Lysis of Human Immunodeficiency Virus Type 1 by a Specific Secreted Human Phospholipase A₂

Abstract

Phospholipase A₂ (PLA₂) proteins affect cellular activation, signal transduction, and possibly innate immunity. A specific secretory PLA₂, sPLA₂-X, is shown here to neutralize human immunodeficiency virus type 1 (HIV-1) through degradation of the viral membrane. Catalytic function was required for antiviral activity, and the target cells of infection were unaffected. sPLA₂-X potently reduced gene transfer of HIV-1 Env-pseudotyped lentivirus vectors and inhibited the replication of both CCR5- and CXCR4-tropic HIV-1 in human CD4⁺ T cells. Virions resistant to damage by antibody and complement were sensitive to lysis by sPLA₂-X, suggesting a novel mechanism of antiviral surveillance independent of the acquired immune system.

Mammalian viruses are subjected to a variety of antiviral host defenses, including recognition by antibody and complement, lysis of producer cells by cytolytic T cells, interferon responses, and possibly antibody-dependent cellular toxicity. However, the mechanisms by which viruses may be subject to recognition and removal at sites of entry, particularly in the mucosa, are poorly defined. Among the gene products synthesized in the gut, phospholipases (PLAs) are secreted in high abundance and play a role in digestion, cellular activation, and nerve signaling. PLA₂s hydrolyze phospholipids at the sn-2 position to release fatty acids and lysophospholipids (3, 18, 24). They are classified into two major groups, the high-molecular-weight cytosolic PLA₂s and the low-molecular-weight secreted PLA₂s (sPLA₂s), which have diverse functions. Human sPLA₂ groups (IB, IIA, IID, IIE, IIF, III, V, VII [platelet-activating factor acetyl hydrolase], X, XIIA, and XIIB) show unique tissue and cellular localizations and enzymatic properties (20, 23). Of these sPLA₂s, group X sPLA₂ (sPLA₂-X) exhibits the highest capacity to hydrolyze phospholipids in intact mammalian cell membranes when overexpressed (16, 17) or added exogenously (4, 8). sPLA₂-X is converted from a catalytically inactive zymogen to a mature, catalytically active form by removal of the N-terminal propeptide after the secretion process, supporting the extracellular action of this enzyme (14). Some sPLA₂s have bactericidal activity, presumably through their effect on membrane integrity, raising the possibility that they may contribute to host antimicrobial defense (7, 11, 12, 26). On the other hand, the role of sPLA₂s in viral infections has not been defined.

In this study, we screened several human sPLA₂s for their potential antiviral effects, and we report that human sPLA₂-X has antiviral activity against lentiviruses due to its catalytic function and its recognition of the virus envelope. This effect was observed even when viruses were resistant to antibody-mediated complement activation. The ability of sPLA₂-X to degrade viruses suggests a novel mechanism of host defense that may provide a barrier to infection independent of the adaptive immune response.

MATERIALS AND METHODS

Cell lines.

The 786-O (human kidney adenocarcinoma) cell line was purchased from the American Type Culture Collection. The HeLa-derived cell line MAGI-CCR5 (a subline of HeLa expressing CCR5) was obtained from the NIH AIDS Research and Reference Reagent Program. The human T-cell leukemia cell line A3R5 (a subline of A3.01 expressing both CCR5 and CXCR4) was a gift from Jerome Kim of the Walter Reed Army Institute of Research. Cells were cultured with Dulbecco's modified Eagle's medium or RPMI 1640 (Invitrogen) containing 10% fetal bovine serum (Sigma) and 100 μg of penicillin-streptomycin/ml.

Construction of expression plasmids.

Human sPLA₂s (PLA₂ groups IIA [GenBank loci NM_000300], IID [NM_012400], III [NM_015715], V [NM_000929], VII [NM_005084], X [NM_003561], and XIIA [NM_03081]) were first PCR amplified from corresponding PLA₂ cDNA clones obtained from Invitrogen or Openbiosytems and then subcloned into the mammalian expression vector CMV/R-mcs. A linker (4 repeats of GGGS) and a six-His tag were added to the carboxy-terminal end of the sPLA₂ group X gene, and a carboxy-terminal six-His tag alone was added to the other genes. Point mutants were constructed by using overlap extension PCR or the QuikChange site-directed mutagenesis kit (Stratagene) according to the manufacturer's protocol. All plasmids were sequenced to verify the coding regions.

The primer sequences for amplification of the sPLA₂ isoforms are as follows: for the IIA isoform, the 5′ sequence is ACCGTTAGCGGCCGCCACCATGAAGACCCTCCTACTGTTGGCAGTGATCATGA and the 3′ sequence is TGCCAGTTCTAGATCAATGATGATGATGATGATGGCAACGAGGGGTGCTCCCTCTGCAGTGTTTATTG; for IID, the 5′ sequence is ACCGTTAGCGGCCGCCACCATGGAACTTGCACTGCTGTGTGGGCTGGTGGTGATGGCTGGTG and the 3′ sequence is TGCCAGTTCTAGATCAATGATGATGATGATGATGGCACCCAGGGGTCTGCCCCCGGCAGTGGGGCC; for III, the 5′ sequence is ACCGTTAGCGGCCGCCACCATGGGGGTTCAGGCAGGGCTGTTTGGGATGCTGGG and the 3′ sequence is TGCCAGTTCTAGATCAATGATGATGATGATGATGCTGGCTCCAGGACTTCTGCTGCCTGT; for V, the 5′ sequence is ACCGTTAGCGGCCGCCACCATGAAAGGCCTCCTCCCACTGGCTTGGTTCCTGGC and the 3′ sequence is TGCCAGTTCTAGATCAATGATGATGATGATGATGGGAGCAGAGGATGTTGGGAAAGTATTGGTAC; for VII, the 5′ sequence is ACCGTTAGCGGCCGCCACCATGGTGCCACCCAAATTGCATGTGCTTTTCTGCC and the 3′ sequence is TGCCAGTTCTAGATCAATGATGATGATGATGATGATTGTATTTCTCTATTCCTGAAGAGTTCTGTAAC; for X, the 5′ sequence is GGTCGACCATGG GGCCGCTACCTGTGTG and the 3′ sequences are GGATCCCCCTCCGCTTCCCCCTCCGCTTCCCCCTCCGCTTCCCCCTCCGTCACACTTGGGCGAGTCCGGCTC (sPLA₂-X-linker) and CAGATCTCAATGGTGATGGTGATGATGGGATCCCCCTCCGCTTCCCC (linker-six-His); and for XIIA, the 5′ sequence is ACCGTTAGCGGCCGCCACCATGGCCCTGCTCTCGCGCCCCGCGCTCACCC and the 3′ sequence is TGCCAGTTCTAGATCAATGATGATGATGATGATGAAGATCAGTTTTTTCTTCATAATGACACCTGCA.

The primer sequences used for point mutants are as follows: for the D47K mutant, the 5′ sequence is GACTGGTGCTGCCATGGCCACAAGTGTTGTTACACTCGAGC and the 3′ sequence is GCTCGAGTGTAACAACACTTGTGGCCATGGCAGCACCAGTC; for the H46N, D47E, and Y50F mutants, the 5′ sequence is CTGCCATGGCAACGAGTGTTGTTTCACTCGAGCTGAGGAGGCCGGCTGCAGCC and the 3′ sequence is GGCTGCAGCCGGCCTCCTCAGCTCGAGTGAAACAACACTCGTTGCCATGGCAGC.

Transfection and Western blot analysis.

293 cells were transfected using calcium phosphate (Promega), and cell culture supernatants were harvested 2 days after transfection and kept at −80°C. Cell culture supernatants were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene difluoride membrane (Bio-Rad). The membrane was incubated with a rabbit polyclonal anti-His antibody (1:1,000; Santa Cruz Biotechnology) for 1 h at room temperature in blocking buffer (Tris-buffered saline, 3% skim milk, 0.5% Triton X-100), followed by washing. The blot was further incubated in blocking buffer with horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (IgG; 1:5,000; Santa Cruz) for 30 min and then washed. Detection was performed with the ECL reagent (Amersham).

Recombinant sPLA₂ protein purification.

The baculovirus expression vector was made following the standard protocol as described by the company (Invitrogen). Briefly, sPLA₂-X cDNA and three sPLA₂-X amino acid mutants (with an H46N, D47E, or Y50F mutation) were cloned into pVL1393 (transfer vector), which has an Autographa californica multiple nucleopolyhedrosis virus polyhedron enhancer-promoter sequence to drive high expression. The recombinant DNA was verified by sequencing. This plasmid was cotransfected with linearized BD baculoGold baculovirus DNA (BD Biosciences) in Sf9 insect cells to make a recombinant baculovirus. The plaque-purified virus was checked for the presence of the PLA₂ gene and was amplified by reinfecting Sf9 insect cells. This high-titer recombinant virus was later used to make PLA₂ protein in High Five (Hi5) cells.

Culture supernatant from 1 liter of Hi5 cells infected with the baculovirus described above was harvested after a 48-h incubation at 27°C. The sample was adjusted to 1× phosphate-buffered saline and 10 mM imidazole with a 1 M stock, filtered (through a 0.45-μm-pore-size polyethersulfone membrane), applied to a 5-ml HisTrap column (GE Healthcare), and eluted with a 20-column-volume linear imidazole gradient to 400 mM. The fractions were analyzed by SDS-PAGE. Final samples were dialyzed to 1× phosphate-buffered saline and concentrated using 10,000-molecular-weight-cutoff Amicon Ultra filtration devices (Millipore).

Wild-type sPLA₂-X and mutant sPLA₂-X (D47K) were also expressed in 293 cells, and the culture supernatants were applied to a 5-ml HisTrap column and eluted as described above.

sPLA₂ enzymatic assay.

To measure sPLA₂ enzymatic activity in the cell culture supernatant from the indicated DNA-transfected cells, an sPLA₂ assay kit (Cayman Chemical) was used according to the manufacturer's recommendations. This assay uses the 1, 2-dithio analog of diheptanoyl phosphatidylcholine as a substrate for sPLA₂s. Upon hydrolysis of the thio ester bond at the sn-2 position by sPLA₂, free thiols are detected using 5, 5-dithio-bis-(2-nitrobenzoic acid) (DTNB) at 405 nm. The specific activity of sPLA₂ was calculated based on the initial slope of the time dependence of absorption at 405 nm, using an extinction coefficient at 405 nm (ɛ₄₀₅) of 12.8 mM⁻ cm⁻.

Viruses.

Luciferase-expressing lentiviral vectors pseudotyped with envelopes from HIV-1_ADA, HIV-1_IIIB, Ebola virus, and Moloney murine leukemia virus (MoMuLV) were prepared by transient cotransfection of 293T cells with calcium phosphate (Promega) (28). Briefly, the packaging vector pCMVΔR8.2 (7 μg), pHR′CMV-Luc (7 μg), and the envelope-expressing vector pSVIII-ADA (10 μg), pRSV-IIIB (10 μg), pVR1012-GP(Z) (50 ng), pNGVL-Env (4070A) (2 μg), or CMV/R-8kb influenza virus H5 [A/Thailand/1(KAN-1)/2004] HA-wt/h (50 ng) were cotransfected. Supernatants were harvested 72 h after transfection, filtered with a 0.45-μm-pore-size syringe filter, and stored at −80°C.

Adenovirus type 5 (Ad5)-luciferase was made as described previously (2). Wild-type HIV-1_BaL and HIV-1_MN stocks were prepared in peripheral blood mononuclear cells (PBMCs) as previously described (13).

Infection of cells with pseudoviruses and luciferase assay.

A total of 30,000 cells were plated into each well of a 48-well dish the day before infection: MAGI-CCR5 cells for HIV-1_ADA, HIV-1_IIIB, and MoMuLV, and 786-O cells for Ebola virus and Ad5. The pseudovirus supernatant (50 to 100 μl) or 1.5 × 10⁷ viral particles of Ad5 (500/cell) were incubated with sPLA₂ or its mutant-transfected cell culture supernatant for 1 h at 37°C and added to the target cells. Cells were replenished with fresh medium at 16 to 18 h postinfection. After 48 h, cells were lysed in 80 μl of cell lysis buffer (Promega) in the plate, and 20 μl of the cell lysate was used in a luciferase assay with luciferase assay reagent (Promega) according to the manufacturer's recommendations. Luciferase assay results were measured using Top Count (Packard).

HIV single-round replication assay.

To assess the effect of sPLA₂-X on live wild-type HIV-1_BaL and HIV-1_MN, the virus (100 ng of p24) was incubated with 53 ng of purified sPLA₂-X (activity, 400 nmol/min) or its H46N, D47E, or Y50F mutant for 60 min at 37°C. A3R5 cells (1 × 10⁶ cells) were added to the above-described mixtures for 2 h, allowing infection. Cells were washed and incubated with fresh medium. After 64 h, cells were stained with a fluorescein isothiocyanate (FITC)-conjugated anti-p24Gag antibody (KC-57-FITC; Beckman Coulter) and analyzed (13).

Analysis of p24 release from virions by use of a density gradient.

Density gradient-purified Ebola pseudovirus (50 μl) or an HIV-1_BaL (2.5 μg of p24)-sPLA₂ mixture was added to the same volume of OptiPrep (Axis-Shield PoC). A density gradient was formed by centrifugation at 421,000 × g for 3.5 h with an NVT100 rotor (Beckman). The collected fractions were weighed, and density was calculated. An equal amount of each fraction (20 μl) was separated on a 4-to-15% SDS-PAGE gel (Bio-Rad), transferred to a polyvinylidene difluoride membrane, and blotted with human anti-HIV-1 IgG or rabbit anti-p24Gag serum (Advanced Biotechnologies). Each lane of the Western blot represents one fraction of the density gradient.

RESULTS

To define the potential of mammalian sPLA₂ to confer protection against viral infection, plasmid expression vectors encoding the human group IIA, IID, III, V, VII, X, and XIIA isoforms were prepared and tagged with a COOH-terminal polyhistidine epitope to facilitate detection. When tested for enzymatic activity, groups IIA, III, VII, and X displayed significant sPLA₂ enzymatic activity compared to control supernatants (vector) (P < 0.05 for IIA; P < 0.01 for III, VII, and X), and sPLA₂-X was the most active (Fig. 1a, top). Expression of each sPLA₂ was also confirmed by Western blotting with an anti-His antibody (Fig. 1a, bottom). The antiviral effects of recombinant human sPLA₂ cell culture supernatants were tested first by measuring the luciferase reporter gene activity of HIV-1 pseudoviruses on MAGI-CCR5 target cells, a human cervical carcinoma (HeLa) cell line expressing CD4 and coreceptors CXCR4 and CCR5. Among the different sPLA₂s, the group X isoform showed marked inhibition of the HIV-1_IIIB pseudotype reporter (Fig. 1b). Though sPLA₂-X displayed the highest enzymatic activity on this substrate, other isoforms were readily detectable by protein expression. There is evidence that different sPLA₂s have different substrate affinities that may determine their biologic effects (20), suggesting that there is specificity for this effect among the isoforms. In addition, the substrate, a 1,2-dithio analog of diheptanoyl phosphatidylcholine, may be useful in predicting antiviral efficacy, possibly because it may be related to viral envelope lipids.

FIG. 1.

The antiviral effect of the sPLA₂-X isoform is specific. (a) (Top) The enzymatic activity of each indicated sPLA₂ gene product in culture supernatant was assessed by a colorimetric assay using an sPLA₂ assay kit. Symbols indicate significant differences from the control: †, P < 0.05; *; P < 0.01. (Bottom) Expression in supernatants was determined by Western blot analysis with an anti-His antibody. (b) An HIV-1_IIIB envelope-pseudotyped lentivirus vector encoding luciferase (100 μl each) was incubated with 1 ml of cell culture supernatant, made from a control or the indicated sPLA₂ isoform from transfected 293 cells, for 60 min at 37°C. The virus-cell culture supernatant mixture was added to MAGI-CCR5 cells and incubated for 16 h. The mixture was removed, and luciferase reporter activity was evaluated 48 h after replacement with fresh medium. Data are averages ± the standard deviations from triplicates and each value is representative of two independent experiments.

To examine whether catalytic activity was required for the inhibitory effect of sPLA₂-X, wild-type, enzymatically active protein and a catalytically inactive point mutant (with the D47K mutation), termed ΔsPLA₂-X, were generated. Though equivalent amounts of proteins were detected, ΔsPLA₂-X showed no catalytic activity (Fig. 2a, left). While enzymatically active sPLA₂-X markedly inhibited reporter gene expression, similar protein concentrations of inactive ΔsPLA₂-X exerted no effect (Fig. 2a, center). sPLA₂-X acted primarily through damage to virions, as evidenced by the fact that treatment of the target cells of infection did not significantly reduce viral gene transfer (Fig. 2a, right).

FIG. 2.

Antiviral effect of sPLA₂-X: dependence of enzymatic activity on the virus and not on the target cells, and specificity of inhibition. (a) sPLA₂-X acts on the virus rather than the target cells. (Top left) The enzymatic activity of purified sPLA₂-X (WT) or the inactive ΔsPLA₂-X (D47K) mutant (Δ) made from 293 cells was assessed by a sPLA₂ assay kit. (Bottom left) Protein amounts in 10 μl are shown by Western blot analysis using an anti-His antibody. (Center) HIV-1_ADA pseudovirions were incubated with sPLA₂-X or inactive ΔsPLA₂-X (0.3 ml) for 60 min at 37°C and ultracentrifuged at 48,400 × g for 1 h to pellet the virus. Viral pellets were resuspended with fresh medium and incubated with MAGI-CCR5 target cells for 17 h. Infectivity was assessed with a luciferase reporter 48 h after replacement with fresh medium. (Right) MAGI-CCR5 target cells were incubated with sPLA₂-X or the catalytically inactive ΔsPLA₂-X (D47K) (0.3 ml) for 2 h at 37°C, washed, and transduced with pseudotyped HIV-1_ADA virions. Cells were again washed at the indicated times to remove the virions and were cultured with fresh medium. Infectivity was assessed by luciferase reporter activity 3 days later. (b) sPLA₂-X exerts specific antiviral activity. Cell culture supernatants (1 ml) made from 293 cells transfected with sPLA₂-X (activity, 33 to 78 nmol/min/ml) or ΔsPLA₂-X (D47K; catalytically inactive mutant) were incubated for 60 min at 37°C with the indicated pseudovirions. The virus-supernatant mixture was added to MAGI-CCR5 cells (for HIV-1_ADA, HIV-1_IIIB, and MoMuLV) or 786-O cells (for Ebola virus and Ad5), incubated for 16 h, and replaced with fresh medium, and luciferase-reporter activity was assayed 48 h later. Data are averages ± standard deviations from triplicates.

The specificity of the sPLA₂-X antiviral effect was assessed on different viral envelopes expressed on lentivirus vectors, including CXCR4-tropic HIV-1_IIIB, CCR5-tropic HIV-1_ADA, amphotropic MoMuLV, Ebola virus glycoprotein (GP), or a nonenveloped viral vector, recombinant Ad5. Wild-type sPLA₂-X showed significant antiviral activity against CCR-5- or CXCR-4-tropic HIV Env, amphotropic MoMuLV, and Ebola virus compared to ΔsPLA₂-X but did not show significant inhibition of reporter gene expression by the nonenveloped virus recombinant Ad5 (Fig. 2b), suggesting that the antiviral activity required the presence of a lipid-containing viral membrane.

The antiviral effect of sPLA₂-X was assessed against HIV-1_BaL (CCR5-tropic) and HIV-1_MN (CXCR4-tropic) stocks produced in PBMCs. Virus preparations were incubated with purified sPLA₂-X or a different catalytically inactive mutant, Δ3sPLA₂-X (H46N, D47E, and Y50F mutations) (9, 19), prior to infection of the human T-cell leukemia cell line A3R5, a subline of A3.01 cells (10) expressing both CCR5 and CXCR4. Flow cytometric analysis of intracellular Gag protein was used to assess viral replication. sPLA₂-X treatment substantially reduced T-cell infection by CCR5-tropic HIV-1_BaL (Fig. 3a, right) compared to the catalytically inactive Δ3sPLA₂-X (Fig. 3a, left). A similar reduction in viral replication was seen when sPLA₂-X was incubated with replication-competent CXCR4-tropic HIV-1_MN (Fig. 3b), suggesting that this antiviral mechanism is effective against diverse lentiviruses with alternative chemokine receptor specificity.

FIG. 3.

sPLA₂-X inhibits productive replication of CCR5- or CXCR4-tropic HIV-1 strains in T cells. HIV-1_BaL (a) or (b) HIV-1_MN (100 ng of p24) stocks were incubated with 53 ng of purified sPLA₂-X (activity, 400 nmol/min) or Δ3sPLA₂-X (H46N, D47E, and Y50F mutations) for 60 min at 37°C. The virus-sPLA₂ mixture was incubated with the human T-cell leukemia cell line A3R5 (a subline of CEM expressing both CCR5 and CXCR4; 1 × 10⁶) for 2 h. Cells were then washed and replaced with fresh medium. HIV-1 replication was analyzed 64 h after infection by flow cytometry, staining for intracellular p24 with an FITC-conjugated anti-p24 antibody. The percentage of p24-positive cells was subtracted from that of mock-infected cells.

To understand the mechanism of the sPLA₂-X antiviral effect, the ability of sPLA₂-X to lyse virus was examined both for pseudotyped lentivirus vectors and for replication-competent HIV-1_BaL derived from PBMCs. For the pseudotyped lentivirus vector, Ebola virus GP pseudotypes were analyzed first, using gradient-purified virions. The presence of p24Gag in different gradient fractions was first confirmed by immunoprecipitation followed by Western blotting, with peak activity at a density of 1.10 (Fig. 4a, right panel, lane 3). Analysis of virions from this purified fraction revealed reactivity with monoclonal antibody 13C6, known to bind Ebola virus GP on virions (27) (Fig. 4a, left panel). This subtype IgG2a antibody has been shown to fix complement (27). Gradient-purified pseudotyped virions were treated with control mouse IgG or 13C6 plus mouse complement. Though virions reacted with this antibody and are able to fix complement, no release of p24Gag was detected, as shown by refractionation through the density gradient (Fig. 4a, right panel, lanes 5 to 7). In contrast, treatment with sPLA₂-X, but not ΔsPLA₂-X (D47K), caused Gag release when these virions were refractionated through a density gradient (Fig. 4a, lower right panel, sPLA₂-X versus ΔsPLA₂-X, lanes 12 to 14). A similar effect was observed with 2F5, a broadly neutralizing human monoclonal antibody of subtype IgG1 that binds HIV-1_BaL (Fig. 4b), confirming its effect on native virus.

FIG. 4.

sPLA₂-X potently damages viral membranes compared to antibody-mediated complement fixation. (a) 13C6, a complement-fixing antibody, binds to Ebola virus pseudovirions but, unlike sPLA₂-X, does not damage the viral membrane. Gradient-purified Ebola virus pseudovirions were incubated with a control mouse IgG or 13C6 for 30 min at 4°C and then immunoprecipitated with protein G-Sepharose. The immunoprecipitate was analyzed for p24 by Western blot analysis using human anti-HIV-1 IgG (left panel). Gradient-purified Ebola virus-pseudotyped virions were incubated with mouse IgG (67 μg/ml) or 13C6 (333 μg/ml) plus mouse complement (10%) for 90 min at 37°C (top right panels) or with 1 ml of sPLA₂-X or ΔsPLA₂-X (D47K) from transfected 293 cell culture supernatants for 60 min at 37°C (bottom right panels). A density gradient was formed by centrifugation using OptiPrep, and the fractions were collected. p24Gag in each fraction is shown by Western blot analysis with anti-HIV-1 IgG. Gag released from damaged virus forms aggregates found in higher-density fractions. (b) 2F5, an antibody known to fix complement, binds to HIV-1_BaL but, unlike sPLA₂-X, does not damage the viral membrane. Purified live HIV-1_BaL was incubated with KZ52 (IgG1) or 2F5 (IgG1) for 30 min at 4°C and immunoprecipitated with protein G-Sepharose. The immunoprecipitate was analyzed for p24 by Western blot analysis using anti-p24 rabbit serum for the presence of 2F5 bound to HIV-1_BaL (left panel). HIV-1_BaL was incubated with 100 μg/ml of monoclonal antibody 2F5 or KZ52 with human complement (10%) (top right panel) or 1 ml of culture supernatants from 293 cells transfected with sPLA₂-X or ΔsPLA₂-X (D47K) (bottom right panel) for 3 h at 37°C. A density gradient was formed by centrifugation using OptiPrep, and the fractions were collected. p24Gag in each fraction is shown by Western blot analysis using rabbit anti-p24Gag serum. Gag released from damaged virus forms aggregates found in higher-density fractions.

DISCUSSION

In this study, the ability of sPLA₂s to inhibit HIV-1 replication has been evaluated. We find that sPLA₂-X displays antiviral activity against diverse lentiviruses by degradation of the viral membrane. sPLA₂-X inhibits replication of both CXCR4- and CCR5-tropic HIV-1 in primary human CD4⁺ cells. This effect was observed despite the resistance of virus preparations to lysis by antibody-mediated complement activation, suggesting that this mechanism acts in cases where the acquired immune response is ineffective.

The antibacterial effects of sPLA₂s against gram-positive bacteria (ranked in order of strength, from highest to lowest, as follows: IIA, X, V, XII, IIE, IB and IIF) and the gram-negative bacterium Escherichia coli have been reported previously; only human group XII displays detectable bactericidal rather than bacteriostatic activity (7, 11, 12, 26). In this study, group X alone exerted an antiviral effect on enveloped lentiviruses, documenting the specificity of this effect. While it may be suggested that the efficacy of sPLA₂-X is due to its high enzymatic activity, it should be recognized that this activity relates to the specificity of the enzyme for the substrate used in this assay and that various sPLA₂s have divergent substrate specificity (20). For example, sPLA₂-XII is not active in this assay yet mediates a significant antimicrobial effect. Taken together, this study raises the possibility of a novel and specific role for this gene product in innate immunity to a specific class of viruses.

sPLA₂-X efficiently hydrolyzes cell membranes primarily because of its high binding affinity for phosphatidylcholine, a phospholipid that is enriched in the outer leaflet of the plasma membrane. The viral membrane of HIV-1 (5) is rich in zwitterionic phospholipids (phosphatidylcholine and sphingomyelin) and may be more susceptible to sPLA₂-X. The sensitivity of lentiviruses to sPLA₂-X despite their resistance to antibody and complement should be noted. This difference suggests that other factors, including the spike density of the virion, glycosylation of envelopes, curvature of the viral particle, and cellular proteins on the viral envelope, may affect susceptibility to different antiviral immune and inflammatory responses.

The enzymatic activity of sPLA₂-X is necessary and sufficient for the antiviral effect. This finding contrasts with a previous report showing that addition of a nonmammalian sPLA₂, derived from bee venom, blocked the entry of HIV-1 by steric inhibition of the chemokine receptor on target cells, in which case catalytic activity was not required (6). Another recombinant sPLA₂-X derived from bacteria inhibited adenovirus plaque formation through its ability to hydrolyze phospholipids on host cell membranes (15), implying a different mechanism of action with a nonphysiological gene product, unlike the report here.

The disruptive action of sPLA₂-X on the viral membrane was strongly confirmed by p24Gag protein redistribution in a high-density gradient fraction (Fig. 4). Antibody and complement did not show p24 release from virions. Although some reports have suggested that HIV-1 is susceptible to complement-mediated lysis (1, 22, 25), these studies have utilized poorly defined sera from HIV-infected subjects or nonphysiological concentrations of polyclonal antibodies and can be explained by factors other than complement that might cause viral lysis, possibly even including sPLA₂, which is found in the circulation as well. Here, purified monoclonal antibodies known to fix the complement on the HIV-1 envelope do not mediate lysis. It is also well known that HIV-1 virions escape complement-mediated lysis in vivo through complement inactivators such as CD46, CD55, and CD59 on their viral membranes (21). Since sPLA₂-X readily degrades such viruses, we suggest that sPLA₂-X may overcome resistance to antibody and complement virolysis. Further, because it is expressed at the highest levels in the intestinal mucosa, a primary site of HIV replication in natural infection, we suggest that sPLA₂-X be studied further to explore its potential role in innate immunity against HIV in the gut. Expression of this recombinant protein, as well as the stimulation of increased endogenous sPLA₂-X, may help to limit viral replication and reduce the incidence of productive replication at sites of primary infection.