HIV-1 Protease Has a Genetic T-Cell Adjuvant Effect Which Is Negatively Regulated by Proteolytic Activity

Abstract

HIV protease (PR) mediates the processing of human immunodeficiency virus (HIV) polyproteins and is necessary for the viral production. Recently, HIV PR was shown to possess both cytotoxic and chaperonelike activity. We demonstrate here that HIV PR can serve as a genetic adjuvant that enhances the HIV Env and human papillomavirus (HPV) DNA vaccine-induced T-cell response in a dose-dependent manner, only when codelivered with DNA vaccine. Interestingly, the T-cell adjuvant effects of HIV PR were increased by introducing several mutations that inhibited its proteolytic activity, indicating that the adjuvant properties were inversely correlated with its proteolytic activity. Conversely, the introduction of a mutation in the flap region of HIV PR limiting the access to the core domain of HIV PR inhibited the T-cell adjuvant effect, suggesting that the HIV PR chaperonelike activity may play a role in mediating T-cell adjuvant properties. A similar adjuvant effect was also observed in adenovirus vaccine, indicating vaccine type independency. These findings suggest that HIV PR can modulate T-cell responses elicited by a gene-based vaccine positively by inherent chaperonelike activity and negatively by its proteolytic activity.

Human immunodeficiency virus protease (HIV PR) is a typical aspartic protease required for processing HIV polyproteins such as Gag and Pol and is essential for HIV production (5). HIV PR consists of three domains: terminal, core, and flap domains, and each has a unique role in the proteolytic process (21). The substrate-binding site is found in the core domain that contains the catalytic triad (DTG). The terminal domain is required for dimerization, which is also important for its proteolytic activity, and the flap domain regulates substrate access (2).

In addition to the above-described processes, HIV PR also cleaves cellular proteins important to cell survival including the antiapoptotic protein, Bcl-2 (26), and procaspase-8 (20), which mediates the apoptosis of HIV-infected cells or of cells transfected with DNA encoding HIV PR. The importance of HIV PR proteolytic activity in mediating cell death was highlighted in studies that demonstrated the inhibition of cell death after mutations to the DTG catalytic site (26) or after the treatment with protease inhibitors such as ritonavir or saquinavir (19).

HIV PR was also shown to have inherent chaperonelike activity mediated by the core domain. The introduction of a mutation that removed the aspartic acid residue of the catalytic site or inhibition of dimerization necessary for proteolytic activity did not affect its chaperonelike activity in vitro, suggesting that the chaperonelike activity was independent of its proteolytic properties (8).

Typical chaperones, such as heat shock protein 70 (Hsp70), Hsp90, and gp96, were reported to chaperone antigenic peptides and mediate cross-priming of cognate antigen-specific CD8 T cells in vivo (1). In addition, the minimal 136-amino-acid peptide binding domain of the mycobacterial Hsp70 efficiently generated CD8 T-cell responses against the complexed peptide. Hsps can also protect processed peptides from further proteosomal degradation and enhance targeting to dendritic cells via the interactions of Hsps with surface molecules, including CD91, TLR4, or CCR5, resulting in CD8 T-cell cross-priming (10).

In the present study, we demonstrated that the codelivery of HIV PR could enhance the T-cell response, but not the humoral response, elicited by DNA and adenoviral vaccines and that the T-cell response was further augmented by HIV PR mutations that inhibited proteolytic activity. Interestingly, the T-cell adjuvant effect of the catalytic mutant was reduced by the introduction of a point mutation that stabilized the flap domain into a closed position and not by a mutation that inhibited dimerization. These data suggested that HIV PR has a T-cell adjuvant effect presumably due to the intrinsic chaperonelike activity which is veiled by its proteolytic activity.

MATERIALS AND METHODS

Construction of plasmids encoding HIV-1 envelope, human papillomavirus (HPV) E6-E7 fusion protein (HPV E67), HIV PR, and its derivatives.

Based on the HIV-1 subtype B consensus sequence published by the Los Alamos National Lab HIV Sequence Database in 2004, a plasmid encoding HIV envelope gene (HIV-1 Env) was synthesized by reverse translation using codons preferred in human cells (Genscript). The codon-optimized env signal sequence was replaced by a signal sequence corresponding to the tissue plasminogen activator, and the env transmembrane region (encoding for amino acids 680 to 702) was deleted to enhance the secretion of the gp120/gp41 complex. The plasmid encoding HPV type 16 E6-E7 fusion protein (HPV E67) was constructed as previously described (23).

The wild-type HIV-1 PR gene (PR) was also codon optimized and synthesized by reverse translation. The catalytically attenuated PR (PR_T26S) and inactivated PR (PR_D25A) were constructed using the site-directed mutagenesis that replaced Thr-26 with Ser or Asp-25 with Ala, respectively. PR_M46I and PR_D25A/M46I were constructed in a similar manner. To generate HIV PR_1-95, the HIV PR 1-95 region was PCR amplified and subcloned into pGX10. Mutations were verified by sequencing.

To test the in vitro expression of plasmids encoding either HIV-1 Env, HPV E67, or HIV PR and its derivatives, 3 μg of each plasmid was transfected into COS-7 cells using Lipofectamine (Invitrogen). At 24 h posttransfection, cell lysates were immunoblotted with anti-HIV gp120 (National Institute for Biological Standards and Control [NIBSC]), anti-HPV E7 monoclonal antibody (Zymed), or anti-HIV PR polyclonal antibody (NIBSC), respectively.

Generation of recombinant adenoviral vector.

The recombinant adenoviral vectors encoding HPV16 E7 protein, HIV PR, or HIV PR_D25A were generated by using an AdEasy vector system (Qbiogene). Briefly, each gene was subcloned into a pShuttleCMV vector that was then cotransformed with pAdEasy into an Escherichia coli strain, BJ5183, by electroporation to generate either pAdEasy-E7, -HIV PR, or -PR_D25A, respectively. The human embryo kidney 293 (HEK293) packaging cell line was transfected with the linearized recombinant pAdEasy construct. The recombinant adenoviruses were amplified in HEK-293 cells, purified using double cesium ultracentrifugation, and titrated using 50% tissue culture infectious doses as previously described (18). The expression of E7 protein, HIV PR, and HIV PRD25A was determined by immunoblotting.

HIV Gag cleavage by HIV PR or its derivatives.

To evaluate the catalytic activity of wild-type HIV PR or its mutants, 2 μg of pGX10-HIV Gag was cotransfected with equal amounts of either pGX10 or pGX10 encoding HIV PR or its derivatives into HEK293 cells. Cell lysates from transfected cells were immunoblotted 24 h later and probed with anti-HIV p24 monoclonal antibody kindly provided by NIBSC. The relative ratio of Gag precursor and MA-CA intermediate were determined to estimate the catalytic activity.

Mice and immunization.

C57BL/6 and BALB/c female mice 6 to 8 weeks old were purchased from Jackson Laboratory (Japan) and housed under specific-pathogen-free conditions in an approved animal facility at Postech Biotech Center. Mice were immunized with 100 μl of plasmid DNA or recombinant adenovirus bilaterally in the anterior tibialis muscles (50 μl per site).

IFN-γ ELISPOT assay.

Splenocytes from the immunized mice were harvested 3 weeks after the last immunization and IFN-γ producing cells were assessed by gamma interferon (IFN-γ) enzyme-linked immunospot (ELISPOT) assay. Briefly, total splenocytes or splenocytes (5 × 10⁵ to 10 × 10⁵ cell/well) depleted of either CD4⁺ or CD8⁺ cells by MACS using magnetic beads (Miltenyi Biotech) in cell culture medium (RPMI 1640 supplemented with 10% fetal bovine serum, 2 mM l-glutamine, 100 U of penicillin/ml, and 100 μg of streptomycin/ml) were added to Multiscreen 96-well plates (Millipore) coated with anti-mouse IFN-γ antibody (Pharmingen). For HIV-1 Env immunogen, the 15-mer peptides (with a 10-amino-acid overlap), which spanned the entire coding region of HIV-1 B subtype consensus envelope (kindly provided by the National Institutes of Health [NIH] AIDS reagent reference repository), were mixed into three pools and used as stimulants for an IFN-γ ELISPOT assay. For HPV16 E6 or E7 immunogen, the CD8 epitopes YDFAFRDL or RAHYNIVTF corresponding to HPV16 E6 and E7 sequences, respectively, were used as a stimulant. After 24 h of incubation at 37°C in a humidified atmosphere containing 5% CO₂, the plates were processed as previously described (15).

ELISA.

To determine the level of antigen-specific IgG responses, sera were harvested 3 weeks after the final immunization and analyzed by enzyme-linked immunosorbent assay (ELISA). Sera (1/50 dilution in blocking buffer containing 5% nonfat milk) from mice were added to 96-well plates (Nunc) coated with 0.5 μg of HIV gp120 (LAI IIIB; NIBSC)/ml or 1 μg of HPV16 E7 recombinant protein/ml. After 4 h at room temperature, the plates were processed as previously described (15).

Flow cytometric analysis.

For the flow cytometric analysis, allophycocyanin (APC)-conjugated anti-CD8, phycoerythrin (PE)-conjugated anti-IFN-γ and fluorescein isothiocyanate (FITC)-conjugated anti-CD107a were obtained from BD Biosciences. The cell surface or intracellular cytokine staining was performed as previously described (18). To detect degranulation by CD107a staining, splenocytes (1 × 10⁶cell/well) were stimulated with either E6 or E7 CD8 epitopes (2 μg/ml) in the presence of FITC-conjugated anti-CD107a (1 μg/well) and brefeldin A (5 μg/ml; Sigma-Aldrich) for 6 h. Stained cells were analyzed using a FACSCalibur flow cytometer and CellQuest Pro software (BD Biosciences).

TC-1 tumor protection experiment.

For the TC-1 tumor protection experiment, immunized mice (n = 9) were challenged with 5 × 10⁵ TC-1 subcutaneously into the right hind flank. The tumor volume and survival rate were monitored twice a week for the indicated period and calculated as previously described (11). Mice bearing tumors that exceeded 1,500 mm³ in volume were sacrificed for ethical reasons according to institutional guidelines.

Statistical analyses.

All values are expressed as the mean ± the standard error of the mean (SEM). A two-tailed Student t test was used to determine significant differences between the experimental groups. The statistical differences in survival between experimental groups were determined by using the Kaplan-Meier log-rank test.

RESULTS

Codelivery of HIV PR could enhance the T-cell response, but not the humoral response, elicited by HIV-1 Env DNA vaccination, which was further increased by its catalytic mutant.

To investigate the adjuvant effect of HIV PR on DNA vaccine-induced immune response, we constructed plasmids encoding a catalytically active HIV PR (PR) and attenuated (PR_T26S) and inactive (PR_D25A) mutants. It was known that mutation of the catalytic triad sequence from DTG to DSG decreased the proteolytic activity ca. 5 to 10 times and that the complete loss of its activity could be achieved by a point mutation changing Asp-25 of HIV-1 PR to Ala-25 (14).

The proteolytic activities of the respective constructs were determined regarding the processing of HIV-Gag in HEK 293 cells cotransfected with plasmids encoding HIV-1 gag protein and an equal amount of the HIV PR constructs (Fig. 1). Cell lysates were examined 24 h after transfection by immunoblotting using anti-p24 (MA) monoclonal antibody. Consistent with the previous reports (12, 14), HIV-Gag was not efficiently processed in the absence of catalytically active HIV PR (vector-only transfected cells). The cotransfection of both PR and PR_T26S induced specific cleavage of HIV-1 Gag to a MA-CA intermediate (p40). However, the generation of this intermediate was decreased when the proteolytic activity was attenuated via T26S mutation. The processing of Gag protein was significantly suppressed in cells cotransfected with PR_D25A showing a cleavage pattern similar to that of the vector control.

FIG. 1.

Processing of HIV Gag polyprotein by HIV protease or its derivatives. (A) Plasmid encoding HIV-1 Gag (pGX10-gag) was cotransfected into HEK- 293 cells with either pGX10 (vector control), HIV PR, or derivatives of HIV PR (PR_1-95, PR_M46I, PR_T26S, or PR_D25A). Cell lysates were examined 24 h later by immunoblotting with anti-p24 monoclonal antibody. Gag (p55) represents the precursor, and MA-CA (p40) corresponds to the proteolytic cleavage product. (B) The relative ratio of band intensity between Gag and MA-CA for each treatment group was determined to estimate proteolytic activity.

When BALB/c mice were intramuscularly immunized twice at 3- week intervals with plasmids encoding HIV-1 envelope (Env) in the absence or presence of either HIV PR, PR_T26S, or PR_D25A, the HIV-1 Env DNA vaccine could induce an HIV gp120-specific IgG response. The codelivery of HIV PR and its catalytic mutants could not enhance the HIV gp120-specific humoral response. However, the codelivery of HIV PR decreased the generation of HIV gp120-specific humoral response induced by DNA vaccination (Fig. 2A). It is likely that the proteolytic activity of HIV PR is inversely associated with gp120-specific IgG response induced by DNA vaccination.

FIG. 2.

Effect of HIV PR and its catalytic mutants on humoral and T-cell immune responses to HIV Env. BALB/c mice were immunized twice at 3-week intervals with a plasmid encoding HIV Env (20 μg) plus an equal amount of either pGX10, PR, PR_T26S, or PR_D25A. (A) At 3 weeks after the final immunization, serum samples were collected and measured for IgG responses against HIV gp120 (n = 10 for naive mice, n = 15 for immunized mice). The absorbance of individual mice or the mean absorbance for each group is indicated by an open circle or bar, respectively. (B) Splenocytes from immunized mice (n = 6) were prepared and examined by IFN-γ ELISPOT assay against three different peptide pools of HIV Env (SFC, spot-forming cells). The data are representative of three independent experiments. *, P < 0.05 (Student t test). (C) Splenocytes from immunized mice were pooled, and either CD4⁺ or CD8⁺ cells were depleted by magnetic separation. HIV Env-specific T-cell responses were examined by IFN-γ ELISPOT assay.

In contrast to the effect on the humoral response, the codelivery of HIV PR enhanced HIV Env-induced T-cell responses as monitored by an IFN-γ ELISPOT assay using three different HIV Env peptide pools. More importantly, the codelivery of HIV PR_T26S and PR_D25A significantly increased the level of HIV Env-specific IFN-γ-secreting T cells. An approximately sixfold higher level of cumulative HIV Env-specific IFN-γ-secreting T cells could be induced by the codelivery of both PR_T26S and PR_D25A (Fig. 2B). To further evaluate whether both CD4 and CD8 T-cell responses could be adjuvanted by the HIV PR catalytic mutant, either CD4⁺ or CD8⁺ cell-depleted splenocytes were used for a IFN-γ ELISPOT assay. The codelivery of PR_T26S could enhance HIV Env-specific CD4 and CD8 T-cell responses (Fig. 3C).

FIG. 3.

Effect of HIV PR and its catalytic mutants on HPV E7-specific CD8 T-cell responses. (A) C57BL/6 mice were immunized twice at 3-week intervals with 50 μg of HPV E67 DNA plus an equal amount of either pGX10, PR, PR_T26S, or PR_D25A. E67/PR_D25A represents mice vaccinated with E67 and PR_D25A at opposite sites of the anterior tibialis. At 3 weeks after the final immunization, splenocytes from the immunized mice (n = 6) were prepared and examined by IFN-γ ELISPOT after stimulation with the HPV E7 CD8 epitope (SFC, spot-forming cells). (B) IFN-γ secretion and degranulation of E7-specific CD8 T cells were assessed using flow cytometry by detecting intracellular IFN-γ-phycoerythrin- and CD107a-FITC-stained cells. (C) Dose-dependent effects of PR_D25A from 0 to 100 μg were evaluated by flow cytometry. All of the data are means ± the SEM and are representative of three independent experiments. *, P < 0.05 (versus the group immunized with E67+pGX10); ‡, P < 0.05 (versus the group immunized with E67+PR_D25A). (D) C57BL/6 mice were immunized with the HPV adenovirus vaccine encoding HPV E7 (10⁶ PFU/mouse) plus an equal amount of either mock, rAd-HIV PR, or rAd-PR_D25A. At 1 week postimmunization, splenocytes from the immunized mice (n = 6 per group) were harvested, and the number of HPV E7- specific CD107a⁺/IFN-γ⁺ CD8 T cells was determined by flow cytometry. *, P < 0.05 (versus the group immunized with Ad-E7+Ad-mock).

These results showed that HIV PR catalytic mutants could enhance HIV Env-specific T-cell response and that the adjuvant effect is restricted to T-cell immune response.

T-cell adjuvant effect of HIV PR catalytic mutant depends on colocalization with antigen-DNA in a vaccine-type-independent and dose-dependent manner.

To exclude the possibility that the T-cell adjuvant effect of HIV PR catalytic mutants is dependent upon antigenic nature and mice strains, C57BL/6 mice were intramuscularly immunized twice with plasmid encoding HPV E67 in the absence or presence of either HIV PR, PR_T26S, or PR_D25A as described above.

As demonstrated for the HIV Env vaccine, HPV E7-specific CD8 T-cell responses were significantly enhanced by the codelivery of two HIV PR catalytic mutants (Fig. 3A), although an HPV E6-specific CD8 T-cell response was undetectable in all groups, since this epitope is poorly immunogenic as previously reported (23).

Interestingly, the spatial separation of two plasmids (HPV E67 and PR_D25A) by the injection of each plasmid into the opposite sites of anterior tibialis did not show any T-cell adjuvant effect, indicating that the coexpression of antigen-DNA with HIV PR_D25A is critical for the T-cell adjuvant effect (Fig. 3A and B).

Similar results were obtained when CD107a (LAMP-1), a marker of CD8 T-cell degranulation, was monitored by flow cytometry, suggesting that both cytotoxicity and IFN-γ secretion were also enhanced by the codelivery of HIV PR catalytic mutants (Fig. 3B). In addition, the codelivery of HIV PR_D25A could increase HPV E7-specific CD107a/IFN-γ-positive CD8 T cells in a dose-dependent manner. Equal amounts of HIV PR_D25A could induce the optimal CD8 T-cell response by DNA vaccination (Fig. 3C).

The efficacy of HIV PR_D25A as a genetic adjuvant was further assessed using recombinant adenoviral vaccines expressing HPV E7. HIV PR_D25A also enhanced the CD8 T-cell response induced by a recombinant adenoviral vaccine expressing HPV E7, suggesting that the T-cell adjuvant effect of the HIV PR catalytic mutant is vaccine type independent (Fig. 3D).

The enhanced HPV E7-specific CD8 T-cell response is correlated with tumor protection efficacy against TC-1 tumor challenge.

To evaluate the T-cell adjuvant effect of HIV PR or its catalytic mutants in the context of tumor protection, immunized mice, as indicated in the legend to Fig. 4, were subcutaneously challenged with 5 × 10⁵ TC-1 tumor cells expressing HPV E6 and E7 antigen 3 weeks after the final immunization.

FIG. 4.

Protective efficacy of HIV PR and its catalytic mutants against TC-1 tumor challenge. C57BL/6 female mice were immunized twice at a 3-week interval with plasmid encoding 50 μg of HPV E67 DNA plus an equal amount of either pGX10, PR, PR_T26S, or PR_D25A. At 3 weeks after the final immunization, mice were subcutaneously challenged with 5 × 10⁵ of TC-1 tumor. Tumor growth (A) and survival (B) were monitored twice a week. *, P < 0.05 (versus the group treated with pGX10); ‡, P < 0.05 (versus the group immunized with E67+PR).

As expected, tumor growth and mortality in mice vaccinated with the HPV E67 DNA vaccine were reduced compared to mock DNA-immunized controls, but mice in both groups succumbed to their tumor burdens by day 45 postchallenge. Immunization with the HPV E67 DNA vaccine combined with HIV PR did not affect tumor growth significantly, and the survival rate was only increased to 22%. In contrast, tumor growth was significantly retarded in the groups adjuvanted with HIV PR_T26S or PR_D25A, leading to enhanced survival rates (66 to 77% survival, respectively) (Fig. 4).

Access of the antigenic peptide to the core domain of HIV PR, but not dimerization, is important for the T-cell adjuvant effect of the HIV PR catalytic mutant.

To clarify the mechanism for the T-cell adjuvant effect of PR_D25A, we evaluated the effect of C-terminal deletion and M46I mutation introduced into either HIV PR or PR_D25A on the T-cell adjuvant effect. Deletion of four amino acids from the C-terminal ends inhibits the dimerization of HIV PR and eliminates proteolytic activity but not the chaperonelike activity of HIV PR (9). The M46I mutation suppressed the access of substrate to the catalytic sites required for both proteolytic and chaperonelike activities of HIV PR by stabilizing flap domains into a closed position.

As shown in Fig. 1, HIV PR_1-95 did not show any proteolytic activity in vitro, and the codelivery of PR_1-95 could significantly enhance the HPV E7-specific CD8 T-cell response as efficiently as PR_D25A, also suggesting that the HIV PR T-cell adjuvant effects were inversely proportional to proteolytic activity (Fig. 5).

FIG. 5.

Effect of the C-terminal deletion or the M46I mutation on HIV PR T-cell adjuvant effects. C57BL/6 mice (n = 6) were immunized twice with the HPV DNA vaccine adjuvanted with either HIV PR or the respective derivatives. Spot-forming cells (SFCs) were determined as described above. The data are means ± the SEM and are representative of three independent experiments. *, P < 0.05 (versus the group immunized with E67+ pGX10); ‡, P < 0.05 (versus the group immunized with E67+PR_D25A).

Interestingly, although the proteolytic activity of PR_M46I was less compared to that of HIV PR, as shown in Fig. 1, introduction of the M46I mutation into either HIV PR or PR_D25A (PR_M46I or PR_D25A/M46I, respectively) suppressed their T-cell adjuvant effects, suggesting that the access of antigenic peptide to the catalytic site required for the HIV PR chaperonelike activity played an important role in mediating these effects.

These results are somehow consistent with a previous report that demonstrated that the monomeric form of HIV PR also had chaperonelike activity and that protease inhibitor treatment suppressed chaperone function by competitively inhibiting substrate binding (8).

DISCUSSION

In the present study, we demonstrated for the first time that the catalytic mutants of HIV PR could act as effective genetic adjuvants with the capacity of significantly enhancing T-cell immune response, but not humoral response, in both DNA and adenovirus vaccination as a consequence of inherent HIV PR chaperonelike activity.

The T-cell adjuvant effect of HIV PR was negatively regulated by its proteolytic activity. Both HIV PR_T26S and PR_D25A had a more potent T-cell adjuvant effect and enhanced HIV Env or HPV E7 DNA vaccine-induced T-cell response more efficiently than did HIV PR, also improving protection against TC-1 tumor challenge.

It was previously reported that HIV PR decreased the expression of the cotransfected reporter gene due to HIV PR-mediated cell death and that the treatment of an HIV protease inhibitor prevented its suppressive effect on gene expression in vitro due to HIV PR-mediated cell death (26). This suggested that the codelivery of HIV PR decreased antigen expression and that either attenuation or elimination of proteolytic activity as a consequence of a catalytic mutation could suppress the deleterious effect of HIV PR on antigen expression. It is likely that HIV PR, but not catalytic mutants, reduced the antibody response due to a decrease in antigen production by the codelivery of PR.

In contrast, HIV PR could slightly enhance T-cell responses against codelivered antigen. Although the codelivery of HIV PR decreases antigen expression, PR-mediated cell death can induce the cross-priming of cell-associated antigen to cognate CD8 T cells, which was known to be more efficient than soluble antigen for cross-priming, thus enhancing the overall CD8 T-cell response (16).

The balance between cell death and antigen expression is important for apoptosis-mediated enhancement of DNA vaccine efficacy, as previously reported (13, 22). The catalytic mutant of caspase-2, in which the cysteine residue in the catalytic site is replaced, induced cell death at low levels and enhanced DNA vaccine induced immune responses more efficiently than did the wild-type caspase-2 (22). Consistently, the codelivery of the caspase gene in the control of the internal ribosome entry site (IRES) could induce a T-cell response against HIV Env better than that of cytomegalovirus-controlled caspase via the reduction of caspase gene expression, allowing sufficient antigen expression (13). A similar mechanism might be involved in the enhancement of a DNA vaccine-induced T-cell response via the codelivery of PR_T26S since the proteolytic activity of PR_T26S, associated with cytotoxicity, is relatively lower than that of HIV PR. However, it is likely that other factors besides of apoptosis-mediated enhancement of DNA vaccine are involved in the T-cell adjuvant effect of HIV PR_D25A because the catalytically inactive HIV PR_D25A did not induce cell death in vitro (26).

Due to their high level of structural and functional similarity, it has been suggested that HIV PR and molecular chaperones evolved from the same ancestral chaperone (8). Proteases such as Clp AP and Clp XP also possess chaperone properties, and the initial interaction of non-natively folded proteins is overlapped for their respective functions (7). In addition, a heat shock protein such as DegP has a protease domain that is also critical for chaperone activity (24).

As previously mentioned, the chaperonelike properties of HIV PR are independent of the proteolytic activity since the elimination of proteolytic functions, by either inhibiting homodimerization via the deletion of C-terminal ends or introduction of catalytically inactivating mutations, did not affect chaperonelike activity (8). Consistent with previous reports describing the inherent chaperonelike activity of HIV PR, we found that the dimerization of PR_D25A is not required for the T-cell adjuvant effect in vivo (data not shown). Codelivery of the PR_1-95 single mutant without the proteolytic function could also enhance the E7-specific CD8 T-cell response as efficiently as PR_D25A.

Furthermore, the introduction of M46I mutation locking the HIV PR flaps into a closed position reduced the T-cell adjuvant effects of both HIV PR and PR_D25A. Although HIV PR_M46I exhibits a slightly decreased level of proteolytic activity compared to HIV PR, the introduction of M46I mutation failed to enhance the T-cell adjuvant effects of either HIV PR or PR_D25A, suggesting that chaperoning of antigenic peptides and the associated cross-priming was more important to the T-cell adjuvant effect. These results further support our suggestion that chaperonelike properties of HIV PR, but not the cytotoxicity that was associated with the proteolytic activity, played a critical role in mediating the T-cell adjuvant effect of HIV PR.

Although the proteolytic function of HIV PR is critical for HIV survival, it is questionable why HIV encodes a protein that enhances T-cell response against itself. Paradoxically, immune activation is necessary for HIV replication and accompanied by progressive HIV infection (4). It was previously shown that HIV preferentially infects HIV-specific CD4 T cells in vivo and subsequently leads to the loss of HIV-specific CD4 T cells (3). Thus, the T-cell adjuvant effect of HIV PR might contribute to the persistent infection via expanding cellular targets for HIV, although this requires further evaluation.

It is noticeable that pathogenic Hsps and those found in mammals are highly homologous and that mycobacterial Hsp65-specific T cells cross-react with mammalian Hsp65, inducing autoimmune diseases such as insulin-dependent diabetes mellitus in NOD mice following either immunization with incomplete Freund adjuvant or adoptive transfer of Hsp65-specific T-cell clones (6). These observations suggested that HIV PR_D25A with chaperonelike activity could be used as an alternative adjuvant for genetic vaccination without safety concerns such as autoimmunity induced by the cross-reactive Hsp-specific T cells (25) or anti-Hsp antibodies (17).

Overall, the data presented in this report demonstrate that HIV PR catalytic mutants are potent T-cell adjuvants in gene-based vaccinations and that their inherent chaperonelike activity, but not cytotoxicity associated with proteolytic activity, plays an important role in mediating T-cell adjuvant effects.