The Vpu Protein of Human Immunodeficiency Virus Type 1 Plays a Protective Role against Virus-Induced Apoptosis in Primary CD4⁺ T Lymphocytes

Abstract

Previous data revealed that primary cultures of peripheral blood mononuclear cells (PBMCs) were killed by apoptosis at higher rates after infection with two CRF01_AE primary isolates of human immunodeficiency virus type 1 (HIV-1) than after infection with five other CRF01_AE primary isolates, five subtype B primary isolates, and two subtype B laboratory strains. Here, we show evidence that mutations at the vpu gene which were exclusively identified only in the two CRF01_AE isolates mentioned above are involved in their abilities to induce massive apoptosis in primary CD4⁺ T lymphocytes. The rates of virus production by these two isolates in the culture media of infected PBMCs were lower (the same as those of the other CRF01_AE isolates) than those of the subtype B isolates. To confirm the correlation between the higher apoptosis-inducing abilities and the mutations at the vpu gene, infectious molecular clone pNL4-3-based vpu mutants were constructed and examined for their apoptosis induction levels. The apoptosis induction levels after introduction of the vpu mutations were greatly increased in primary CD4⁺ T lymphocytes. In contrast, the apoptosis induction abilities of these vpu mutants were lower in human T-cell line MT-4. Thus, the Vpu protein of HIV-1 could play a protective role against virus-induced apoptosis in primary CD4⁺ T lymphocytes.

Human immunodeficiency virus type 1 (HIV-1) infection is characterized by a major decline in circulating CD4⁺ T cells, resulting in susceptibility to opportunistic infections (13). This phenomenon is presumably one of the key factors contributing to the virus-induced impairment of the host immune response. Such a decline in immune cells is generally considered to be caused by apoptosis, although theories about the mechanisms by which HIV-1 causes cell death are controversial (8, 29).

Recent models of high cell turnover kinetics have shown that a low steady-state level of infected cells during clinical latency is no longer incompatible with a continuous inexorable decline in CD4⁺ T cells (19, 41, 63). Two distinct mechanisms for cell killing by HIV-1 have been observed in culture systems: one is the direct killing of infected cells and the other is the induction of apoptosis in uninfected bystander cells, including CD4⁺ and CD8⁺ T cells (4). Apoptosis in bystander cells could be caused by HIV-1 Env gp120, which may induce aberrant T-cell signaling through binding to CD4 molecules on uninfected cells (38). It has also been shown that HIV-1 particle adsorption, even of defective particles, can induce efficient activation-dependent apoptosis in bystander CD4⁺ and CD8⁺ T cells (22-24). In addition, secreted HIV-1-encoded proteins, such as Tat, Nef, or Vpr, and apoptosis-inducing factors released from HIV-1-infected cells, such as Fas ligand, tumor necrosis factor alpha, or tumor necrosis factor-related apoptosis-inducing ligand, have all been shown to trigger apoptosis in uninfected bystander cells (4). On the contrary, several hypotheses derived from in vitro studies have investigated possible mechanisms for the direct killing of infected cells, as follows. Apoptosis induction by recombinantly expressed Env (5, 7, 18, 32, 38, 46, 47), Tat (6, 30, 35, 39, 42, 48, 64), Nef (45, 68), Vpr (11, 56-58), and Vpu (2, 10, 44) has been demonstrated. However, much of the evidence on mediators of HIV-1-induced cell death is contradictory. For example, Tat has been associated with apoptosis induction in some studies (6, 30, 39, 48, 64), while other studies have shown its protective role against apoptosis (15, 35, 68). The wide variability of these findings may be due, at least in part, to the fact that most of these studies have examined the effects of large amounts of recombinant HIV-1 proteins artificially expressed in various kinds of cell lines.

In a previous report, Komoto et al. assayed several HIV-1 primary isolates for their abilities to induce apoptosis in healthy donor-derived peripheral blood mononuclear cells (PBMCs) (27). The apoptosis induction levels were highly variable among individual isolates. Among a total of 12 primary isolates of subtype B and CRF01_AE, only two CRF01_AE isolates induced massive apoptosis, preferentially in CD4⁺ T cells. The increase in p53 protein in infected cells was initiated before virus production, as reported previously for CD4⁺ T cells infected with an HIV-1 laboratory strain (14). The level of p53 protein was almost proportional to the rate of apoptosis induction by an individual isolate. Furthermore, treatment with Z-VAD-FMK, a blocker of apoptosis mediated by caspases (7), significantly decreased cell mortalities, indicating that caspases are involved in the cascade leading to apoptosis.

In this study, we further characterized the two CRF01_AE primary isolates mentioned above to clarify the possible reason(s) for their high rates of apoptosis induction in PBMCs. We detected unique premature stop codon mutations in the vpu gene as a solo common feature in those two isolates. Based on this observation, we examined the possible involvement of such Vpu mutations in apoptosis induction in CD4⁺ T cells. The introduction of similar Vpu mutations into the wild-type laboratory strain, which has a low apoptosis induction ability in primary CD4⁺ T cells, significantly increased its ability to a rate similar to those of the two CRF01_AE primary isolates with high rates of apoptosis induction. In contrast, the same Vpu mutations had the opposite effect in human cell lines, such as MT-4. Thus, this study clearly showed the variable effects of HIV-1 vpu mutations on apoptosis induction in different host cell systems. Furthermore, in primary CD4⁺ T lymphocytes, the Vpu protein could play a protective role against apoptosis.

MATERIALS AND METHODS

Viruses.

HIV-1 laboratory strains NL4-3 (CXCR4 tropic; derived from molecular clone pNL4-3) and Ba-L (CCR5 tropic), both of which are subtype B, were used as controls. A total of 12 HIV-1 primary isolates were used: 5 (0-55-1, 0-4-26, 0-63-1, 17-7-5, and 17-3-6) belonged to subtype B, and 7 (95TNIH022 EMBL accession number AB032740] [3], 95TNIH047 [EMBL accession number AB032741] [3], CU98-28, CU98-26, CU98-29, CU98-31, and 0-47-1) belonged to CRF01_AE (27). HIV-1 vpu mutants derived from infectious molecular clone pNL4-3 (1)—pNL4-3(NL-vpu-TM), pNL4-3(022-vpu-TM), and pNL4-3(Δvpu), which were prepared as described below—were also used.

Site-directed mutagenesis and plasmid construction.

Three vpu mutants were constructed: pNL4-3(NL-vpu-TM), containing the transmembrane (TM) portion of vpu derived from NL4-3; pNL4-3(022-vpu-TM), containing the TM portion of vpu derived from 95TNIH022; and pNL4-3(Δvpu), with a complete deletion of vpu. These vpu mutants were constructed by PCR-based mutagenesis of the 2.7-kb SalI-BamHI fragment, which was inserted into pBluescript II SK(+) (Stratagene, La Jolla, Calif.). To construct pNL4-3(NL-vpu-TM) and pNL4-3(Δvpu), two pairs of primers (Table 1) were used to amplify two fragments that overlap and share an EcoRI site at the location of the desired mutations. The 375- and 292-bp SalI-EcoRI fragments were amplified with primers 1 and 6 and primers 1 and 7 for pNL4-3(NL-vpu-TM) and pNL4-3(Δvpu), respectively. The 279- and 354-bp EcoRI-NdeI fragments were amplified with primers 2 and 8 and primers 3 and 9 for pNL4-3(NL-vpu-TM) and pNL4-3(Δvpu), respectively. Then, individual sets of PCR products were used as templates to amplify the 620-bp SalI-NdeI fragments with primers 1 and 8. To construct pNL4-3(022-vpu-TM), five primers were used. At the first PCR step, 311- and 50-bp fragments were amplified with primers 1 and 9 and primers 4 and 10, respectively. The 311-bp fragment has a SalI site at its 5′ terminus. The 5′ terminus of the 50-bp fragment overlaps the 3′ terminus of the 311-bp fragment. The first PCR products were used for the second amplification step with primers 1 and 10. The amplified 352-bp fragment has a SalI site at its 5′ terminus. At the third PCR step, the 300-bp fragment was amplified with primers 5 and 8. The resultant 300-bp fragment has an NdeI site at its 3′ terminus, and its 5′ terminus overlaps the 3′ terminus of the 352-bp fragment. The second and third PCR products were used as templates to amplify 641-bp SalI-NdeI fragments with primers 1 and 8.

TABLE 1.

PCR primers used for the construction of HIV-1 vpu mutants

Primer	Nucleotide positions in pNL4-3	Nucleotide sequence (restriction enzyme site)a
1	5785-5811	5′-cgGTCGACATAGCAGAATAGGCGTTACTC-3′ (SalI)
2	6112-6142	5′-GCAATAGTTGTGTAATAAGAATTCATCATAG-3′ (EcoRI)
3	6051-6079	5′-AGTACATGTAATGTAATAAGAATTCGTAG-3′ (EcoRI)
4	6073-6097	5′-GCAAATTAGTGCAATAGTAGGACTGATAG-3′
5	6108-6149	5′-CTTAGCAATAGTAGTGTAATAAGAATTCATCATAGAATATAG-3′ (EcoRI)
6	6144-6113	5′-TTCTATGATGAATTCTTATTACACAACTATTG-3′ (EcoRI)
7	6085-6052	5′-CTATTGCTACGAATTCTTATTACATTACATGTAC-3′ (EcoRI)
8	6404-6381	5′-gagaCATATGCTTTAGCATCTGATGCAC-3′ (NdeI)
9	6078-6052	5′-CTAATTTGCAAAGGAGTCATTACATGTAC-3′
10	6118-6089	5′-CTATTGCTAAGATTAGCGCTACTATCAGTC-3′

^aThe restriction enzyme sites are underlined. Lowercase letters indicate nonviral nucleotides added to the 5′ terminus of the primers.

Each SalI-NdeI fragment was digested with restriction enzymes and cloned into the corresponding site of pSB. Finally, SalI-BamHI fragments from individual pSB plasmids containing mutant vpu sequences were reinserted into pNL4-3 at the corresponding sites to create the three pNL4-3-based HIV-1 vpu mutants.

Preparation of healthy donor-derived PBMCs and CD4⁺ T cells.

Healthy donor-derived PBMCs and CD4⁺ T cells were prepared by centrifugation through Ficoll-Paque (Amersham Pharmacia Biotech, Buckinghamshire, United Kingdom). A CD4⁺-T-cell subpopulation was also prepared from PBMCs (20). For this, B cells, monocytes, CD8⁺ T cells, dendritic cells, early erythroid cells, platelets, and basophils were indirectly magnetically labeled by using a cocktail of hapten-conjugated CD8, CD11b, CD16, CD19, CD36, and CD56 antibodies and MACS MicroBeads coupled to an antihapten monoclonal antibody (Miltenyi Biotec, Gladbach, Germany). The magnetically labeled cells were depleted by retention on a MACS separator (Miltenyi Biotec).

Cell cultures, infection, and transfection.

The PBMCs and PBMC-derived CD4⁺ T cells were cultured at 2 × 10⁶/ml in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) (complete medium) in the presence of phytohemagglutinin A (PHA) (2 μg/ml) (Sigma, St. Louis, Mo.) for 3 days.

The PHA-stimulated PBMCs and CD4⁺ T cells were used for the evaluation of apoptosis induction rates after infection with HIV-1 primary isolates, laboratory strains, and pNL4-3-derived vpu mutants. The PHA-stimulated PBMCs were infected with HIV-1 at 110 ng/ml, which was adjusted by using a Gag p24 antigen capture enzyme-linked immunosorbent assay (ELISA) (Zeptometrix, Buffalo, N.Y.). After adsorption for 2 h at 37°C, the cells were washed with phosphate-buffered saline and then cultured at 10⁶/ml in complete medium with recombinant interleukin-2 (50 U/ml) (Becton Dickinson, Mountain View, Calif.). The cells were adjusted to 10⁶/ml in fresh medium every 2 days. In addition, MT-4 cells, a human CD4⁺-T-cell line, were also used for the evaluation of apoptosis induction rates after infection with the same HIV-1 inocula as those described above. The cells were adjusted to 5 × 10⁵/ml and cultured in fresh complete medium.

For transfection with infectious HIV-1 molecular clones, 293T cells, a human kidney cell line, cultured in Dulbecco's modified Eagle's medium containing 10% FBS, were grown to near confluence in 60-mm dishes. The cells were transfected with a total of 10 μg of plasmid DNAs by using a calcium phosphate transfection kit (Invitrogen, Carlsbad, Calif.). The culture supernatants were harvested approximately 48 h after transfection, filtered, and then used as stocks of wild-type or vpu mutant viruses to infect PBMCs, PBMC-derived CD4⁺ T cells, or MT-4 cells. For the assessment of apoptosis induction after transfection with the vpu mutants, 293T cells were similarly transfected with wild-type or individual vpu mutant plasmids and cultured in Dulbecco's modified Eagle's medium containing 10% FBS for 3 days, after which apoptosis induction rates were evaluated. To evaluate transfection efficiency, pEGFP-N1 (BD Biosciences Clontech, Palo Alto, Calif.) was used as a marker with green fluorescence.

Apoptotic cell detection.

The level of apoptosis was monitored by staining with terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) (Medical and Biological Laboratories, Nagoya, Japan), followed by flow cytometry, as described previously (27). In addition, the level of apoptosis was determined by using trypan blue exclusion and staining with the nuclear dye Hoechst 33342 (Sigma) (27).

Viral protein analysis.

The level of HIV-1 production in conditioned media sequentially obtained from infected cells was measured by using the HIV-1 Gag p24 antigen capture ELISA. HIV-1 proteins in infected cells were visualized by Western blotting and an immunofluorescence assay (IFA) with serum from an HIV-1-infected individual. For Western blotting, HIV-1-infected cells (10⁶) were lysed in sodium dodecyl sulfate-polyacrylamide gel electrophoresis sample buffer and subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (27). The separated proteins in the gel were transferred to polyvinylidene difluoride membranes. The membranes were probed with a sheep polyclonal antibody to the Vpu cytoplasmic domain (Aalto Bio Reagent, Dublin, Ireland) or serum from an HIV-1-infected individual. The antigen-antibody complexes were detected by using an enhanced chemiluminescence system (Amersham Pharmacia Biotech). For the IFA, cells from individual cell cultures were smeared, fixed with cold acetone, and then stained with serum from an HIV-1-infected individual, followed by the second antibody, fluorescein isothiocyanate-conjugated goat anti-human immunoglobulin G (Jackson ImmunoResearch, West Grove, Pa.). On average, 300 to 500 cells for each culture were counted to determine the sequential percentage of HIV-1 antigen-positive cells.

Nucleotide sequencing of the HIV-1 vpu gene.

Total cellular DNA was extracted from PHA-stimulated PBMCs after infection with primary HIV-1 isolates as described previously (27). The HIV-1 gene fragment including vpu was amplified from the DNA by nested PCR. The first PCR step was performed with outer primers TAT1 (5′-CAACTGCTGTTTATCCATTTCAGAA-3′) and VPU2 (5′-ATATGCTTTAGCATCTGATGCACAAAA-3′). The first PCR products were used for the second amplification step with inner primers TAT3 (5′-GTGTCGACATAGCAGAATA-3′) and VPU4 (5′-CCATAATAGACTTGTGACCCACA-3′) (61). The PCR products were cloned in vector pCR2.1 by using an Original TA cloning kit (Invitrogen). Seven of the cloned DNAs for each isolate were sequenced on both strands with LI-COR4000 (LI-COR, Lincoln, Nebr.). The nucleotide sequences were aligned and then translated to amino acid sequences with GENETYX-MAC, version 9.0 (Software Development, Tokyo, Japan). The amino acid sequence multiple alignment was performed with CLUSTAL W (60).

RESULTS

Correlation between the vpu mutations in the HIV-1 primary isolates and their higher apoptosis-inducing abilities in PBMCs.

In a previous report, Komoto et al. compared the rates of apoptosis induction in primary PBMCs for 12 primary isolates (5 subtype B and 7 CRF01_AE isolates) and representative laboratory strains (NL4-3 and Ba-L as CXCR4-tropic and CCR5-tropic HIV-1 strains, respectively) (27). PBMCs infected with two CRF01_AE isolates (95TNIH022 and 95TNIH047) showed higher rates of apoptosis than PBMCs infected with the other isolates or the laboratory strains. By the flow cytometric TUNEL technique, the rates of apoptosis induction in PBMCs on day 7 after infection were calculated, as shown on the right side of Fig. 1. The higher rates were exclusively observed in PBMCs infected with isolates 95TNIH022 and 95TNIH047 and not with the other primary isolates or laboratory strains. Similar results were obtained by trypan blue dye exclusion, staining of cells with the nuclear dye Hoechst 33342, and examination of cell morphological changes (data not shown).

FIG. 1.

vpu sequences of five subtype B and seven CRF01_AE primary isolates in addition to the NL4-3 and Ba-L subtype B laboratory strains. The Vpu amino acid sequences of the HIV-1 primary isolates were determined as described in Materials and Methods. The percentages of apoptosis induction in PBMCs on day 7 after infection with these primary isolates are shown at the right (means and standard deviations of triplicate experiments). An asterisk indicates the appearance of a stop codon.

To determine which viral gene(s) of 95TNIH022 and 95TNIH047 contributed to the increased ability to induce apoptosis in PBMCs, we carefully compared the full genome sequences of those two isolates (3) with that of NL4-3, which has intact open reading frames in all genes (1) and is a representative isolate with a lower rate of apoptosis induction in PBMCs. The results revealed (Fig. 1) the existence of a sole unique feature in the 95TNIH022 and 95TNIH047 genomes: premature stop codon mutations at the vpu gene. There were no such mutations at the vpu gene in the other 10 isolates or the 2 laboratory strains. Thus, a positive correlation was suggested between vpu mutations and an increased ability to induce apoptosis in infected PBMCs.

Since the premature stop codon mutations at the vpu gene exist just before the boundary between the TM and cytoplasmic domains in both 95TNIH022 and 95TNIH047, the expected Vpu products of these primary isolates are truncated, with 22 and 27 amino acid residues, respectively (Fig. 2A). Because there was an insertion of one amino acid, commonly in CRF01_AE-specific sequences at the Vpu N termini, the corresponding NL4-3 Vpu products were 21 and 26 amino acid residues long, respectively. In fact, the expression of the full-length Vpu (16-kDa) protein was not detected in PBMCs infected with 95TNIH022 or 95TNIH047 by Western blotting with an anti-Vpu antibody recognizing the cytoplasmic domain, which is the only available commercial antibody (Fig. 2B, left panel). Western blotting of the same infected PBMCs with serum from an HIV-1-seropositive individual showed no differences in the expression of Gag proteins (Fig. 2B, right panel). Thus, truncation at the Vpu cytoplasmic domain in 95TNIH022 and 95TNIH047 was found to be their unique feature. Therefore, we next examined the relationship between such vpu premature stop codon mutations and virus replication and virion release rates as well as apoptosis induction rates.

FIG. 2.

Mutations at vpu gene in 95TNIH022 and 95TNIH047. (A) Schematic representations of vpu sequences of 95TNIH022 (amino acid residues 1 to 22, with one amino acid insertion in the TM domain and no cytoplasmic domain) and 95TNIH047 (amino acid residues 1 to 27, with one amino acid insertion in the TM domain and no cytoplasmic domain). As a control, NL4-3 containing wild-type vpu is shown. LTR, long terminal repeat. (B) Western blotting of Vpu protein in PBMCs infected with 95TNIH022, 95TNIH047, and NL4-3. (Left blot) An anti-Vpu antibody recognized the cytoplasmic domain. (Right blot) A similarly prepared blot was reacted with serum from an HIV-1-seropositive individual.

Replication and rates of release of virions from 95TNIH022 and 95TNIH047 in PBMCs.

To detect the virological characteristics of 95TNIH022 and 95TNIH047, we compared the virus replication kinetics of these two isolates with those of NL4-3 as a control. After adsorption, the infected PBMCs were sequentially harvested and separated into media and cell fractions. The virus particles released into the media were quantified by the Gag p24 antigen capture ELISA (Fig. 3, upper panels), and the HIV-1 proteins within the cells were detected by Western blotting with serum from an HIV-1-seropositive individual (Fig. 3, middle panels). The sequential percentages of infected cells were determined by IFA with the same serum (Fig. 3, lower panels). The two primary isolates showed peaks similar to those of NL4-3 with regard to virion release and the accumulation of viral proteins within infected cells. The peaks for newly synthesized viral proteins, such as the Gag precursor p55 and its cleaved products, p41, p24, and p17, in infected cells were similar at 4 days after infection. In addition, the percentages of HIV-1 antigen-positive cells were also similar in infected cells at the same peaks (Fig. 3, lower panels). However, the amounts of virions released into the media from PBMCs infected with 95TNIH022 and 95TNIH047 were much lower than that released by NL4-3 at 4 days after infection—only 45 and 23% the amount for NL4-3, respectively (Fig. 3, upper panels). Thus, as expected, the reduced rates of release of virions from 95TNIH022- and 95TNIH047-infected cells were confirmed. In contrast, viral proteins were continuously accumulated in PBMCs infected with these two isolates.

FIG. 3.

Comparison of the kinetics of HIV-1 protein synthesis and the release of virions from PBMCs infected with 95TNIH022 or 95TNIH047 and with NL4-3 as a control. PBMCs were infected with 95TNIH022, 95TNIH047, or NL4-3. After adsorption for 2 h, the infected PBMCs were cultured for 12 days. The HIV-1 particles released into the culture media obtained every 2 days were quantified by the Gag p24 antigen capture ELISA (upper panels). The data shown are the means and standard deviations of triplicate experiments. The viral proteins in the infected cells were visualized by Western blotting (middle panels) and IFA (lower panels) with serum from an HIV-1-seropositive individual. The infected cells used for the experiments shown in the middle and lower panels were derived from one of the triplicate experiments shown in the upper panels. IFA data represent the sequential percentages of infected cells.

Construction of HIV-1 molecular clones with various vpu mutations.

To investigate the contributions of the premature stop codon mutations at the vpu gene in 95TNIH022 and 95TNIH047 to their high apoptosis induction rates in infected PBMCs, three kinds of vpu mutants were constructed in the genetic background of pNL4-3 (Fig. 4A). pNL4-3(NL-vpu-TM) had a premature stop codon mutation corresponding to amino acid residue 22 in pNL4-3, the position of which corresponds to the mutation site in 95TNIH022. Since the homology for Vpu between NL4-3 and 95TNIH022 was only 43% at the amino acid level (3), pNL4-3(022-vpu-TM), which was designed to express premature truncated Vpu derived from 95TNIH022 with the other gene products of pNL4-3, was also constructed. In addition, pNL4-3(Δvpu), with a complete deletion of vpu in pNL4-3, was also constructed. Individual viruses produced from 293T cells after transfection with the vpu mutant constructs were used as inocula for PBMC infection. The Vpu and Gag expression of the three pNL4-3-based vpu mutants was confirmed by Western blotting: no detection of Vpu protein in these vpu mutants by an anti-Vpu cytoplasmic domain antibody and no differences in Gag protein expression between these vpu mutants and NL4-3 (Fig. 4B).

FIG. 4.

Construction of NL4-3-based vpu mutants. (A) Three vpu mutants were constructed by introducing three different vpu mutant genes into pNL4-3: the NL4-3-derived gene for the TM region spanning amino acid residues 1 to 21, designated NL4-3(NL-vpu-TM); the 95TNIH022-derived vpu gene, designated NL4-3(022-vpu-TM); and a complete deletion of the vpu gene, designated NL4-3(Δvpu). As a control, wild-type NL4-3 [NL4-3(WT)] was used. LTR, long terminal repeat. (B) Expression of Vpu protein in PBMCs infected with these three vpu mutants and NL4-3(WT), as detected by Western blotting with an anti-Vpu antibody recognizing the cytoplasmic domain and with serum from an HIV-1-seropositive individual as a control antibody.

Augmentation of HIV-1-induced apoptosis in PBMCs and primary CD4⁺ T cells by vpu mutations.

PBMCs were infected with individual vpu mutant viruses and the original NL4-3 wild-type virus as a control. All three vpu mutants induced massive apoptosis in PBMCs, and the levels of apoptosis induction were almost the same as that seen with 95TNIH022 (Fig. 5A, left panel). The mutant with a complete vpu deletion and the mutant expressing truncated Vpu with only the TM domain showed similar increases in the levels of apoptosis induction. When we used a PBMC-derived CD4⁺-T-cell subpopulation, we found that all of the vpu mutants significantly augmented the apoptosis induction rates (Fig. 5A, right panel). A reduction in the amounts of virions released into the media was observed in PBMCs (Fig. 5B, left panel) and CD4⁺ T cells (Fig. 5B, right panel) infected with these vpu mutants and was compatible with the reduction seen with 95TNIH022 infection. In addition, there were no apparent differences in the percentages of HIV-1 antigen-positive cells between wild-type NL4-3 and vpu mutants at the peaks in PBMCs (Fig. 5C, left panel) and CD4⁺ T cells (Fig. 5C, right panel). The same results were obtained with independently prepared PBMCs or CD4⁺ T cells derived from another three healthy donors (data not shown).

FIG. 5.

Augmentation of apoptosis induction by NL4-3 in human primary CD4⁺-T-cell cultures after the introduction of vpu mutations. (A) PBMCs (left panel) or a PBMC-derived CD4⁺-T-cell subpopulation (right panel) was mock infected (□) or infected with NL4-3 (▵), 95TNIH022 (○), NL4-3(NL-vpu-TM) (▾), NL4-3(022-vpu-TM) (⧫), or NL4-3(Δvpu) (▪). The apoptosis induction levels were measured every 2 days by TUNEL staining. (B) Simultaneously, the Gag p24 production levels in the culture media were also measured by the antigen capture ELISA. The data shown in panels A and B are the means and standard deviations of triplicate experiments. (C) Sequential percentages of infected cells, as determined by IFA. The data shown are representative of three independent cell cultures.

Cell type dependence for the augmentation of HIV-1-induced apoptosis by vpu mutations.

We also examined the contribution of the vpu mutations to the augmentation of HIV-1-induced apoptosis in cell line MT-4. Since it was reported that HIV-1 infection reduces MT-4 cell numbers drastically by single cell killing rather than by the induction of syncytia, a mechanism similar to that observed in HIV-1-infected PBMCs (25), we expected an accurate measurement of single cell death by apoptosis. Surprisingly, augmentation of the apoptosis-inducing ability of NL4-3 by the introduction of vpu mutations, as was seen in primary PBMCs and CD4⁺ T cells (Fig. 5), was not observed in MT-4 cells (Fig. 6, left panel). The data obtained for MT-4 cells were completely opposite, revealing decreased levels of apoptosis induction, which were similar in infections with the three vpu mutants. The amounts of virions released into the media from vpu mutant-infected MT-4 cells were smaller than those released by wild-type NL4-3 (Fig. 6, middle panel). In addition, the progression of infection with these vpu mutants was retarded compared to that with wild-type NL4-3 (Fig. 6, middle and right panels).

FIG. 6.

Reduction of apoptosis induction ability of NL4-3 in human CD4⁺-T-cell line MT-4 by the introduction of vpu mutations. MT-4 cells were mock infected (□) or infected with NL4-3(WT) (▵), NL4-3(NL-vpu-TM) (▾), NL4-3(022-vpu-TM) (⧫), or NL4-3(Δvpu) (▪). The apoptosis induction levels were measured every 2 days after infection by TUNEL staining (left panel). Simultaneously, the Gag p24 production levels in the culture media (middle panel) and the sequential percentages of infected cells (right panel) were measured by the antigen capture ELISA and IFA, respectively. The data shown are the means (and standard deviations, for the left and middle panels) of triplicate experiments.

Next, we transfected 293T cells to examine a single round infection (Fig. 7). Vpu is believed not to play a role in the replication cycle before integration. Therefore, we examined the effects of vpu mutations on apoptosis induction in cells transfected with the infectious molecular clone during a single round of replication, probably on events following viral transcription. Apoptosis was found in about 25% of 293T cells transfected with pNL4-3 on day 3 after transfection but in only 16% of the cells transfected with pNL4-3(NL-vpu-TM) (Fig. 7A). The latter level was almost the same as that in controls, such as mock transfection and transfection with pUC18. On the other hand, the rate of virus production in the culture media, as measured by the Gag p24 antigen capture ELISA, was significantly reduced in 293T cells transfected with the vpu mutant, although the virus production kinetics were almost the same for pNL4-3 and pNL4-3(NL-vpu-TM) (Fig. 7B, upper panels). In addition, Western blotting also showed that there were no differences in the amounts of cell-associated viral proteins at all points between pNL4-3 and pNL4-3(NL-vpu-TM) (Fig. 7B, middle panels). To evaluate the transfection efficiency for these HIV-1 plasmids, we used pEGFP-N1, and the result was 68% (data not shown). In addition, the results of IFA with an anti-HIV-1 antibody revealed almost the same kinetics for HIV-1 antigen-positive cells in pNL4-3 and pNL4-3(NL-vpu-TM) (Fig. 7B, lower panels). The percentage of vpu mutant-infected cells was higher than that of wild-type pNL4-3-infected cells, as was seen for PBMCs (Fig. 5).

FIG. 7.

Reduction of apoptosis induction ability of NL4-3 in human kidney cell line 293T by the introduction of vpu mutations. (A) 293T cells were mock transfected or transfected with pNL4-3(WT), pNL4-3(NL-vpu-TM), or pUC18 as a negative control in a single-round infection assay. The apoptosis induction levels were measured 3 days after transfection by TUNEL staining, followed by flow cytometric analysis. The data shown are the means and standard deviations of triplicate experiments. (B) (Upper panels) Simultaneously, the Gag p24 production levels in the culture media from the cells transfected with pNL4-3 (▵) and pNL4-3(NL-vpu-TM) (▾) were measured by the antigen capture ELISA. (Middle and lower panels) The viral proteins in the same transfected cells were visualized by Western blotting and IFA, respectively, with serum from an HIV-1-seropositive individual. The data shown in the upper and lower panels are representative of three independent experiments (means and standard deviations).

Consequently, the contributions of vpu mutations to apoptosis induction were very different between PBMCs or PBMC-derived CD4⁺ T cells and human cell lines, such as MT-4 or 293T, despite the reduction in virus production rates that was consistently observed in cells infected or transfected with the vpu mutants.

DISCUSSION

In this study, we showed that Vpu protein could play a protective role against apoptosis induction in infected human primary CD4⁺ T lymphocytes. Two (95TNIH022 and 95TNIH047) of 12 primary isolates of HIV-1, including subtype B and CRF01_AE isolates, that showed massive apoptosis induction in PBMCs (compared with other results [27]) contained the solo unique feature of mutations at vpu. The contribution of the vpu mutations to the massive apoptosis induction in primary CD4⁺ T cells was demonstrated by infection with newly constructed vpu mutants in the genetic background of pNL4-3, which produces wild-type HIV-1 with a lower apoptosis-inducing ability in primary CD4⁺ T cells.

We examined several pNL4-3-based vpu mutants with only the TM domain (subtype B derived as well as CRF01_AE derived) or with a complete deletion of the vpu gene. However, all of the mutant viruses showed almost similar levels of apoptosis induction in PBMCs. When we infected a PBMC-derived CD4⁺-T-cell-enriched subpopulation with the same vpu mutants, the apoptosis induction levels were greatly augmented. The cell death that was observed with the vpu mutants occurred in uninfected rather than infected cells (Fig. 5). This finding would be consistent with the notion of Vpu increasing cell-cell fusions, which could be the cause of the increased cell death in the culture (17). In contrast, when we infected a human CD4⁺-T-cell line (MT-4) with the same vpu mutants or a transfected human kidney cell line (293T) with pNL4-3-based plasmids with mutant vpu sequences, the apoptosis induction levels were significantly reduced compared to that obtained with wild-type NL4-3. This latter result was consistent with a previous report that a pNL4-3-based mutant virus with a frameshift at vpu could generate more survivors in infected MT-4 cells after acute infection, establishing a persistent infection with HIV-1 (25). Thus, the augmentation of HIV-1-induced apoptosis by vpu mutations was dependent on the type of cells used. It was very clear that Vpu contributed in opposite ways to apoptosis induction in human cell lines and primary CD4⁺ T cells. On the other hand, Rapaport et al. (44) reported that only a slight reduction in the level of apoptosis induction in PBMCs was obtained with NL4-3 containing mutations at the vpr, vpu, and nef genes, leading to the conclusion that these accessory genes are not responsible for apoptosis in infected cells. Therefore, mutations at vpu only may be involved in HIV-1-induced apoptosis in PBMCs in a manner different from that of mutations at vpu together with mutations at vpr and nef.

HIV-1 Vpu is a small integral membrane protein (34, 59) that is found only in HIV-1 and simian immunodeficiency virus (SIV) from chimpanzees (SIV_cpz) (21). Simian-human immunodeficiency viruses (SHIVs) containing regions of the HIV-1 tat, rev, env, and vpu genes in an SIV_mac genetic background represent a model system for analyzing the roles of HIV-1 gene products, such as Vpu, in viral pathogenesis. Data on Vpu functions in monkey systems showed that in strains of nonpathogenic SHIVs that later became pathogenic during passage in monkeys, a functional vpu gene may have been involved, allowing mutations in the nef and env genes to develop, and that after the development of such mutations, vpu became less important in inducing severe depletion of CD4⁺ T cells and consequent fatal opportunistic infections (31, 33, 36, 37). Thus, Vpu seems to be only indirectly involved in CD4⁺-T-cell depletion, at least in monkeys.

Several important biological activities of HIV-1 Vpu have been demonstrated in vitro, including the abilities to facilitate the release of virions from infected cells (16, 26, 51, 59, 67), induce CD4 degradation in the endoplasmic reticulum (9, 43, 65, 66), and regulate transport of the viral Env protein from the endoplasmic reticulum to the Golgi apparatus (62). These functions of Vpu are mechanistically distinct (40, 49, 50, 55). The degradation of CD4 involves the cytoplasmic domain of Vpu, and this Vpu function is dependent on the phosphorylation of two conserved serine residues (49, 53, 54); in contrast, the involvement of Vpu in virion release depends on the TM domain (50). An ion channel activity of the Vpu TM domain was also identified (12, 28) and shown to be involved in the regulation of virus release from infected cells (52). In fact, the levels of HIV-1 virions in the culture media from 95TNIH022- and 95TNIH047-infected cells were much lower than those in media from NL4-3-infected cells. Consequently, viral proteins gradually accumulated inside the cells. Thus, the involvement of Vpu in virion release was confirmed in our system with PBMCs, but we did not examine the involvement of Vpu in CD4 degradation because it is difficult to differentiate among Env, Nef, and Vpu functions in CD4 down-regulation or degradation in the culture system of PBMCs infected with HIV-1. The vpu mutants containing a complete deletion and only the TM domain similarly augmented apoptosis in PBMCs and CD4⁺ T cells, indicating that the cytoplasmic domain may contribute to this augmentation. However, we could not detect the expression of the TM domain in NL4-3(NL-vpu-TM)- and NL4-3(022-vpu-TM)-infected cells, since an antibody to the Vpu TM domain was not available. Thus, the Vpu domain contributing to the augmentation of apoptosis in primary CD4⁺ T cells could not be elucidated in this study. However, a large amount of retained HIV-1 virions in infected PBMCs could become a trigger for enhanced apoptosis of the infected cells.

Although several Vpu functions have been elucidated in HIV-1 pathogenesis, whether HIV-1 lacking an intact vpu gene can still cause disease or increase virulence in humans is currently unknown. In this study, we identified vpu mutations only in some CRF01_AE isolates (two of seven; 29%) and not in subtype B isolates (zero of five) that were correlated with massive apoptosis induction. Therefore, it might be helpful to characterize the frequency of vpu mutant generation in CRF01_AE-infected individuals to gain an understanding of the possible differential mechanisms of AIDS pathogenesis among different HIV-1 subtypes and to determine whether vpu mutations contribute to apoptosis in vivo.