virus; lanes 11 to 15, virus 060; lanes 16 to 20, virus 063. The locations of the gp160 and gp120 proteins are indicated, along with the Rainbow markers (M/kD) (Promega).

The processing of gp160 by PBj6.6 in transfected cells was efficient, leading to significant quantities of gp120 within 2 h (Fig. 5A, lanes 1 and 2). The relative amount of gp160 decreased at later time points as it was processed to gp120 protein (lanes 3 to 5). In comparison, although PBj6.9 gp120 was also observed by 2 h, the relative amount of gp160 remained higher than the amount of gp120, suggestive of a reduction in gp160 processing. These gp160 processing patterns were reproduced in the envelope exchange viruses. For virus 060 (PBj6.6 virus containing PBj6.9 gp120) a processing pattern like that of the PBj6.9 virus was observed (lanes 11 to 15 compared to lanes 6 to 10). Conversely, for virus 063 (PBj6.9 virus containing PBj6.6 gp120), the gp160 processing resembled that of the PBj6.6 virus (lanes 16 to 20 compared to lanes 1 to 5), as gp120 protein remained detectable in the cell lysate. Thus, it appeared that the single amino acid change of aspartate to glycine (in virus 060) was sufficient to produce gp160 processing that resembled that of the PBj6.9 virus. Correspondingly, when the G119 residue was restored to D119 for virus 063, a PBj6.6 gp160 processing pattern resulted.

The release of free, non-virion-associated gp120 into the culture supernatant was also examined at each time point of the pulse-chase assay (Fig. 5B). Supernatant samples were first centrifuged to remove the virions by pelleting, and then a RIPA was performed on the supernatant fraction. The amount of free gp120 shed from the cells (and virions) over the 8-h chase period is shown in Fig. 5B. There was an increase in the amount of free gp120 released into the supernatant with time for PBj6.9 virus-producing cells compared to that observed for cells producing PBj6.6 virus (Fig. 5B, lanes 6 to 10 compared with lanes 1 to 5). This relative level of free gp120 was dependent on the amino acid present at position 119 of the gp120 protein. For virus 060 (glycine replacing aspartate) there was an increased amount of free gp120 compared to the gp120 present for virus 063 (aspartate replacing glycine) over the 8 h examined (Fig. 5B, compare lanes 11 to 15 with lanes 16 to 20). A RIPA on the pelleted virions produced during the time course indicated that similar levels of virus were produced, suggesting that the observed results were not a reflection of varying transfection efficiencies. These results suggested that there was a difference in the noncovalent association between the gp120 and gp41 envelope protein subunits for the PBj6.6 and PBj6.9 viruses. A glycine at position 119 of gp120 in PBj6.9 appeared to destabilize the interaction between the gp120 and gp41 proteins, leading to a reduction in the amount of gp120 remaining associated with gp41 at the plasma membrane of the cell. Correspondingly, higher levels of free gp120 were detected in the supernatant from PBj6.9-producing cells than that observed from cells producing the PBj6.6 virus.

Decreased virion-associated gp120 and increased spontaneous shedding of PBj6.9 gp120.

To further examine the association between the gp120 and gp41 envelope subunit and transmembrane proteins, spontaneous gp120 shedding from radiolabeled virions was examined. As previously reported for HIV-1 (31, 38), radiolabeled virions were isolated from transfected culture supernatants by centrifugation, resuspended in PBS, and divided into two equal volumes. The viruses were incubated at 37°C for 0 or 48 h and then repelleted by centrifugation to separate free gp120 in the supernatant from the gp120 associated with the virion (virus pellet). A RIPA was performed on each fraction as shown in Fig. 6, analyzed by SDS-PAGE, and quantitated by PhosphorImager analysis. The amount of free gp120 in the supernatant was expressed as a percentage of the total gp120 (free gp120 plus virion-associated gp120) present at that time point.

FIG. 6.

Spontaneous shedding of gp120 from the PBj parental and gp120 exchange viruses. ³⁵S-labeled virions were pelleted from transfected 293 cells and resuspended in medium. This suspension was split into two, and each half was incubated for 0 or 48 h at 37°C and then repelleted. The free gp120 released into the supernatant (SN; even-numbered lanes) and that remaining associated with the virion (odd-numbered lanes) were analyzed by RIPA using a polyclonal serum. Lanes 1 to 4, PBj6.6 virions; lanes 5 to 8, PBj6.9 virions; lanes 9 to 12, 060 virions; lanes 13 to 16, 063 virions. The percentage of free gp120 present at each time point is shown, along with the molecular weight markers (Promega).

As shown in Fig. 6, there was an increase in spontaneous gp120 shedding by PBj6.9 (18.4 to 36.9%) compared to that by PBj6.6 (8.1 to 29.5%). Similar results were obtained for the gp120 exchange viruses (Fig. 6, lanes 9 to 16). For virus 060 (PBj6.6 virus with PBj6.9 gp120) the amount of free gp120 at time zero and 48 h later was now more similar to that for PBj6.9 virus than for PBj6.6, a direct result of replacing aspartate 119 with glycine. Correspondingly, virus 063 (PBj6.9 virus with PBj6.6 gp120) now demonstrated a pattern virtually identical to that of wild-type PBj6.6 virus. Interestingly, for both the PBj6.9 and 060 viruses (G119) there was an increased amount of free gp120 present at time zero when compared to that for the PBj6.6 and 063 viruses. These results demonstrated that the identity of the amino acid at position 119 of gp120 was important for determining the stability of the gp120-gp41 association on the virion envelope, with the PBj6.9 virus being more likely to spontaneously shed gp120 than PBj6.6.

The relative amounts of gp120 present on PBj6.6 and -6.9 virions purified through a sucrose gradient, rather than pelleted culture supernatant, was also estimated. Radiolabeled virions were centrifuged through a 20 to 60% sucrose gradient, fractionated, and resolved by PAGE. To determine the relative amount of gp120 on the purified virions, the gp120 band was quantitated by phosphorimager analysis as a percentage of the amount of p27 protein present in the peak fractions from the gradient. This percentage was 2.4% for PBj6.6 and 1.3% for PBj6.9 virions (data not shown). This result is consistent with earlier observations and suggests an approximately twofold reduction in the amount of gp120 protein on PBj6.9 compared to that on PBj6.6 virions.

Consistent with previous studies of soluble CD4 (sCD4)-induced shedding of gp120 for SIV (36), there was no increase in gp120 shedding induced by incubation of PBj6.6 or PBj6.9 virions with increasing concentrations of sCD4 protein (data not shown). Therefore, the relative affinity for CD4 binding of the PBj6.6 and PBj6.9 envelopes as they were present on the virion could not be directly determined.

Mutagenesis of aspartate 119 in the PBj gp120 protein.

Sequence analysis of viral quasispecies during the development of the SIVsmmPBj14 virus (43) revealed that the aspartate residue at position 119 of Env arose concurrently with the development of pathogenicity. To further characterize the role of residue 119 within the gp120 protein, D119 of the PBj6.6 virus was changed to arginine to give rise to virus 120 (R119) and to glutamate to give virus 127 (E119). The effect of these point mutations on Env polyprotein precursor processing and the association of the gp120-gp41 interactions in cell lysates was then evaluated as previously described (Fig. 7A).

FIG. 7.

gp160 processing and gp120-gp41 association for the parental PBj and mutated viruses. (A) RIPA of gp160 processing in cell lysates. Odd-numbered lanes were immunoprecipitated with a polyclonal serum, and even-numbered lanes were immunoprecipitated with the SIVsm Fab 201 IgG. PBj6.6 virus (lanes 1 and 2), PBj6.9 virus (lanes 3 and 4), and the point-mutated viruses 120 (D119R; lanes 5 and 6) and 127 (D119E; lanes 7 and 8) are shown. (B) RIPA for virion pellets of PBj6.6 virus, PBj6.9 virus, virus 120, and virus 127 is shown in lanes 1 to 4, respectively. (C) Free gp120 protein, a combination of that shed from virions and from the surfaces of transfected cells, is shown (lane designations are the same as for panel B). For both panels B and C a polyclonal serum was used to immunoprecipitate the viral proteins. The migration of molecular mass markers (M/kD) is also indicated (Promega).

Comparing relative amounts of gp160 and gp120 revealed a reduction in the amount of gp120 associated with the cellular material for the PBj6.9 virus (Fig. 7A, lanes 3 and 4) when compared to the PBj6.6 virus (lanes 1 and 2), as seen earlier. Introduction of the positively charged arginine at position 119 of PBj6.6 (virus 120) shifted the distribution of gp160 and gp120 (lanes 5 and 6) to a pattern similar to that observed for the PBj6.9 virus. Introduction of a negatively charged residue at this position, E119 in virus 127 (lanes 7 and 8), produced a pattern of gp160 processing similar to that of the PBj6.6 virus (lanes 1 and 2).

The distribution of the virion-associated and free gp120 for the mutated viruses is shown in Fig. 7B and C, respectively. No significant difference was observed in the levels of gp120 incorporation into virions of mutants 120 and 127. In contrast, the level of free gp120 present was increased for PBj6.9 (lane 2) and virus 120 (R119) (lane 3). Consistent with earlier observations, there was a reduced level of free gp120 for both the PBj6.6 (lane 1) and 127 (E119; lane 4) viruses when a negatively charged amino acid was present in position 119 of the gp120 protein.

The results suggested that the character of the amino acid at position 119 of gp120 was an important determinant of the gp120-gp41 association at the cell surface. This also has consequences for gp120 incorporation into the virus particle and for spontaneous shedding of gp120 from the virion. The most likely explanation for this behavior was structural changes introduced into the envelope proteins due to each amino acid. Therefore, it is likely that a negatively charged residue at position 119 of gp120 is important for preserving optimum envelope structure and function.

Differences in pathogenesis of parental PBj6.6 and PBj6.9 strains.

Three pigtailed macaques were inoculated intravenously with 2 × 10³ TCID₅₀s of either PBj6.6, PBj6.9, 060, or 063 virus to determine whether the defects in env biology identified in vitro altered pathogenesis in vivo. All the animals became infected, as evidenced by the rescue of infectious virus from their PBMC (Table 2). As expected, animals inoculated with PBj6.6 virus (1163, 7108, and 7109) developed diarrhea, generalized skin rash, lethargy, and appetite loss. The clinical condition of these animals necessitated euthanasia at 6 (1163 and 7108) and 8 (7109) days. A high virus load in the lymph nodes was indicated following the isolation of virus from as few as 1,000 LNMCs obtained at day 5 by a lymph node biopsy. Consistent with this high virus titer, rapidly increasing levels of viral RNA levels were measured in the plasma associated with a rapid decline in the total lymphocyte count (Fig. 8 and 9). Peak plasma viral RNA levels of 10⁷ to 10⁹/ml were achieved by days 5 to 7 after inoculation. Histopathologic analysis of tissues of these animals revealed lesions characteristic of SIVsmPBj. Marked pathologic changes observed in the intestine included massive infiltration of mononuclear cells into the lamina propria and gastrointestinal-associated lymphoid tissues, villus blunting, fusion, and ulceration of the epithelial surface of the intestinal tract (Fig. 10). These changes were associated with large numbers of SIV-expressing cells detected by in situ hybridization (>100 SIV-positive cells per high-power field). A similarly large number of SIV-expressing cells was observed in the mesenteric lymph nodes.

TABLE 2.

Infection of pigtailed macaques in vivo with PBj6.6, -6.9, and the gp120 chimeras 060 and 063

Virus	Animal designation	Disease outcome (time)	Plasma p27 antigenemia (ng/ml)a at day:						Virus isolation by coculture of:
			5	6	7	8	9	10	PBMCb	LNMCc
PBj6.6	1163	Death (day 6)	1.8	35.3					+	103
	7108	Death (day 6)	0.4	6.4					+	103
	7109	Death (day 8)	0.4	8.7	29.4	53.4			+	103
PBj6.9	1154	Euthanatized (week 3)	−	−	−	−	−	0.2	+	x
	7103	Euthanatized (week 3)	−	−	−	−	−	−	+	x
	7122	Euthanatized (week 3)	−	−	−	0.7	−	−	+	x
060	358	Euthanatized (week 3)	−	−	−	−	−	−	+	105
	1152	Euthanatized (week 3)	−	−	−	−	0.2	0.1	+	105
	93P027	Euthanatized (week 3)	−	−	−	−	−	−	+	104
063	1031	Death (day 10)	−	−	−	4.0	5.4	4.7	+	104
	7100	Death (day 11)	−	−	−	−	0.2	0.3	+	104
	93P009	Euthanatized (week 4)	−	−	−	0.7	2.8	2.8	+	104

^ap27 antigenemia was determined by an antigen capture ELISA (Retrotek) and is shown for the days indicated after inoculation of the virus into the animal. −, a titer of less than 0.1 ng/ml. 

^b+, cocultures of PBMC isolated from experimentally infected animals with CEM 174 cells became positive for RT activity. 

^cThe minimum number of LNMCs from which RT activity was isolated by coculture with CEM 174 cells maintained for a 4-week period. x, no virus was recovered from the coculture of up to 10⁵ LNMCs for PBj6.9 virus. 

FIG. 8.

Plasma viremia in pigtailed macaques following infection with parental PBj6.6, PBj6.9, and chimeras 060 and 063 is shown graphically. Deaths of animals are indicated by crosses.

FIG. 9.

Absolute peripheral lymphocyte counts following inoculation of pigtailed macaques with parental PBj6.6, PBj6.9, and chimeras 060 and 063. Deaths of animals are indicated by crosses.

FIG. 10.

Pathologic lesions observed in macaques infected with the highly pathogenic PBj6.6 virus (A and B) and the envelope chimera 063 (C and D). Representative fields of hematoxylin and eosin-stained sections of the ileum and the same sections hybridized with an SIV-specific digoxigenin-labeled antisense probe are shown for two representative animals. Animal 1163 was inoculated with PBj6.6 (A and B), whereas animal 1031 was inoculated with the 063 chimera (C and D). Characteristic blunting and fusion of villi and infiltration of mononuclear cells are seen in both animals. More erosive features are evident in intestinal sections of the PBj6.6-inoculated macaque, and the level of virus expression was higher than that observed in the 063-inoculated macaque.

In contrast, the pathogenesis of the PBj6.9 virus was attenuated in macaques 1154, 7103, and 7122. Although these animals became infected, as evidenced by virus rescue from PBMC, plasma antigenemia was markedly reduced compared to that in animals inoculated with PBj6.6 (Table 2). Virus load in the lymph nodes at day 5 was lower than that for PBj6.6-infected animals, as virus remained undetected from a coculture of up to 10⁵ LNMCs with CEM 174 cells. Correspondingly, the kinetics of plasma viremia were slower than those observed in PBj6.6-inoculated macaques. Plasma viral RNA reached peak levels of 10⁵ to 10⁷/ml by day 12, by which time all of the PBj6.6-inoculated macaques had been euthanatized. Only a transient decrease in total lymphocyte count was observed in PBj6.9-infected animals, and all three recovered from the acute phase of the infection (Fig. 8). These animals remained free of any observable symptoms of PBj infection and were euthanatized at 3 weeks. No specific lesions were observed by histopathology. These results were similar to results observed in an earlier PBj6.9 animal infection study (33), confirming the more attenuated pathogenesis of this virus.

Comparison of gp120 chimeras of PBj6.6 and PBj6.9.

Inoculation of three animals (358, 1152, and 93P027) with virus 060 (PBj6.6 virus with PBj6.9 gp120; D119G) resulted in infection, as evidenced by the recovery of infectious virus from PBMC (Table 2). Coculture of 10⁴ to 10⁵ LNMCs with CEM 174 cells allowed recovery of virus 060 from the lymph node tissue of these animals. All three animals survived the acute phase of infection without any clinical symptoms and were euthanatized at the end of the experiment. The kinetics of plasma viremia in these animals were similar to those observed for PBj6.9-inoculated macaques, and only a transient lymphopenia was observed (Fig. 9). Therefore, the pathogenesis of 060 appeared to be very similar to that observed with PBj6.9 virus. Clearly, changing a single amino acid, aspartate 119 in PBj6.6 to glycine (producing virus 060), had a dramatic effect on the pathogenesis of infection in vivo.

Infection of animals (1031, 7100, and 93P009) with the reciprocal chimera, virus 063 (PBj6.9 virus with PBj6.6 gp120; G119D), led to a consistent p27 antigenemia in two of the three animals. Peak plasma viral RNA levels were achieved by day 8 and ranged from 10⁶ to 10⁸/ml. Tissue-associated virus load (10⁴ LNMCs) was higher than for macaques inoculated with either PBj6.9 or 060 virus. The kinetics and levels of plasma viremia for one of the animals (1031) were indistinguishable from those observed in PBj6.6-inoculated macaques (Fig. 8). Animals 1031 and 7100 showed some lymphopenia (Fig. 9), although not to the same extent as that seen in PBj6.6-infected animals, and were euthanatized at days 10 and 11 due to typical PBj symptoms. Interestingly, animal 93P009 had a detectable antigenemia and a similar lymphopenia to the other 063-infected animals but survived the acute phase of infection and was later euthanatized at 4 weeks. Sequence analysis of the virus isolated from this animal at two time points did not show any reversion of the aspartate at position 119 (data not shown).

The pathologic changes induced by the 063 mutant (Fig. 10A and B) were similar to those observed for PBj6.6 but were significantly less severe and lacked the ulcerative features (Fig. 10C and D). The number of SIV-expressing cells observed in the mesenteric lymph nodes or affected areas of intestine was reduced at least 10-fold (5 to 20 per high-power field) compared with representative sections from macaques inoculated with PBj6.6. Thus, the pathogenesis appeared to be similar for these two viruses, but the pathology induced by the 063 mutant was reduced compared to that induced by PBj6.6. These results indicated that replacing glycine 119 in the PBj6.9 virus with aspartate (producing virus 063) produced a more pathogenic virus than the PBj6.9 parental virus. However, virus 063 remained less pathogenic than the PBj6.6 virus, suggesting that other amino acid substitutions may need to be introduced into the PBj6.9 background to completely reconstruct PBj6.6 pathogenesis.

DISCUSSION

The construction and evaluation of chimeric viruses are a valuable approach which has provided new insights into viral pathogenesis. In an earlier study based on SIVsmPBj6.6 and SIVsmmH4 chimera construction, env in combination with an additional unknown gene was identified as a determinant of pathogenesis (33). We have extended these observations by the analysis of chimeras derived from two highly related PBj parental viruses to more precisely identify sequence differences that specify the phenotype of each virus. Chimeric viruses were constructed between the acutely pathogenic PBj6.6 and less pathogenic PBj6.9 viruses. Of the five amino acid differences that distinguish these parental viruses, the single amino acid exchange of aspartate to glycine at position 119 of gp120 (D119G) was identified as an important component of PBj pathogenesis. The present study demonstrates the importance of the envelope glycoprotein of primate lentiviruses such as SIV in viral pathogenesis. A single substitution within gp120 of the SIVsmPBj6.9 molecular clone was responsible for reducing the relative infectivity of viral particles in vitro and completely abrogated the virulence of SIVsmPBj6.6 in vivo. This specific mutation was clearly responsible for the attenuated phenotype of the less pathogenic SIVsmPBj6.9 molecularly cloned virus. However, this PBj6.6 gp120 substitution alone was not sufficient to restore full virulence to SIVsmPBj6.9. Therefore, SIVsmPBj6.9 must contain an additional attenuating mutation(s) relative to SIVsmPBj6.6.

The retroviral envelope glycoprotein is an important determinant of both viral replication and tropism. Env is synthesized as a gp160 polyprotein precursor which is glycosylated and then cleaved by a host protease, producing the extracellular gp120 subunit and the gp41 transmembrane viral protein. These proteins remain associated together by a noncovalent interaction and become incorporated into nascent virions at the plasma membrane of the cell (reviewed in reference 11). Previous reports have shown that amino acid changes within the V1-V2 region of the HIV-1 envelope glycoprotein affect both gp160 polyprotein precursor processing and the association between the gp120 and gp41 cleavage products (11, 28, 42). However, it is uncertain what the effect of altered envelope processing or changes in the affinity of the gp120-gp41 interaction may have on the pathogenesis of the virus in vivo.

Introduction of the D119G exchange into the PBj6.6 virus led to a reduction in virus infectivity as reflected in slower replication kinetics in MAGI-CCR5 cells and PBMC in vitro (Fig. 2 and 3 and Table 1). Correspondingly, introduction of the reciprocal G119D exchange into the PBj6.9 virus increased the infectivity of the chimera relative to the parental PBj6.9 virus. The stability of the gp120-gp41 interaction was analyzed for the PBj viruses. The presence of aspartate at position 119 in the PBj6.6 and 063 viruses produced an optimal gp160 processing pattern in cells, with the incorporation of gp120 into virions as well as a quantity of shed, or free, gp120 (Fig. 4, 5, and 6). When glycine replaced aspartate in viruses PBj6.9 and 060, gp160 processing was less efficient, with a reduction in the amount of gp120 remaining associated with the surface of the virus-producing cells and a corresponding increase in the level of free gp120 shed from cells and virions.

The decrease in the stability of the gp120-gp41 association probably accounts for the reduced ability of PBj6.9 and 060 viruses to infect cells relative to that of PBj6.6 and 063 (Fig. 3 and Table 1). Reduced infectivity due to increased shedding of gp120 has been described for HIV-1 (31). Altered physical and functional properties of gp120 have been shown to directly affect viral infectivity involving subunit association (42, 44, 49). Attempts to better characterize the affinity of the PBj gp120-gp41 interaction by analyzing sCD4-induced shedding of PBj gp120 were unsuccessful (data not shown), as sCD4 does not induce gp120 shedding for SIVs, unlike HIV-1 env (36). Although sCD4 has been shown to increase the infectivity of SIVs (1), recent studies indicate that the mechanism may be by the sCD4-induced increase in binding of gp120 to CCR5 rather than by exposure of the fusogenic domain in gp41 as a result of gp120 shedding as proposed for HIV-1 (36).

Comparing the mutated viruses 120 (R119) and 127 (E119) to PBj6.6 (D119) and PBj6.9 (G119), it was clear that a negatively charged amino acid was preferred at this position of gp120 (Fig. 7). The consequences of replacing aspartate (PBj6.6) with the small and highly polar glycine (PBj6.9) as well as with the oppositely charged arginine (virus 120) would be expected to alter the structure of the local V1 region or perhaps the entire gp120 protein. When the more conservative glutamate substitution was made (virus 127), gp160 processing efficiency and the gp120 distribution profile were similar to those observed for the PBj6.6 virus, suggesting that the negative charge was essential in this region if not the aspartate itself. However, aspartate is highly conserved at this position for many SIV isolates (data not shown). Similarly, it was shown for HIV-1 that a single and highly conserved aspartate residue between the V1 and V2 region was critical for the early stages of viral replication (47).

Using the SIV animal model allows the pathogenic consequences of the defects in PBj envelope biology identified in vitro to be examined in vivo. PBj6.6- and PBj6.9-infected animals demonstrated a predictable and expected pathogenesis. PBj6.6 was highly pathogenic, resulting in the death of two animals at day 6 and of the remaining animal at day 8 from the typical PBj enteric syndrome. However, those macaques that received PBj6.9 virus had attenuated pathogenesis; all PBj6.9-inoculated animals survived the initial acute phase of the disease with only superficial clinical symptoms, as observed previously (33). However, infection of animals with the env exchange viruses produced a clear change in pathogenesis. Animals that were infected with virus 060 (D119G) displayed a pathogenesis that was very similar to that observed for PBj6.9. These animals survived the initial phase of the infection and failed to demonstrate any major clinical symptoms, even though they were genuinely infected. The only difference between virus 060 and the parental PBj6.6 was the D119G exchange. Therefore, the complete loss of acute pathogenicity by introducing the D119G mutation into PBj6.6 demonstrates the importance of this amino acid and the V1 region of gp120.

In contrast, virus 063 (G119D) infection produced a more severe pathogenesis than that observed for its PBj6.9 parent. Two of the three animals (1031 and 7100) succumbed to virus 063 infection, at days 9 and 11 postinfection. The pathogenesis appeared similar to that seen with PBj6.6, although it was less aggressive and somewhat delayed. The kinetics and pattern of plasma viremia were shifted earlier than those observed in macaques inoculated with the PBj6.9 parent. However, there was more variability in the kinetics of viremia and lower peak levels than were observed in PBj6.6-inoculated macaques. Furthermore, in situ analysis of the pathologic changes observed in representative gut tissue sections associated with viruses 063 and PBj6.6 indicated that although the pathology was similar, there were less extensive cytopathic effects observed for the 063-infected animals. Therefore, virus 063 had a much more pathogenic phenotype than the PBj6.9 parental background as a direct result of introducing the single amino acid change G119D, but chimera 063 was not as pathogenic as the PBj6.6 virus. Animal 93P009, which was also infected with virus 063, survived the infection despite showing an initial steep decline in total lymphocyte subsets and clinical symptoms. Sequencing analysis of the virus recovered from animal 93P009 on two separate occasions demonstrated that the mutation (G119D) of gp120 had been retained. While the introduction of D119G into PBj6.6 clearly reduced the pathogenicity of the 060 chimera, the inability to fully reconstruct PBj6.6 pathogenesis in the PBj6.9 background by making the reciprocal G119D exchange (virus 063) suggests that additional amino acid exchanges are required in Nef or Vpx.

It is now well established that PBj Nef, with the essential 17YE amino acid sequence as part of an immunoreceptor tyrosine-based activation motif, is a major determinant of PBj pathogenesis (8, 9). Mutagenesis of SIVmac239 to SIVmac239YE resulted in a virus with typical PBj properties; however, these infected animals mostly survived the initial acute phase of the disease. Correspondingly, mutagenesis of the Nef YE sequence in the PBj6.6 virus abrogated pathogenesis (37), and disruption of this sequence was found to occur in surviving animals following a mucosal route of infection with SIV PBj14 (39). However, it is likely that the YE-encoding nef mutation alone is insufficient to specify the most severe PBj pathogenesis, because both PBj6.6 and PBj6.9 viruses have an identical Nef YE sequence. The only sequence difference between these two viruses is at position 252 of Nef. The phenylalanine residue is only present in PBj6.6 (data not shown) and is seemingly an unimportant sequence difference, as indicated by the in vitro assays performed here. In addition, replacing the nef 3′ LTR region of the nonpathogenic simian-human immunodeficiency virus strain PPc (SHIV_PPc) with the PBj14 nef LTR region produced virus SHIV_PPc PBjnef, which was able to induce PBMC activation in vitro but failed to replicate productively in vivo. This indicates that PBj nef alone was insufficient to cause a pathogenic infection in vivo, indicating that an interaction between the PBj nef and PBj env was required for pathogenesis (41). Our results are consistent with the involvement of PBj env and nef, i.e., the reduced pathogenicity of virus 060 with the D119G gp120 mutation and the enhanced pathogenicity of virus 063 as a result of the G119D mutation, although both viruses had the same Nef 17YE sequence.

An alternative amino acid exchange that may be involved in PBj disease is that encoded by the vpx gene. Vpx has been shown to be required for replication in macrophages in vitro (27), and more recently, for acute pathogenesis of SIVsmPBj in vivo (20). A vpx deletion mutant PBj virus that was impaired for macrophage infection failed to establish a pathogenic infection in vivo, indicating that a macrophage-dependent mechanism may be a prerequisite for vpx-dependent viral amplification. However, previous studies have demonstrated that SIVsmPBj6.9 is capable of infecting macrophages in vitro. Therefore, if the vpx substitution is important for modulating the acute pathogenesis of SIVsmPBj6.9, it would not appear to be through effects on replication in macrophages. When the vpx exchange alone was introduced into PBj6.6 and PBj6.9 parents, generating viruses 034 and 035, respectively, no alteration in the biologic behavior relative to the parental viruses was observed with in vitro assays. Of course, this may be because these assays were not for vpx function per se and instead were a readout of virus infectivity, of which the gp120 exchanges would be dominant. Therefore, it is possible that along with the G119D mutation introduced into PBj6.9, it may also be necessary to make the Vpx R89C exchange to compensate for any defect in vpx function. Perhaps the reduced total number of SIV-positive cells and the reduced cytopathicity observed by in situ hybridization in addition to the delayed acute phase pathogenesis observed in vivo were due to a lack of viral spread involving the vpx gene.

In conclusion, the identification of important PBj pathogenic determinants must attempt to identify the order of precedence and the combinations required of the various amino acid changes needed to completely reconstruct PBj6.6 pathogenesis in a PBj6.9 background. Our approach has clearly indicated that the function of env is an important component of PBj pathogenesis, where the stability of the gp120-gp41 interaction appears to influence viral infectivity. It has recently been shown by antiviral intervention treatment of PBj-infected animals that viral replication is an important part of the acute phase of the disease (23), which is consistent with our observations. However, this study indicates that it may be the interplay and combination of the env gene, the YE Nef protein, and even the vpx gene that together are necessary and sufficient to specify acute or severe PBj pathogenesis. These results indicate the important consequences that single amino acid changes, which occur frequently during replication of the viral genome by reverse transcriptase, can contribute to virus pathogenesis.