Cyanovirin-N Binds to gp120 To Interfere with CD4-Dependent Human Immunodeficiency Virus Type 1 Virion Binding, Fusion, and Infectivity but Does Not Affect the CD4 Binding Site on gp120 or Soluble CD4-Induced Conformational Changes in gp120

Abstract

Cyanovirin-N (CV-N), an 11-kDa protein isolated from the cyanobacterium Nostoc ellipsosporum, potently inactivates diverse strains of human immunodeficiency virus type 1 (HIV-1), HIV-2, and simian immunodeficiency virus. While it has been well established that the viral surface envelope glycoprotein gp120 is a molecular target of CV-N, the detailed mechanism of action is of further interest. We compared matched native and CV-N-treated virus preparations in a panel of assays that measure viral replication, assessing successive stages of the viral life cycle. CV-N-treated virions failed to infect cells as detected by p24 production and quantitative PCR for HIV-1 reverse transcription products, whereas treatment of the target cells did not block infection, confirming that CV-N acts at the level of the virus, not the target cell, to abort the initial infection process. Compared to native HIV-1 preparations, CV-N-treated HIV-1 virions showed impaired CD4-dependent binding to CD4⁺ T cells and did not mediate “fusion from without” of CD4⁺ target cells. CV-N also blocked HIV envelope glycoprotein Env-induced, CD4-dependent cell-cell fusion. Mapping studies with monoclonal antibodies (MAbs) to defined epitopes on the HIV-1 envelope glycoprotein indicated that CV-N binds to gp120 in a manner that does not occlude or alter the CD4 binding site or V3 loop or other domains on gp120 recognized by defined MAbs and does not interfere with soluble CD4-induced conformational changes in gp120. Binding of CV-N to soluble gp120 or virions inhibited subsequent binding of the unique neutralizing MAb 2G12, which recognizes a glycosylation-dependent epitope. However, prior binding of 2G12 MAb to gp120 did not block subsequent binding by CV-N. These results help clarify the mechanism of action of CV-N and suggest that the compound may act in part by preventing essential interactions between the envelope glycoprotein and target cell receptors. This proposed mechanism is consistent with the extensive activity profile of CV-N against numerous isolates of HIV-1 and other lentiviruses and supports the potential broad utility of this protein as a microbicide to prevent the sexual transmission of HIV.

Currently, more than 30 million people are infected with human immunodeficiency virus (HIV) worldwide (51). The main route of transmission of the virus is through heterosexual contact, which accounts for 75 to 85% of all HIV infections (51). The highly mutable nature of HIV and the daunting complexities of developing a broadly protective vaccine against the multiple clades of HIV are increasingly apparent (12, 27, 44). With no vaccine on the horizon, there is a pressing need to develop anti-HIV microbicides to prevent the sexual transmission of HIV. The World Health Organization, the U.S. Department of Health and Human Services, and the U.S. National Institute of Allergy and Infectious Diseases have stated that the development of female-controlled topical virucides is an urgent global priority (20, 30).

HIV infection of a cell is a stepwise process beginning with the adsorption of virions to cells and binding to the CD4 receptor, leading to CD4-binding-induced conformational changes, engagement of the appropriate coreceptor, and fusion of the virion envelope with the cell membrane. Later steps include transfer of the viral capsid into the cytoplasm, uncoating of the virus to render the viral genome accessible for reverse transcription and integration, and subsequent production of viral transcripts, proteins, and progeny virions (16). Upon virion binding to CD4, the HIV envelope glycoprotein (Env) undergoes conformational changes (57, 68) that appear to be required for subsequent interactions between the chemokine receptor binding surface of gp120 and a member of the chemokine receptor family (54, 68). This conformational change is thought to expose the transmembrane gp41 protein of the virion to initiate target membrane fusion (15, 16, 38, 56). A compound that could irreversibly block one or more of these initial steps (binding, CD4-induced conformational changes, fusion, or uncoating) of the HIV infection process might serve as an effective microbicide to block infection. One such recently discovered compound that inhibits the infectivity and cytopathic effects of HIV is cyanovirin-N (CV-N) (10).

CV-N, an 11-kDa (101-amino-acid) protein with potent HIV-inhibitory activity, was originally isolated from an aqueous extract of the cyanobacterium Nostoc ellipsosporum as part of a project to identify novel natural products with anti-HIV activity (10, 25). Recombinant CV-N has also been produced in Escherichia coli and is indistinguishable from natural CV-N (10, 45). In contrast to soluble CD4 (sCD4) and most known neutralizing antibodies (Abs) that bind gp120, CV-N exerts broad virucidal activity at nanomolar concentrations, against both primary isolates and laboratory-adapted strains of primate immunodeficiency retroviruses. These include T-lymphocyte-tropic, macrophage-tropic, and dual tropic primary clinical isolates of HIV type 1 (HIV-1), as well as laboratory-adapted strains of HIV-1, HIV-2, and simian immunodeficiency virus (SIV) (10). Previous results have shown that gp120 is a molecular target of CV-N (10, 46), while other findings indicated that CV-N did not visibly disrupt the virion ultrastructure (40). CV-N is extremely resistant to physicochemical degradation and can withstand treatment with denaturants, detergents, organic solvents, multiple freeze-thaw cycles, and heat (up to 100°C) with no apparent loss of antiviral activity (10). The nuclear magnetic resonance structure of recombinant CV-N has been solved, revealing a largely β-sheet protein with twofold pseudosymmetry (8). However, CV-N has no sequence or structural homology with known proteins, and its physiological function in the cyanobacterium is unknown. The mechanism underlying the HIV-inhibitory activity of CV-N has not been fully elucidated, although initial results in certain binding assay formats indicated that CV-N is able to bind diverse gp120 molecules, despite the known extensive sequence variation between virus isolates (10, 40).

To better understand the molecular mechanism(s) of CV-N inactivation of HIV, we used a panel of assays intended to track successive stages of the viral life cycle. These included (i) infectivity cultures, (ii) a quantitative PCR-based viral entry assay, (iii) a virus-induced “fusion from without” assay, (iv) an Env-mediated cell fusion assay, (v) a flow cytometric whole-particle virus binding assay, and (vi) epitope mapping assays in multiple formats to determine if CV-N binding affected exposure of defined epitopes on the envelope glycoprotein.

These studies demonstrate that CV-N binds to gp120 in a manner that occludes or alters the 2G12 epitope and prevents CD4-dependent virion binding, fusion, and infectivity. However, CV-N does not detectably alter the primary CD4 binding site (CD4bs) on gp120, nor does it affect the binding of sCD4 to virions or subsequent sCD4-induced conformational changes in the envelope glycoprotein. These data suggest that the mechanism of action of CV-N may involve interference with essential interactions between the viral envelope glycoprotein and target cell receptors. CV-N should be a valuable reagent to further examine the early steps of virion binding and fusion and appears promising as a candidate microbicide to prevent the sexual transmission of HIV and AIDS.

MATERIALS AND METHODS

HIVs.

HIV-1_MN/H9 clone 4 and HIV-1_IIIB were propagated in H9 cells, as described elsewhere (49). Where indicated, concentrated virus preparations (12,500 ng of p24^CA per ml) were produced by sucrose gradient banding in a continuous-flow centrifuge (7). All virus stocks were stored at −70°C or in vapor-phase liquid nitrogen until use.

Virus infectivity assays.

Virus infectivity assays were performed essentially as described previously (42), with AA2 cells (14, 65). Briefly, 2 × 10⁶ indicator cells in 3-ml volumes were inoculated with native or CV-N-inactivated (see below) virus stocks. Cells were cultured in RPMI 1640 with 10% heat-inactivated fetal bovine serum, 2 mM l-glutamine, 100 U of penicillin G per ml, and 100 μg of streptomycin sulfate per ml (complete medium); 200 μl of medium was replaced twice weekly. On days 0, 3, 6, 9, 12, and 15 postinoculation, supernatants were harvested and tested for p24^CA content as an index of productive infection, by a capture enzyme-linked immunosorbent assay (ELISA) (AIDS Vaccine Program, National Cancer Institute-Frederick Cancer Research and Development Center, Frederick, Md.).

CV-N inactivation of HIV-1_MN and HIV-1_IIIB.

For all procedures, frozen virus stocks were quickly thawed at 37°C in a water bath. For inactivation with CV-N, a stock solution of CV-N (10 μM in phosphate-buffered saline [PBS]) was prepared and added directly to virus to produce the desired CV-N concentration. Virus preparations were treated for 60 to 90 min at 4°C. For the viral entry assay shown in Fig. 1, free CV-N was removed by ultrafiltration, with a centrifugal filtration device with a 500-kDa-cutoff membrane (Centriprep 500; Amicon, Beverly, Mass.). Control virus preparations were mock treated with bovine serum albumin (BSA) and processed in parallel with inactivated samples.

FIG. 1.

CV-N interacts with HIV virions but not host cells to inhibit HIV infection. HIV-1_MN was either mock treated or pretreated with CV-N (200 nM) at 4°C for 90 min before performance of filtration dialysis twice through a 500-kDa-cutoff membrane to remove free CV-N. AA2 cells were either mock treated (—◊—) or pretreated with CV-N (—□—) (200 nM) at 37°C for 90 min and washed three times before being plated in duplicate at 0.5 × 10⁶ cells per well in a total volume of 3 ml and infected with HIV-1_MN. CV-N-treated HIV-1_MN (—●—) or HIV-1_MN and CV-N (—▴—) were added to AA2 cells on day 0 and left in for the duration of the 10-day assay. (A) p24 accumulation. One hundred microliters of culture supernatant was sampled every 3 days to assay for virus production as measured by capsid p24 content. CV-N (200 nM) was nontoxic to the cells (by trypan blue exclusion [data not shown]). Nonincreasing amounts of p24 seen for CV-N-treated HIV or simultaneous addition of CV-N and HIV reflect residual virus inoculum. (B) Viral entry assay (PCR). The experiment in panel A was set up in duplicate, and the cells were harvested 24 h postinoculation and assayed for gag DNA by quantitative real-time PCR as described in Materials and Methods. ∗, no gag DNA detected (<30 copies per reaction). Data are representative of three independent experiments. PBGD, porphobilinogen deaminase.

Cell lines.

The H9, A3.01, and Sup-T1 cell lines were obtained from the National Institute of Allergy and Infectious Diseases AIDS Research and Reference Reagent Program (Rockville, Md.). The AA2 cell line (65) was provided by R. Benveniste (National Cancer Institute-Frederick Cancer Research and Development Center). All cell lines were mycoplasma negative (PCR mycoplasma detection kit; American Type Culture Collection, Manassas, Va.) and were cultured in complete medium.

Ab reagents.

The generation and characterization of polyclonal goat Abs against human major histocompatibility complex (MHC) class I and MHC class II and a polyclonal goat antiserum raised against microvesicles prepared from cultures of H9 cells have been described previously (3, 7). The murine monoclonal antibodies (MAbs) 48.d and 17b (62) are directed against CD4-induced conformational determinants on gp120. The polyclonal antiserum W0/07, which recognizes an epitope in the V3 domain (29); the CD4bs-directed MAbs F105 (52) and immunoglobulin G1b12 (IgG1b12) (13), which recognize discontinuous conformational epitopes; and sCD4 protein (2) were all obtained from the AIDS Research and Reference Reagent Program. MAb 2G12, which recognizes a glycosylation-dependent conformational epitope comprised of residues from the C3, C3, V4, and C4 regions of gp120 (64), and MAb 2F5, which recognizes an epitope in gp41 (48), were graciously provided by H. Katinger. The C108G MAb, which recognizes a glycan-dependent epitope on V2, has been previously described (67). MAbs SC258 and G3-4 recognize V2 epitopes (69), MAbs 50.23 and 110.5 recognize V3 epitopes (6), and MAbs G3.299 and G3-42 recognize C4 epitopes on gp120.

Whole-virion immunoprecipitation assay.

A whole-virion immunoprecipitation assay was performed essentially as described elsewhere (3, 55). Briefly, comparable input amounts of native or CV-N-treated virus preparations (25 ng of p24^CA per ml) were incubated overnight at 4°C on a rocker with empirically optimized concentrations of each Ab. For some studies, sCD4 (0.05 to 1.0 μg/ml) was added to virus preparations after CV-N treatment but before addition of Abs. Formalin-fixed Staphylococcus aureus Cowan (GIBCO, Grand Island, N.Y.) (25 μl) was then added, and after incubation at 20°C for 30 min, virions with bound Ab were immunoprecipitated by centrifugation (2,000 × g, 30 min). The residual virus content of the supernatant after immunoprecipitation was determined by p24 capture immunoassay and compared to the particle-associated (ultracentrifugation-pelletable, 100,000 × g, 1 h) p24 content of the same virus preparation prior to immunoprecipitation to calculate the percent clearance of viral particles. Clearance by a particular Ab in this assay is indicative of the presence of intact antigen on the surface of the virions (3).

ELISA studies for gp120 Ab or CV-N binding.

Ninety-six-well assay plates (Nunc Life Technologies, SARL, Ceigy Pontoise, France) were coated overnight at 4°C with 50 μl of a 10-μg/ml solution per well in bicarbonate buffer (pH 8.6) of MAb D7324, which is directed against the carboxy-terminal region of gp120 (60). Coated plates were washed twice in wash buffer (PBS, 0.5% Nonidet P-40), and nonspecific binding sites were blocked with 100 μl of 2% (wt/vol) BSA in wash buffer for 1 h at room temperature followed by two washes. Soluble gp120 HIV-1_IIIB/LAI (sgp120) was then added (50 μl of a 1-μg/ml solution per well) and incubated for 2 h at room temperature followed by two washes. Fifty microliters of CV-N (125 nM) or PBS per well was then incubated for 1 h at room temperature followed by two washes. Serial dilutions of gp120-reactive MAbs (50 μl/well) were then tested for binding to the captured control or CV-N-treated sgp120. After 1 h of incubation at room temperature, plates were washed five times and bound gp120-reactive Ab was detected with mouse (Jackson Immunoresearch, West Grove, Pa.) or human (Immunotech, Westbrook, Maine) specific anti-IgG horseradish peroxidase conjugate at a 1/1,000 final dilution for 1 h at room temperature. After washing five times, 200 μl of peroxidase substrate was added for 30 min at room temperature, and optical density was read at 450 nm.

Separate ELISAs were performed to compare binding interactions of MAb 2G12 and CV-N with sgp120 pretreated with CV-N or 2G12, respectively. Fifty nanograms (50 μl of a 1-μg/ml solution) of sgp120 from HIV-1_IIIB (Intracel, Issaquah, Wash.) per well was captured as described above. Fifty microliters of CV-N (33.2 nM), MAb 2G12 (66.4 nM), or PBS per well was added to the captured sgp120 and incubated at room temperature for 2 h, followed by two washes. To evaluate the effect of prior CV-N binding to sgp120 on the ability of 2G12 to bind to sgp120, 50 μl of serial dilutions of 2G12 (0.00125 to 1.25 μg/ml) per well was incubated with CV-N-pretreated or control sgp120 for 30 min at room temperature. After three washes, MAb 2G12 bound to captured sgp120 was detected with a goat anti-human IgG alkaline phosphatase conjugate (Boehringer Mannheim, Indianapolis, Ind.). To assess the effect of prior MAb 2G12 binding to sgp120 on subsequent binding of CV-N to sgp120, 50 μl of serial dilutions of CV-N (0.026 to 83.0 nM) per well was incubated with 2G12-pretreated or control sgp120 for 30 min at room temperature. After three washes, CV-N bound to captured sgp120 was detected with a rabbit anti-CV-N polyclonal antiserum, followed by three washes and incubation with a goat anti-rabbit IgG alkaline phosphatase conjugate (Boehringer Mannheim) (10). Following incubations with alkaline phosphatase-conjugated anti-IgG reagents, substrate was added and optical density was read at 405 nm.

HIV-1 virion binding assay.

An immunofluorescence flow cytometric-based, whole-particle virion binding assay was performed as described elsewhere (55), with modifications to a previously reported assay format (58, 66). The A3.01 cell line (22) expresses CD4 and CXCR4 but does not express HLA-DR. Virus propagated in HLA-DR-expressing cells incorporates host cell-derived HLA-DR into the viral envelope (7). Thus, acquisition of HLA-DR reactivity by A3.01 cells following incubation with HLA-DR-containing virions can be used to quantitate virion binding. A3.01 cells (3 × 10⁵ per condition) were preincubated at 4°C for 30 min with either staining buffer (calcium- and magnesium-free PBS, with 1% [wt/vol] BSA) or unlabeled anti-Leu3a MAb (10 μg/ml; Becton Dickinson Immunocytometry Systems, San Jose, Calif.) and then washed once in staining buffer. Cells were incubated with 100 μl of staining buffer, native virus, or CV-N (0.02 to 2,000 nM)-treated or MAb 2G12-treated (50 μg/ml) virus preparations at 37°C for 25 min and washed twice. Immunofluorescent staining was performed (4°C for 30 min) with a phycoerythrin-conjugated MAb to HLA-DR and fluorescein isothiocyanate-conjugated MAbs to the Leu3a or OKT4 epitopes on CD4, with nonspecific Ab binding measured with isotype MAbs of irrelevant specificity, conjugated to the appropriate fluorochromes. All Abs were from Becton Dickinson Immunocytometry Systems, except for the anti-OKT4 MAb (Ortho Diagnostics, Raritan, N.J.). Following Ab staining, cells were washed three times and fixed with 2% paraformaldehyde for 60 min at 4°C prior to analysis on a FACScan flow cytometer, with CellQuest software (Becton Dickinson Immunocytometry Systems). Cells were gated by forward and 90° light scatter, and at least 10,000 events were acquired for each sample.

Fusion from without assay.

To test the ability of CV-N-treated virus to mediate CD4-dependent, HIV-1 envelope glycoprotein-mediated fusion from without (18, 55), we incubated Sup-T1 cells (10⁵ cells/well/50 μl in 96-well flat-bottom plates), which are highly susceptible to HIV-1-induced cell fusion, at 37°C with matched concentrated preparations of HIV-1 or CV-N-treated (200 nM) virions (HIV-1_MN [6,400 ng of p24 per ml], 50 μl/well). The presence of characteristic syncytia was evaluated by inverted phase-contrast microscopy 1 to 3 h following virus addition. Syncytia present at this time are due to fusion from without (that is, due to the input virus inoculum), since this is insufficient time for infection to result in cell surface expression of envelope glycoproteins and resulting syncytia (fusion from within). As a positive control for inhibition of CD4-dependent fusion, target cells were pretreated with the anti-Leu3a MAb (10 μg/ml) for 30 min, prior to addition of virus.

Cell-cell fusion assay.

The effect of CV-N on HIV-1 Env-mediated cell-cell fusion was analyzed by using a previously described reporter gene activation assay (11, 48a). Env proteins were expressed from the following plasmids containing a synthetic early-late vaccinia virus promoter: pCB-41, LAV Env (11); pCB-43, Ba-L Env (11); pGA13-89.6, 89.6 Env (1a); and pCB-16, Unc Env (11).

Effector cells were prepared by transfecting NIH 3T3 cell monolayers with designated Env-encoding plasmids followed by infection in suspension with the recombinant vaccinia virus vTF7-3 (23), which encodes the bacteriophage T7 RNA polymerase gene driven by a vaccinia virus promoter. Target cells were prepared by infecting PM1 cells (expressing endogenous CD4⁺-CXCR4⁺-CCR5⁺) (39) in suspension with the recombinant vaccinia virus vCB-21R-LacZ (1), which contains the E. coli lacZ gene linked to the T7 promoter. Following overnight incubation at 31°C to allow protein expression, effector and target cells were each washed and resuspended in Eagle’s minimum essential medium–2.5% fetal bovine serum. Effector cells (100 μl, 10⁶ cells/ml) were added to duplicate wells of 96-well plates and preincubated for 30 min at 37°C with 50 μl of PBS containing different concentrations of CV-N to reach final concentrations of 0, 10, 100, and 1,000 nM. Target cells (50 μl, 2 × 10⁶ cells/ml) were then mixed with these effector cells, and the plates were incubated for 2.5 h at 37°C. The cells were lysed with Nonidet P-40, and β-galactosidase activity was measured at 570 nm with chlorophenol-red-β-d-galactopyranoside as a substrate.

Viral entry assay.

To determine the stage of the viral life cycle at which infectivity was arrested for CV-N-treated virus preparations, we performed a viral entry assay in which reverse-transcribed viral DNA species were quantified by a real-time PCR assay. Briefly, AA2 cells were inoculated with native or CV-N-treated virus and cultured in complete medium, and aliquots were harvested at 24 to 36 h postinfection. The washed, dry cell pellets were cryopreserved at −70°C until processing and analysis. Pellets were lysed and total DNA was extracted with commercial reagents (PureGene kit; Gentra Systems, Minneapolis, Minn.) according to the manufacturer’s recommendations. HIV-1 gag DNA, indicative of completion of first-strand DNA synthesis, was quantitated by real-time PCR assay on an ABI Prism 7700 sequence detection system. The underlying principles and operation of this instrument are reviewed in detail elsewhere (26, 37, 61). For the present assays, the following reagent sets were used: Gag forward primer, 5′-GiC ATC AiG CAG CCA TGC AAA T-3′ (1366 to 1387); Gag reverse primer, 5′-CAT iCT ATT TGT TCi TGA AGG GTA CTA G-3′ (1507 to 1480); probe, 5′-(R)TCA ATG AGG AAG CTG CAG AAT GGG AT(Q)-3′ (1402 to 1427) (based on the reference sequence for HIV-1, isolate HXB2, GenBank accession no. K03455), where i indicates inosine residues, R indicates the reporter fluorochrome (6-carboxy-fluorescein), and Q indicates the quencher dye 6-carboxy-tetramethyl-rhodamine conjugated through a linker arm nucleotide (37). Fluorescent probes for HIV-1 gag DNA were obtained from Syngen Research, Inc, San Diego, Calif. In addition, each specimen was analyzed for the copy number for a unique sequence from the coding region for porphobilinogen deaminase (24), with a fluorescent probe from the Applied Biosystems Division of Perkin-Elmer (Foster City, Calif.). Since this sequence is present at two copies per diploid cell, and there are no pseudogene sequences, quantitative analysis for this sequence in a given specimen provides an internal control, allowing normalization of HIV copy number relative to the number of diploid genome equivalents of DNA present in the specimen (60a). The average interassay coefficient of variation for the real-time PCR assays for HIV-1 gag and strong-stop and porphobilinogen deaminase DNA was <15%, with a threshold sensitivity of 30 DNA copy equivalents per reaction.

RESULTS

CV-N interacts with HIV virions, not target cells, to block infection.

Previous findings indicated that the anti-HIV activity of CV-N was mediated at least in part by high-affinity interactions with gp120 (10, 46). In those studies, pretreatment of HIV-1 virions with CV-N, followed by washing to remove CV-N, completely prevented subsequent infection of CEM-SS cells (10), as measured by a quantitative infectivity (syncytium formation) assay (31). Pretreating the cells with CV-N did not prevent their subsequent infection with the untreated virus (10). However, other experiments have indicated some direct binding of CV-N to cells in the absence of gp120 (reference 47 and data not shown). To further evaluate the possibility that CV-N binding to cell surface molecules could inhibit HIV infectivity, we pretreated either the cells or virus with CV-N and removed unbound CV-N before adding the HIV virions to the target cells. Infection was assessed by viral p24 core antigen levels in culture supernatants. Figure 1A shows cumulative p24 production in AA2 cells that were infected with HIV-1_MN. Untreated cells were productively infected as detected by the increase in p24 production and formation of syncytia (data not shown) as early as 3 days postinoculation. Similarly, pretreating AA2 cells with 200 nM CV-N had little or no effect on virus infection, as comparable levels of p24 were detected 3 to 10 days postinfection (Fig. 1A). In contrast, pretreating the HIV-1 virions with CV-N, or adding CV-N to the culture at the time of virus addition, completely blocked productive infection as measured by p24 production (Fig. 1A). These results confirmed that CV-N was interacting with HIV-1 virions, rather than target cells, to prevent productive infection.

CV-N inhibits the virus life cycle prior to reverse transcription.

In parallel assays to establish whether CV-N was blocking HIV infection at a pre- or postentry step of the virus life cycle, we used a quantitative PCR assay to measure the number of reverse-transcribed gag DNA copies in target cells that had been either mock treated or treated with CV-N and inoculated with either HIV-1_MN or CV-N-treated HIV-1_MN. Using parallel samples from the same experiment whose results are shown in Fig. 1A, we isolated cells 24 h after virus inoculation, extracted total DNA, and performed a real-time quantitative PCR assay to measure reverse-transcribed viral DNA (26, 37, 61). As shown in Fig. 1B, mock treatment of the virus led to a high level of infection within 24 h. In agreement with our p24 ELISA data (Fig. 1A), pretreating the AA2 cells with CV-N did not block infection (Fig. 1B). Pretreating the virus preparation with CV-N or adding CV-N to the cells at the time of virus addition completely blocked infection, as no gag DNA copies were detected (Fig. 1B). These data further confirm and extend previous findings (10) and indicate that CV-N interacts with the virus, and not target cells, to block HIV infection at an early stage in the viral life cycle.

CV-N inhibits HIV-1 Env glycoprotein-mediated fusion.

To determine whether CV-N was preventing infection by blocking fusion of virions with target cell membranes, we examined the ability of CV-N to prevent fusion from without (18, 55). As shown in Fig. 2B, mock-treated virus fused with the Sup-T1 target cells within 2 h of inoculation, as detected by the formation of syncytia. Similarly, Sup-T1 cells which had been incubated with CV-N and then washed also formed syncytia within 2 h of virus addition (data not shown). In contrast, virions that had been treated with CV-N (Fig. 2C) or cultures where CV-N was added simultaneously with HIV (data not shown) did not undergo fusion from without, as no syncytia were detected within 24 h. Pretreating the cells with the anti-Leu3a MAb, which recognizes the gp120 binding site on CD4, completely blocked fusion (Fig. 2D), confirming that the cell fusion observed in this system was authentic, HIV envelope glycoprotein-induced, CD4-dependent cell fusion (35, 36).

FIG. 2.

CV-N blocks HIV-1-mediated fusion from without. (A) Untreated Sup-T1 T cells, highly susceptible to CD4-dependent, HIV-1 envelope-mediated cell fusion. (B) Following a 2-h incubation with concentrated native HIV-1 (6,400 ng of p24^CA per ml), characteristic syncytia are seen, reflecting virion-mediated fusion from without. (C) CV-N pretreatment (200 nM) of virions blocks fusion. (D) Fusion mediated by native virions is inhibited by prior incubation of cells with anti-Leu3a MAb (10 μg/ml).

CV-N also blocked HIV-1 Env-mediated cell fusion in a concentration-dependent fashion when preincubated with effector cells (Fig. 3). Comparable dose-response effects were observed with LAV (X4, T-cell line adapted) and 89.6 (R5X4, dualtropic primary) Envs. Similar effects of CV-N were obtained without preincubation of the effector cells and against the Ba-L Env (data not shown). These data clearly showed that CV-N prevented HIV envelope glycoprotein-induced cell fusion but left open the possibility that CV-N-treated virions were still capable of binding to the cell but were unable to undergo postbinding conformational changes required to fuse with the target cell.

FIG. 3.

Effect of CV-N on Env-mediated cell fusion. The vaccinia virus-based reporter gene cell fusion assay was used (see Materials and Methods). The indicated CV-N concentrations represent those in the final fusion mixture. The background β-galactosidase activity value (0.5), obtained with the uncleavable nonfusogenic Env mutant, Unc, was subtracted from each value obtained with the active Envs. For each Env, 100% is defined as the β-galactosidase activity obtained in the absence of CV-N (optical density/minute × 1,000; LAV and 89.6, 36.8). Error bars indicate standard deviations of the mean values obtained from duplicate samples.

CV-N blocks CD4-dependent virus binding to cells.

To determine if CV-N-treated virions were able to bind to CD4⁺ T cells, we performed a flow cytometry-based, virion-binding assay (55, 58, 66). The A3.01 cell line expresses CD4, demonstrated by the staining of both anti-Leu3a (Fig. 4A) and anti-OKT4 (data not shown) MAbs, but does not express HLA-DR (Fig. 4B). After incubation of A3.01 cells with native virions produced from HLA-DR-positive H9 cells, the A3.01 cells became HLA-DR positive (Fig. 4B), with a concomitant decrease in the availability of the Leu3a epitope (gp120 binding epitope) (Fig. 4A), but with little change in OKT4 (non-gp120-binding epitope) staining (Fig. 5A), all consistent with virions binding to the target cells. The HLA-DR staining reflects the HLA-DR determinants present on the surface of the virions, which are in turn bound to the target cells. Availability of the Leu3a epitope is decreased in the presence of the virions (Fig. 4A), as this CD4 epitope is involved in binding to virion-associated gp120 (4). Preincubation of the A3.01 cells with unlabeled anti-Leu3a MAb inhibited acquisition of approximately 56% of the HLA-DR signal (Fig. 4B), indicating that approximately 56% of the HIV-1_MN binding detected by acquisition of HLA-DR staining was CD4 dependent while the remaining 44% of the binding was CD4 independent (43). However, all fusion and infectivity were blocked by the anti-Leu3a MAb (Fig. 2 and data not shown). Pretreatment of virions with 200 nM CV-N, a concentration that completely blocked infectivity (Fig. 1A and B and Fig. 5B), inhibited virion binding as assessed by acquisition of HLA-DR signal (Fig. 4B). Importantly, inhibition of overall binding seen by HLA-DR staining (Fig. 4B) reflected largely a decrease in the CD4-dependent component of binding as the anti-Leu3a MAb signal decreased only slightly when CV-N-treated virions were added to the cells (Fig. 4A) (13% decrease compared to 37%). Consistent with this interpretation, binding of CV-N-treated virions to anti-Leu3a-pretreated cells was comparable to binding of native virions to anti-Leu3a MAb-pretreated cells (Fig. 4B). The lack of an additive inhibitory effect suggests that CV-N treatment of virions did not inhibit CD4-independent binding of virions to cells.

FIG. 4.

Binding of HIV and CV-N-treated virions to A3.01 cells. A3.01 (CD4-positive) cells were either mock treated or treated with 10 μg of anti-Leu3a MAb per ml for 30 min at 4°C and washed twice before addition of HIV-1_MN, which was either mock treated or treated with CV-N (200 nM). The virions were added to A3.01 cells, and binding was assessed by quantitating virion-associated HLA-DR signal by flow cytometry, where HLA-DR acquisition indicates overall virion binding, and by Leu3a staining, where loss of availability of the Leu3a epitope is an indirect indication of CD4-dependent virion binding. Numbers in figure keys are MFI values. (A) CD4-Leu3a signal (gp120 binding epitope on CD4). A3.01 cells express the Leu3a epitope on CD4 (black trace), and saturating unlabeled anti-Leu3a MAb pretreatment blocks binding of fluorescein isothiocyanate-labeled anti-Leu3a MAb (compare blue and black traces). HIV binding blocks the Leu3a epitope (compare black and green traces). Cells with bound CV-N-treated virions have greater availability of Leu3a epitopes than do cells with bound untreated virions (compare red and green traces). (B) HLA-DR signal. A3.01 cells are HLA-DR negative (black trace). Addition of untreated HIV results in acquisition of HLA-DR staining (compare black and green traces). Approximately 56% of virion binding is CD4 dependent (compare green and blue traces). CV-N treatment of virions inhibits overall virion binding (compare green and red traces). CV-N inhibits CD4-dependent binding of virions; note the lack of increased inhibition of virion binding for CV-N-treated virions on unlabeled anti-Leu3a MAb-pretreated cells (compare blue and orange traces). At least 10,000 events were acquired for each sample.

FIG. 5.

CV-N concentration-dependent effects on virion binding (A) and infectivity (B) and effects of MAb 2G12 and CV-N on CD4-dependent binding (C). A3.01 (CD4-positive) cells were either mock treated with PBS or treated with 10 μg of unlabeled anti-Leu3a MAb per ml for 30 min at 4°C and washed twice before addition of HIV-1_MN (17,000 ng of p24 per ml), which was either mock treated or treated with CV-N at various concentrations ranging from 0.02 to 2000 nM. (A) Percent CD4-dependent virion binding (anti-HLA-DR [—▴—]) was calculated by subtracting the MFI for CD4-independent binding (virion binding in the presence of the anti-Leu3a MAb) from total binding with 100% binding equal to binding in the absence of CV-N. MFI values for untreated virions: total binding to untreated cells, 286; binding to unlabeled anti-Leu3a MAb-pretreated cells, 127. Percent Leu3a epitope availability (—●—) and percent OKT4 epitope availability (—■—) were calculated by dividing the anti-Leu3a or anti-OKT4 signal in the presence of untreated virions by the anti-Leu3a MAb or anti-OKT4 MAb signal on untreated cells. MFI values for untreated cells: anti-Leu3a, 19.5; anti-OKT4, 33.1; for HIV-1 bound to cells, anti-Leu3a, 12.7; anti-OKT4, 30. (B) Parallel samples used for the binding assay shown in panel A were used in a quantitative PCR-based viral entry assay. ∗, no gag DNA detected (<30 copies per reaction). PBGD, porphobilinogen deaminase. (C) Effect of MAb 2G12 (50 μg/ml) alone or with CV-N (200 nM) on CD4-dependent HIV-1_MN virion binding. Anti-HLA-DR MAb MFI values for virions (untreated cells and anti-Leu3a-treated cells, respectively): HIV-1-bound cells, 318 and 106 (67% CD4-dependent binding); CV-N-treated virions, (219 and 163; 2G12-treated virions, 222 and 102; CV-N–2G12-treated virions, 252 and 180. Percent CD4-dependent binding was calculated as for panel A. Each data point represents at least 10,000 acquired events.

This is perhaps better appreciated by examining Fig. 5A, which graphically summarizes results for CD4-dependent virion binding based on measurements of mean fluorescence intensity (MFI) or overall brightness of staining for positive cells in the experiments described above in the absence or presence of unlabeled, anti-Leu3a MAb pretreatment of the A3.01 cells. Treatment of virions with greater than 200 nM CV-N, concentrations producing essentially complete inactivation of viral infectivity in the high-input virus assay systems described here, inhibited acquisition of HLA-DR signal, reflective of CV-N blocking the binding of virions to the target cells (Fig. 5A). This was paralleled by an increase in the availability of the Leu3a epitope, reflecting a decrease in the blockade of the epitope associated with CD4-dependent binding of virions to target cells (Fig. 5A). Treatment of HIV-1 virions with virucidal concentrations of CV-N thus appears to block CD4-dependent binding of virions to target cells. The ability of CV-N to prevent loss of the Leu3A epitope upon exposure of CD4⁺ cells to treated virions and the observation that the anti-fusion peptide DP178 (17) did not decrease the HLA-DR signal observed in our binding assay (21a) indicate that the HLA-DR signal is dependent on virion binding without a requirement for fusion and that CV-N blocks this CD4-dependent binding. Neither binding of native virions nor CV-N-treated virions interfered with detection of the OKT4 epitope on A3.01 cells (no concentration-dependent inhibition [Fig. 5A]), revealing that bound virions were not masking this epitope or inducing CD4 degradation or endocytosis.

Because the concentrations of CV-N required to fully neutralize HIV-1_MN in the present assay systems were almost 100-fold higher than those previously reported in other assays (10), we performed parallel infectivity experiments to determine if the reduction in CD4-dependent binding correlated with inhibition of infectivity (Fig. 5B). Parallel virus samples from Fig. 5A were used in an infectivity assay to determine the concentrations of CV-N required to neutralize HIV-1_MN. Using the quantitative PCR infectivity assay described above, we observed that concentrations of CV-N required to neutralize the virus in the present experiments (Fig. 5B) coincided with those associated with inhibition of CD4-dependent virion binding (Fig. 5A). The greater concentrations required to block infectivity were most likely due to the fact that, to obtain detectable signals in the virus binding assay, we were using concentrations of input virus several logs higher than those typically used in infectivity assays. Nevertheless, there was a clear correlation between the concentrations of CV-N required to block infection and concentrations associated with inhibition of CD4-dependent virion binding.

In view of results from immunoprecipitation and ELISA format assays suggesting possible similarities between the interaction sites on virions for CV-N and the MAb 2G12 (see below), we also evaluated the effect of pretreatment of virions with MAb 2G12 on virion binding. Both CV-N-treated (200 nM) and MAb 2G12-treated (50 μg/ml) virions showed impaired binding to target cells compared to that for untreated virions (Fig. 5C). Simultaneous pretreatment of virions with both CV-N and MAb 2G12 resulted in no incremental inhibition of virion binding over that seen for virions pretreated with CV-N alone (Fig. 5C).

CV-N occludes or alters a distinct neutralizing epitope on gp120.

To better understand how CV-N affected the integrity of viral and cellular proteins on the surface of virions, we examined the expression of several gp120 and cell-derived protein epitopes in a whole-virion immunoprecipitation assay (3), comparing native and CV-N-treated virions. A panel of both MAbs and polyclonal Abs reactive with host cell-derived and viral proteins on the virion surface were tested, including some MAbs to defined conformationally sensitive epitopes on gp120, in an effort to map the site(s) on the virions where CV-N was interacting.

CV-N treatment of virions did not interfere with precipitation by a polyclonal antiserum that recognizes class II MHC (HLA-DR) or class I MHC or precipitation by an anti-H9 serum that recognizes only partially characterized cellular proteins on the surface of the virion (Fig. 6A). CV-N treatment of virions also had negligible effects on the ability of the MAb F105 to recognize a discontinuous epitope near the CD4bs and on the ability of the neutralizing MAb IgG1b12 to bind to the CD4bs on gp120 (Fig. 6B). In addition CV-N did not significantly inhibit binding of a monospecific antiserum (W0/07) which recognizes a linear epitope in the V3 domain of gp120 (Fig. 6B).

FIG. 6.

Whole-virion immunoprecipitation of CV-N-treated virions. HIV-1_MN (A, B, C, E, and F) or HIV_IIIB (D) was either mock treated or treated with CV-N for 90 min at 4°C and precipitated with MAb or antisera to virion surface proteins. Values shown are means ± standard deviations for triplicate measurements in one experiment, representative of three independent experiments with similar results. (A) Polyclonal Ab to host cell-derived virion surface proteins, including MHC class I, MHC class II, and an antiserum raised against microvesicle preparations derived from H9 cells (7). (B) gp120 with IgG1b12 (CD4bs), W0/07 (V3), or F105 (CD4bs). (C) CD4i conformational epitope 48.d. After virions were pretreated with CV-N, sCD4 or 1% BSA was added, and the virions were precipitated with the 48.d MAb. (D) 2G12 MAb distinct neutralizing epitope on HIV-1_MN and HIV-1_IIIB. (E and F) Parallel samples from panels A to C (HIV-1_MN) were used in a quantitative PCR-based viral entry assay (E) (∗, no gag DNA detected [<30 copies per reaction]) and a p24^CA infectivity ELISA (F). Neg. Cont., negative control. PBGD, porphobilinogen deaminase.

We also examined whether CV-N treatment of virions prevented sCD4 binding-induced conformational changes in gp120. As described previously (62), pretreatment of virions with sCD4 resulted in increased exposure of the epitope recognized by the MAb 48.d (Fig. 6C), the result of a conformational change in gp120 triggered by sCD4 binding. Virions were pretreated with CV-N, prior to sCD4 exposure and immunoprecipitation with MAb 48.d. As shown in Fig. 6C, both expression of the 48.d epitope and the induction of increased expression of this epitope as a consequence of conformational changes triggered by sCD4 binding were maintained in CV-N-treated virions, even at the highest concentration of CV-N tested (2,000 nM). Thus, CV-N treatment did not affect sCD4-triggered conformational changes in gp120, even at concentrations well in excess of those required to neutralize HIV. CV-N treatment also had no effect on the ability of the 2F5 MAb to recognize a neutralizing epitope on gp41 (data not shown).

Lastly, we examined the effect of CV-N treatment of virions on the unique glycosylation-dependent, neutralizing epitope recognized by MAb 2G12 (28, 64, 70). Treatment with neutralizing concentrations of CV-N (Fig. 6E and F) potently interfered with the ability of the MAb 2G12 to immunoprecipitate HIV-1_MN virions (Fig. 6D), indicating that the CV-N treatment of virions occludes or otherwise alters this epitope, perhaps by steric inhibition due to binding at or near the same site recognized by MAb 2G12. To extend this result, we tested the highly glycosylated isolate HIV-1_IIIB/LAI, which also expresses the conserved epitope recognized by MAb 2G12. Precipitation of the HIV-1_IIIB/LAI virus by MAb 2G12 was inhibited to an even greater extent by CV-N treatment of the virions than was precipitation of HIV-1_MN virions (Fig. 6D). At the highest concentrations, CV-N reproducibly increased the magnitude of virion precipitation. This unexplained phenomenon may have been due to intervirion cross-linking by CV-N or CV-N binding nonspecifically to the 2G12 MAb.

We also evaluated the effect of CV-N on the reactivity of a panel of MAbs with monomeric gp120 in an ELISA format assay. As shown in Fig. 7A to F, these studies demonstrated that CV-N interfered strongly with recognition of monomeric gp120 by the MAb 2G12 but minimally with recognition of gp120 by MAbs that recognize the V2 loop, the V3 loop, the C4 region, the CD4bs, or CD4-induced epitopes (Fig. 7A to E). Additional studies confirmed that CV-N pretreatment of gp120 blocked subsequent binding of MAb 2G12 to gp120 (Fig. 7G) and demonstrated that MAb 2G12 pretreatment of gp120 did not block subsequent binding of CV-N to gp120 (Fig. 7H).

FIG. 7.

Effect of CV-N on binding of anti-gp120 MAbs to monomeric gp120: CV-N blocks 2G12 binding, but 2G12 does not block CV-N binding. HIV_IIIB monomeric gp120 was applied as a coating to ELISA plates, and the ability of CV-N to compete with a panel of MAbs was examined. (A) V2 loop Abs; (B) V3 loop Abs; (C) C4 region Abs; (D) CD4bs Ab; (E) CD4i epitope Abs; (F) 2G12 MAb; (G and H) captured gp120 was pretreated with CV-N prior to incubation with 2G12 and detection of bound MAb (G) or pretreated with 2G12 prior to incubation with CV-N and detection of bound CV-N (H). O.D., optical density. The results for panels G and H were obtained by using the second assay protocol described under “ELISA studies” in Materials and Methods.

DISCUSSION

The potent inactivating activity of CV-N against a broad range of HIV isolates has been well established, and the compound is promising as a potential topical microbicide for the prevention of sexual transmission of AIDS. The present studies, consistent with earlier observations (10), show that CV-N acts at the level of the virion, but not the target cell (Fig. 1A). Pretreatment of virions eliminated HIV-1 infectivity, while real-time, quantitative PCR studies demonstrated that infectivity was blocked at a step in the viral life cycle prior to reverse transcription (Fig. 1B). A virus-induced fusion-from-without assay showed that CV-N-treated virions did not mediate this CD4-dependent fusion process (Fig. 2C) and that CV-N blocked Env-mediated cell fusion (Fig. 3). A flow cytometric, whole-particle, virion binding assay showed that CV-N treatment of virions inhibited CD4-dependent binding required for productive infection.

Studies in multiple assay formats with MAbs to defined epitopes on the HIV envelope glycoprotein suggested a structural basis for the interaction of CV-N with HIV virions and the resulting inhibition of infectivity (Fig. 6 and 7). CV-N treatment of virions did not disrupt the epitopes recognized by MAbs reactive with the V3 region or the CD4bs epitope recognized by the MAb IgG1b12, consistent with previous observations in which CV-N did not block anti-gp120 MAb binding to CV-N-treated recombinant monomer gp120 in an ELISA (10). However, results in both whole-particle immunoprecipitation and ELISA format assays consistently demonstrated that CV-N treatment interfered with MAb binding to the epitope recognized by the neutralizing MAb 2G12 (Fig. 6D and 7F). This MAb reacts with an only partially characterized, glycosylation-dependent epitope comprised of elements from the C2, C3, C4, and V4 domains of gp120 (64). In the recently solved crystal structure of the Ab-complexed gp120 core of the HIV-1_IIIB (HXB2c) laboratory isolate of HIV-1, a large fraction of the predicted accessible surface of gp120 in the trimer is composed of variable, heavily glycosylated core and loop structures that surround the CD4bs and the coreceptor binding regions (28, 70, 71). The 2G12 epitope overlies the stem of the V3 loop and the V4 variable regions and is characterized by high-mannose sugars (70). It is noteworthy that the 2G12 MAb is unique among Abs characterized to date with respect to recognizing this epitope, despite the fact that the epitope is conserved across viral isolates. In addition, the gp120 crystal structure indicates that there is a large, heavily glycosylated, immunologically silent domain (28, 70, 71). It is possible that, in addition to binding at or near the 2G12 epitope, CV-N might bind other glycosylated domains in this immunologically silent region.

The highly conserved 2G12 epitope is located on a relatively variable surface on the gp120 outer domain, opposite and approximately 25 Å away from the CD4bs (70). This may help to explain why sCD4 is able to bind to the CD4bs on gp120 and induce conformational changes in virions treated with CV-N (Fig. 6) or 2G12 (64). However, it was initially somewhat hard to envision how sCD4 could bind to CV-N-treated virions (Fig. 6C), while CD4-dependent binding of CV-N-treated virions to target cells was impaired. One possibility is that CV-N may form aggregates on the virions that sterically hinder virion binding to target cells but still allow binding to sCD4. The steric context may be quite different for virion gp120 binding to sCD4 compared to binding to membrane-bound CD4 (50).

The crystal structure predicts that the 2G12 epitope is oriented toward the target cell upon CD4 binding (28, 70), suggesting that the 2G12 MAb and CV-N may sterically impair interactions of the oligomeric envelope glycoprotein complex with CD4 or other host cell moieties, such as chemokine receptors used as retroviral coreceptors. Indeed, it is possible that a portion of the HLA-DR signal measured as an index of virion binding in our flow cytometric assay (Fig. 4 and 5) reflects virions stabilized by interactions between the chemokine receptor binding domain on gp120 and the cognate chemokine receptor expressed on the target cell, in addition to gp120 binding to CD4. Since optimal exposure of the chemokine receptor binding domain on gp120 is dependent on prior conformational changes induced in gp120 by binding to CD4, such stabilized binding observed in the virion binding assay would be dependent on gp120-CD4 interactions. If CV-N bound to gp120 interferes directly or sterically with interactions between the chemokine receptor binding domain on gp120 and the receptor, then CV-N-treated virions might show impaired CD4-dependent binding in our flow cytometric assay (Fig. 4 and 5), in spite of CV-N not directly blocking gp120-sCD4 binding or postbinding-induced conformational changes (Fig. 6 and 7). Consistent with this interpretation is the observation that CV-N can potently inactivate the feline immunodeficiency virus (18a), a virus which is entirely CD4 independent but which can utilize the CXCR4 chemokine receptor (19). This model would predict that CD4-independent strains of HIV and SIV (21, 41), in which the chemokine receptor binding domain of gp120 is already sufficiently exposed to obviate the need for initial gp120-CD4 binding-induced conformational changes, should be susceptible to inhibition by CV-N.

This interpretation is also consistent with studies performed with the 2G12 MAb showing that the 2G12 epitope comprises a unique competition group (64) that does not interfere with the binding of monomeric gp120 to either CD4 or CCR5 (63) but does block HIV-1 virion binding to CD4⁺ T cells (66). Our own results, in which both CV-N and MAb 2G12 inhibited CD4-dependent binding of virions to target cells (Fig. 5D) while CV-N did not interfere with interactions between virion-associated gp120 and sCD4 (Fig. 5 and 6), extend these findings. The lack of incremental blockade of CD4-dependent binding to target cells of virions treated with both CV-N and MAb 2G12, compared to virions treated with CV-N alone (Fig. 5D), suggests that the compounds may interact with virions in similar ways. If CV-N were interacting with gp120 in a manner similar to that of the glycosylation-dependent MAb 2G12, this would also be in agreement with the finding that CV-N binds less well to nonglycosylated recombinant gp120 than to native gp120 (10). The fact that the 2G12 MAb reacted less well against the HIV-1_MN isolate in the viral precipitation assay than against the highly glycosylated HIV-1_IIIB/LAI would be predicted since MAb 2G12 was the least active against the HIV-1_MN isolate of the many primary and tissue-culture-adapted viruses tested (64). Interestingly, CV-N somewhat less potently inactivated HIV-1_MN than it did HIV-1_IIIB/LAI, which was more readily neutralized by CV-N (10) as well as by 2G12 (64).

However, reciprocal cross-blocking studies in which CV-N pretreatment prevented subsequent MAb 2G12 binding to gp120 but MAb 2G12 pretreatment did not prevent subsequent CV-N binding to gp120 underscore the fact that the two agents do not act identically. This observation is formally consistent with the possibility that CV-N and MAb 2G12 bind to essentially the same site, with CV-N having a much higher affinity. Indeed, initial studies indicated very tight binding of CV-N to gp120 (10). However, the nonreciprocal blocking of CV-N by MAb 2G12 may also reflect the possibility that CV-N binds to multiple CHO-associated sites on gp120, in a manner that renders the 2G12 epitope inaccessible to the Ab. Indeed, preliminary studies suggest that more than one CV-N molecule can bind to a single gp120. Other evidence suggests that, at high concentrations, CV-N may cause aggregation of gp120, which may underlie an increase in CD4-independent binding of CV-N-treated virions seen in the flow cytometric binding assay (21a). This may reflect multiple modes of CV-N interaction with virions, perhaps with differing affinities. This factor, and the effects of CV-N on both total and CD4-dependent binding, will be important in interpreting the effects of CV-N in different assay systems (40). Future structural studies of CV-N complexed with gp120 will hopefully define the binding stoichiometry of CV-N and gp120 and reveal the precise molecular mechanism of CV-N interaction with gp120.

Many reports have documented the importance of glycosylation for the infectivity and pathogenesis of HIV and SIV (5, 9, 32, 34, 53, 59, 60). There are approximately 24 potential sites for N-linked glycosylation on gp120, and carbohydrate constitutes approximately 50% of the mass of gp120 (33). Several groups have proposed that these carbohydrate residues act as a shield to protect the virion from the humoral immune response (12, 44, 53, 71). Because only a few human Abs are able to neutralize primary HIV isolates and many of these work only in a type-restricted manner, identifying ways to elicit Abs to the carbohydrate domains on the virion or identifying compounds that interfere with these epitopes might provide effective means to inhibit infection. The 2G12 epitope is unique in that it is highly conserved across many HIV-1 isolates and Abs to this site neutralize the virus (64). Identifying compounds such as CV-N that avidly bind or otherwise block this site on the virion may provide new strategies to prevent and treat HIV infection and AIDS. In previous studies, CV-N did not block infectivity of other enveloped nonlentivirus viruses including herpesvirus 1, cytomegalovirus, and adenovirus type 5 (10). However, if the mechanism of blockade of HIV and SIV infection involves interactions with viral envelope glycoprotein carbohydrate, it is possible that the CV-N may be active against other viruses with envelopes having significant CHO content. Such activity might broaden CV-N’s potential clinical utility.