INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

Infection with human immunodeficiency virus-1 (HIV-1),¹ which progresses to AIDS, represents a disease pandemic that cannot yet be controlled by vaccination (1). Because the pathology associated with HIV infection is not fully understood, there are concerns that vaccines made from attenuated strains of HIV could be unsafe, whereas subunit vaccines might not be complex enough to generate an appropriate immune response. Further, it is unlikely that immunity produced against one strain of HIV-1 will be effective in providing large scale protection against such a mutable virus. Studies of infected individuals termed nonprogressors have indicated that a combined response of neutralizing antibody and reactive cytotoxic T-cells can retard the onset of AIDS (2). Therefore, the development of a stable vaccine, which could elicit both humoral and cellular responses and be flexible enough to incorporate sequences from many HIV-1 strains, is desirable. Current vaccine approaches range from the development of recombinant viruses (3, 4) through the use of purified envelope proteins (5) to the evaluation of conjugates composed of viral peptides (6-8). Despite much effort, a consensus vaccine candidate has not emerged.

Here we characterize a series of recombinant V3 loop-toxin chimeric proteins. The V3 loop was chosen because it represents the major neutralizing epitope of HIV-1. Pseudomonas exotoxin (PE) was chosen because it has been used as an effective carrier-adjuvant in vaccines directed against bacterial pathogens (9, 10) and because its functional domains have been well characterized (11). Because of the latter, it was possible to identify a location where the introduction of a V3 loop insert was unlikely to cause a major disruption in toxin structure.

Previous work has shown that the third variable (V3) loop of the envelope protein, gp120, contains the principal neutralizing domain of HIV-1 (12-14). Further, immunization with a recombinant form of gp120 appears sufficient to protect chimpanzees from infection by HIV-1 challenge (15). Thus, gp120 and its parent molecule, gp160, have been used to vaccinate human volunteers (16, 17). Because of the importance of the envelope protein and specifically its V3 loop for HIV infectivity, our efforts have focused on the development of chimeras composed of strain-specific V3 loop sequences. PE has a nonessential subdomain, termed Ib, which is composed of a small loop flanked by a disulfide bond. Chimeras were designed to introduce V3 loop amino acids in this location where they could form a loop and be exposed at the surface of the protein. Although V3 loops vary considerably among the various HIV-1 strains (18), specific antibodies to this region have been shown to neutralize infectivity of the virus and to prevent viral cell fusion in vitro (6). Because our vector allows for the introduction of many different sequences, we anticipate that a chimera displaying the V3 loop of any isolate could be produced.

PE is secreted by Pseudomonas aeruginosa as a 67-kDa protein composed of three prominent globular domains (Ia, II, and III) and one small subdomain (Ib) connecting domains II and III (19). Domain Ia of PE binds to the low density lipoprotein receptor-related protein (LRP) (20, 21), also known as the ₂- macroglobulin receptor. LRP is expressed on the surface of most mammalian cells and tissues, including those of the immune system (22). Domain II mediates translocation to the cytosol, and domain III has ADP-ribosylating activity (23). Once bound to LRP, the toxin traffics via coated pits to an endosomal compartment, where it is cleaved by the protease, furin, to generate a 37-kDa C-terminal fragment composed of domains II, Ib, and III (24, 25). This fragment translocates to the cell cytosol, ADP-ribosylates elongation factor 2, and shuts down protein synthesis. When glutamic acid 553 in domain III is deleted (E553), the toxin still gains access to the cytosol (26) but is rendered nontoxic since this mutation eliminates ADP-ribosylating activity (27). In the construction of our chimeras, we deleted much of subdomain Ib from the (E553) version of PE and substituted V3 loop sequences of various sizes in its place. The resulting chimeras were characterized structurally, evaluated for display of V3 loop sequences, and tested as immunogens for the generation of antiviral responses.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Plasmid Construction-- Plasmid pMOA1A2VK352 (24), encoding PE, was digested with SfiI and ApaI (residues 1143 and 1275, respectively) and then re-ligated with a duplex containing a novel PstI site. The coding strand of the duplex had the following sequence: 5'-TGGCCCTGACCCTGGCCGCCGCCGAGAGCGAGCGCTTCGTCCGGCAGGGCACCGGCAACGACGAGGCCGGCGCGGCAAACCTGCAGGGCC-3'. The resulting plasmid encoded a slightly smaller version of PE that lacked much of domain Ib. The PstI site was subsequently used to introduce duplexes encoding V3 loop sequences flanked by cysteine residues. To make nontoxic proteins, vectors were modified by subcloning to introduce the enzymatically inactive domain III from pVC45E553. An additional subcloning, from pJH4 (23), was needed to produce a vector that lacked a signal sequence. Insertion of duplexes and subcloning modifications were initially verified by restriction analysis, and final constructs were confirmed by dideoxy double strand sequencing.

Chimera Protein Expression and Purification-- All ntPE-V3 loop chimeric proteins were expressed in Escherichia coli SA2821/BL21(DE3) using a T7promoter/T7 polymerase system (28). SA2821/BL21(DE3) cells were transformed with the appropriate plasmid and grown to an absorbance of 1.0 (600 nm) in medium containing ampicillin. To induce high level protein expression, isopropyl--D-thiogalactoside (1 mM) was added to the culture and incubated for an additional 90 min. E. coli cell pellets were resuspended in 50 mM Tris, 20 mM EDTA, pH 8.0 (TE buffer) and dispersed using a Tissue Miser. Cell lysis was accomplished with lysozyme (200 µg/ml; Sigma). Membrane-associated proteins were removed by washing with 2.5% Triton X-100 and 0.5 M NaCl.

Biochemical Characterization-- Chimeric proteins were separated by SDS-PAGE using 8-16% gradient polyacrylamide gels (Novex, San Diego, CA), and visualized by staining with Coomassie Blue. For Western blot analysis, proteins were transferred onto Immobilon-P membranes (Millipore Corp, Bedford, MA) and exposed to either an anti-PE mouse monoclonal antibody (M40-1; Ref. 29) or an anti-gp120 mouse monoclonal antibody (1F12 for MN sequences or 1B2 for Thai-E sequences; Genentech, Inc., South San Francisco, CA). The primary antibody was detected by a secondary anti-mouse antibody conjugated to horseradish peroxidase. Reactive products were visualized by the addition of diaminobenzadine and hydrogen peroxide. Immunocapture experiments were performed for 30 min at 23 °C using anti-gp120 monoclonal antibodies. Antibody-chimeric protein complexes were recovered with protein G-Sepharose beads (Amersham Pharmacia Biotech) and separated using SDS-PAGE (as above). Recombinant forms of gp120 derived from HIV-1-MN (gp120/MN, subtype B; Genentech, Inc.) and a Thai subtype E isolate (gp120/Thai-E-Chiang Mai; Advanced Biotechnologies, Columbia, MD) were used as standards.

Biophysical Characterization-- Circular dichroism (CD) spectra were collected on an Aviv 60DS spectropolarimeter. Near-UV CD spectra (400 nm to 250 nm) were obtained in 0.2-nm increments with a 0.5-nm bandwidth and a 5-s time constant (150 readings/s averaged) for samples in a 1-cm path length cell. Far-UV spectra (250 nm to 190 nm) were collected in 0.2-nm increments with a 0.5-nm bandwidth and a 3-s time constant in a 0.05-cm path length cell. Each spectrum was digitally smoothed using the Savitsky-Golay algorithm (30), corrected for concentration, and normalized to units of mean residue weight ellipticity (_MRW) using the following relationship.

(Eq. 1)

_obs is the observed ellipticity, M_r(monomer) is the molecular weight of the monomer, n_monomer is the number of amino acids in the monomer, d is the path length of the cell (cm), and c is the concentration of the sample in the cell (mg/ml).

Cell-based Cytotoxicity Assays-- Human A431 (epidermoid carcinoma) cells were seeded in 24-well tissue culture plates at 1 × 10⁵ cells/well in RPMI 1640 medium supplemented with 5% fetal bovine serum. After 24 h, cells were treated for 18 h at 37 °C with 4-fold dilutions of either wtPE or toxic forms (with a glutamic acid residue at position 553 and capable of ADP-ribosylating elongation factor 2) of the chimeric proteins. Inhibition of protein synthesis was assessed by monitoring the incorporation of [³H]leucine.

Immunizations-- Rabbits were immunized subcutaneously at four sites with 200 µg (total) of each ntPE-V3 chimeric protein. The first injection was administered with complete Freund's adjuvant. All subsequent injections (at 2, 4, and 12 weeks) were given with incomplete Freund's adjuvant. Venous bleeds were obtained weekly after the third injection and screened by immunoblotting against gp120.

ELISA-- Ninety-six-well plates (Pierce) were coated with gp120 (100 ng/well) derived from either an MN strain or a Thai strain (gp120/Chiang Mai from ABI; see above). Gp120 in PBS was added for 2 h at room temperature followed by a blocking solution of 1% bovine serum albumin and 0.2% Tween 20. Subsequent washes were with 0.2% Tween 20 in PBS (TPBS). The appropriate dilution of primary antibody in TPBS was added to each well for 30-60 min at room temperature. To detect the primary antibody, donkey anti-rabbit IgG conjugated to horseradish peroxidase (Amersham Pharmacia Biotech) diluted 1:1000 in TPBS was added for 30 min. Bound peroxidase was visualized using 3,3',5,5'-tetramethyl benzidine (Pierce) mixed with 0.01% hydrogen peroxide. H₂SO₄ was added to quench the reaction and allow absorbance to be measured at 450 nm. Prebleed sera were used to determine nonspecific binding in the assay. To determine specific binding, absorbance with a prebleed serum sample was subtracted from the corresponding value obtained with immune sera.

Viral Neutralization Assay-- One assay utilized MT4 cells as a indicator of HIV-1-mediated cell death (31). Duplicate serial dilutions of antiserum were incubated with HIV-1/MN (provided by M. Norcross, Food and Drug Administration and grown in FDA/H9 cells (32)) and the mixture added to MT4 cells for 7 days. Viral-mediated cell death was assessed using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide dye assay (33) and spectrophotometric analysis at 570 nm. The serum 50% inhibitory concentration was calculated and reported as the neutralization titer. A second assay used p24 production of as an indicator of viral growth (34). Primary virus was first titrated to determine the amount that reproducibly yielded significant but submaximal amounts of p24. Virus preparations were incubated for 1 h at 37 °C with various dilutions of rabbit sera, either immune or prebleed, and this mixture was then added in quadruplicate to 2.5 × 10⁵ PBMCs. The culture continued for 3 days at which time cells were washed and resuspended in medium containing interleukin 2. Accumulation of p24 was detected by an ELISA.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Construction of Chimeras-- Wild-type (wt) PE is composed of 613 amino acids and has a molecular mass of 67,122 Da. Deletion of a glutamic acid 553 (E553) results in a nontoxic version of PE (27), referred to as ntPE (Fig. 1A). To generate chimeric proteins, we replaced much of subdomain Ib with V3 loop sequences (Fig. 1B) from either an MN (subtype B) or Thai (subtype E) strain of HIV-1. The MN sequence is from a T-cell-tropic strain, and the Thai-E sequence comes from a macrophage-tropic strain. Plasmids were constructed by inserting oligonucleotide duplexes encoding V3 loop sequences into a PE-based vector that was designed with a novel PstI site (see "Experimental Procedures"). In an effort to produce a V3 loop of similar topology to that found in gp120, the 14- or 26-amino acid inserts were flanked by cysteine residues (Fig. 1C, bold type). Construction of the vector resulted in several changes in the amino acid sequence of ntPE near the insertion point of the V3 loop (Fig. 1C, italics). Insertion of an irrelevant 16-amino acid sequence resulted in the construction of a control chimera referred to as ntPE-FP16. Removal of the Ib loop (6 amino acids) and modification of flanking amino acids adjacent to the V3 loop insert resulted in a small increase in molecular mass compared with wild-type PE (Fig. 1C).

View larger version (44K): [in this window] [in a new window]   Fig. 1.   A, a schematic depiction of PE showing its major structural domains: Ia, II, and III. Also shown is the minor domain, Ib, and the approximate location of the single amino acid deletion (E553) to ablate PE toxicity. B, a PE-V3 loop chimera showing where domain Ib has been replaced by the V3 loop. C, amino acid sequences, represented with single-letter code, which replaced the Ib loop of wild-type PE with a V3 loop sequence of gp120 (bold type) from either the MN-B or Thai-E (Th-E) strains of HIV-1. The insert was bordered by two cysteine residues to allow for the formation of a disulfide bond at the base of the loop. The insertion of a unique PstI restriction site into the toxin vector resulted in several modifications to the sequence of wild-type PE (italics). An irrelevant insert was prepared as a control and is designated ntPE-FP16. Calculated molecular masses are shown for full-length expressed proteins.

Characterization of Chimeras-- SDS-PAGE analysis (Fig. 2A) of purified ntPE-V3 loop chimeras was consistent with calculated masses (Fig. 1C). Western blots, using monoclonal antibodies raised against gp120/MN (1F12) or gp120/Thai-E (1B2), showed strain-specific reactivity with the MN- and Thai-E-V3 loop chimeras (Fig. 2B). Free sulfhydryl analysis of purified ntPE-V3 loop chimeras failed to demonstrate any unpaired cysteines (data not shown), suggesting that the purified ntPE-V3 loop chimeras had refolded and oxidized to form a novel disulfide bond at the base of the V3 loop (Fig. 1B). We anticipated that the formation of this disulfide bond would result in the exposure of the V3 loop at the surface of the chimeras. This was tested directly by immunocapture studies (Fig. 2C). The 1F12 and 1B2 monoclonal antibodies selectively captured the soluble MN and Th-E chimeric proteins confirming that the V3 loops were accessible to antibody probes. Despite the fact that the 1F12 antibody reacted strongly with ntPE-V3MN14 in Western blots (Fig. 2B), it captured only a small amount of soluble protein (Fig. 2C, lane 3), suggesting that the reactive epitope was not completely accessible when only 14 amino acids were inserted.

View larger version (58K): [in this window] [in a new window]   Fig. 2.   Characterization of ntPE-V3 loop chimeras after separation by SDS-PAGE. A, Coomassie Blue staining of purified ntPE-V3 loop chimeras following separation by SDS-PAGE. Approximately 1 µg of protein was loaded on each lane. B, Western blot analysis of ntPE-V3 loop chimeras. After transfer to Immobilon P membranes, proteins were probed with monoclonal antibodies raised against intact gp120/MN (1F12) or gp120/Thai-E (1B2). C, immunocapture studies, using either 1F12 or 1B2 immobilized on protein G-Sepharose, were used to characterize the exposure of V3 loop sequences on the surface of the various chimeric proteins. Proteins were visualized by staining gels with Coomassie Blue. gp120 and FP16 (with an irrelevant insert) were used as positive and negative controls, respectively. The capture of PE-V3 loop proteins is indicated with a single arrowhead and of gp120 by a double arrowhead. The left panel shows the presence of the antibody heavy chain (hc) only since the light chain (lc) was run off the gel. The right panel shows both chains.

View larger version (30K): [in this window] [in a new window]   Fig. 3.   V3 loop amino acid sequence insertions do not significantly alter the secondary structure of wild-type PE. Near-UV (A) and far-UV (B) CD spectra (mean of three scans following background spectrum subtraction) were digitally smoothed, corrected for concentration, and normalized to units of mean residue weight ellipticity. C, secondary structure calculations were performed using the SELCON fitting program. *, calculated -helix content agrees with values determined from changes in observed ellipticity at 222 nm.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   An enzymatically active PE-V3 loop chimera is toxic for cells. Inhibition of protein synthesis, assessed by [3H]leucine incorporation, was determined in human A431 cells following an 18-h exposure to various concentrations of either wild-type PE or a toxic form (with a glutamic acid residue at position 553 and capable of ADP-ribosylating elongation factor 2) of PE-V3MN26.

Immunization of Rabbits with Chimeras-- To investigate the immune response to the chimeras, rabbits were injected subcutaneously with 200 µg of either the MN or Thai-E version of the protein. In Western blots, serum samples from rabbits immunized with the ntPE-V3MN proteins exhibited a strong reactivity for immobilized recombinant gp120/MN (Fig. 5A). Reactive titers increased with time. At 6 weeks, reactivity was noted at 1:200 dilution; at 12 weeks, reactivity was noted at 1:5,000 dilution; and at later times, reactivity could be detected at 1:25,000. Anti-V3 loop/MN sera were either less reactive or not reactive with gp120/Thai-E (Fig. 5A). Rabbits injected with the ntPE-V3ThE produced reactive sera for gp120/Thai-E with little or no reactivity for gp120/MN (Fig. 5A). Sera from rabbits injected with nontoxic PE (i.e. ntPE with no insert) exhibited no reactivity for gp120 of either subtype (data not shown).

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Characterization of rabbit sera following immunization with either ntPE-V3MN26 or ntPE-V3ThE26. A, Western blot reactivity of rabbit antisera diluted 1:1000 for recombinant gp120/MN and gp120/Thai-E was assessed following SDS-PAGE and the transfer of proteins to Immobilon P membranes. Reactive primary antibody was detected by a secondary anti-rabbit antibody conjugated to horseradish peroxidase. B, sera obtained from a rabbit injected with ntPE-V3MN26 were diluted 1:10,000 and then incubated with various concentrations of either soluble gp120/MN or ntPE-V3MN26. The residual reactivity for immobilized gp120/MN is shown. Dotted column, no competitor; shaded column, gp120/MN; striped column, ntPE-V3MN26. C, sera obtained from a rabbit injected with ntPE-V3ThE26 were diluted 1:10,000 and then incubated with various concentrations of either soluble gp120/Thai-E or ntPE-V3ThE26. The residual reactivity for immobilized gp120/Thai-E is shown. Gp120/MN and ntPE-V3MN26 (at 20 nM) were also tested as potential competitors. Dotted column, no competitor; bold diagonal striped column, gp120.Thai-E; horizontal striped column, ntPE-V3ThE26; shaded column, gp120/MN; light diagonal striped column, ntPE-V3MN26.

Characterization of Anti-V3 Loop Response-- Using an ELISA format, sera from rabbits immunized with either ntPE-V3MN26 or ntPE-V3ThE26 were characterized further. Wells were coated with gp120 derived from either the MN or Thai-E strain and then probed with various dilutions of rabbit antisera. Specific reactivity to both envelope proteins (after background subtraction; see "Experimental Procedures") was detected down to a dilution of 1:300,000 with a linear response in the dilution range of 1:10,000 to 1:100,000. Specifically, we wished to compare the reactivity for the V3 loop of gp120 with that of the V3 loop inserted within the toxin. However, because most of the antibodies produced against the ntPE-V3 chimeras were to toxin epitopes, it was not useful to probe immobilized chimeric protein directly. Instead we devised a competitive ELISA that allowed comparisons of antibody responses to soluble gp120 with those to soluble ntPE-V3 chimeras. For technical reasons, it was first necessary to remove anti-PE antibodies. This was done by passing immune rabbit sera over a wtPE affinity column and using the flow-through (see "Experimental Procedures").

HIV-1 Neutralizing Activity-- Sera from immunized rabbits neutralized HIV-1 infectivity in two separate in vitro assays (Figs. 6 and 7). Pre-immune sera showed no protection of a human T-cell line, MT4, from killing by HIV-1 MN. However, following immunization with ntPE-V3MN26, sera obtained from one rabbit at weeks 8 and 27 were protective against viral challenge with 50% neutralization occurring at approximately a 1:400 dilution (Fig. 6). Sera from a second rabbit exhibited 50% neutralization at 1:30 dilution. Based on the immunization schedule, week 5 sera reflected the response in animals immunized and boosted once, while week 8 sera were from animals boosted twice and week 27 sera came from animals boosted three times. These data suggested that subcutaneous injections of ntPE-V3 loop chimeras could result in the production of neutralizing antibodies.

View larger version (30K): [in this window] [in a new window]   Fig. 6.   A ntPE-V3 loop chimera administered to rabbits produces an antibody response capable of neutralizing HIV-1 infectivity in vitro. Rabbits were immunized subcutaneously with 200 µg of ntPE-V3MN26 and boosted similarly after 2, 4, and 12 weeks. Sera collected up to 27 weeks after the initial administration were evaluated for the ability to protect a human T-cell line, MT4, from killing by HIV-1 MN, as assessed by an 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide dye conversion assay. Values represent triplicate readings normalized against a control MT4 incubation not challenged by virus.

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Neutralization of clinical HIV isolates. Postvaccination sera from rabbits injected with ntPE-V3MN26 were mixed with either a B or E subtype virus. After a 1-h incubation at 37 °C, viral infectivity was determined by adding treated virus to PBMCs for another 3 days. Inhibition of viral growth was evaluated by measuring p24 levels. , p24 antigen (uninfected); , p24 antigen + prebleed sera; , p24 antigen + immune sera (24 weeks).

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

A vaccine to prevent HIV infection and/or retard viral progression within an infected individual is clearly needed. Many approaches are being considered, and some of these have been tested in clinical trials, but no superior candidates have yet emerged. Here we characterize V3 loop-toxin chimeras and report on their ability to produce strain-specific (MN-subtype B versus Thai-subtype E) antibody responses in rabbits (Fig. 5). When injected subcutaneously, the MN chimera produced, in at least one rabbit, neutralizing antibodies to both a subtype B and E isolate of HIV-1. Despite the fact that the MN26 and ThE26 inserts were ~50% identical at the amino acid level (Fig. 1C), injection of ntPE-V3ThE26 produced sera that consistently failed to recognize the MN insert or gp120/MN. In contrast, some rabbits injected with ntPE-V3MN26 produced sera that recognized both the MN and Thai-E sequences (albeit with a lower titer to the Thai-E sequence) and neutralized a representative clinical isolate from each subtype. The basis for this differential response remains to be determined.

Neutralizing antibodies were generated following the injection of proteins that had been refolded from inclusion bodies expressed in E. coli. In part, such favorable responses might have been due to the three-dimensional structure adopted by the V3 loop sequences within these chimeras. This view is supported by immunocapture data showing that soluble chimeric proteins were recognized by strain-specific anti-V3 loop monoclonal antibodies (Fig. 2C). Further, the reactivity of immune sera for immobilized gp120 could be completely absorbed with excess soluble gp120 (Fig. 5, B and C). Thus anti-V3 loop antibodies were primarily raised against epitopes that are exposed on gp120. Studies on the secondary structure of V3 loop peptides have shown that a turn-turn-helix motif is evident with circular (constrained by a disulfide bond) but not linear peptides (35). Also, circular peptides are recognized more readily by anti-V3 loop monoclonal antibodies than linear ones (35). Our data support the contention that V3 loops of PE chimeras have similar topologies to those found in gp120.

When enzymatically active versions of the chimeras were added to cells they inhibited protein synthesis with a similar dose response to that of wtPE (Fig. 4). Since the V3 loop sequences were inserted into the toxin's 37-kDa translocating fragment, one can assume that V3 loop sequences were co-transported to the cytosol. Toxin-mediated delivery to the cytosol may result in the generation of viral peptides and presentation via major histocompatibility complex class I antigens (36). This is under investigation for the ntPE-V3 loop chimeras. Previously, PE was shown to mediate the delivery of influenza peptides to major histocompatibility complex class I antigens (37). PE has also been shown to transport various peptides and enzymes to the cytosol (38, 39). Therefore, we believe these chimeras may be useful for the generation of HIV-reactive cytotoxic T-cells.

Sexual transmission of HIV-1 is via exposure at mucosal surfaces to both virus and virally infected cells. Ideally, a vaccine for HIV would produce not only a neutralizing systemic response, and possibly a cytotoxic T lymphocyte response but could also be used to generate mucosal immunity. The importance of mucosal immunization for defense against HIV has been proposed (40, 41). In cystic fibrosis patients infected with P. aeruginosa, there is a strong IgA anti-PE response (42). Our preliminary data indicate that PE-V3 loop chimeras, administered at mucosal surfaces, can elicit an IgA anti-V3 loop response (data not shown).

Other chimeras containing components of HIV-1 have been constructed and their immunogenic properties evaluated. These include a poliovirus antigen containing an epitope of the gp41 transmembrane glycoprotein from HIV-1 (43), a mucin protein containing multiple copies of the V3 loop (44), a genetically modified cholera B chain with V3 loop sequences (45), and a chemically detoxified PE-V3 loop peptide conjugate (46). These various approaches were each capable of producing anti-HIV responses. However, distinguishing features of our approach include the display of V3 loop sequences in near-native conformation, the ease with which multiple strain-specific chimeras can be produced (which will allow vaccine formulation with mixtures of chimeric proteins), and the potential to elicit both antibody and cellular responses from the same protein. In addition, because our chimeras are wholly recombinant constructs, there is no need to attach V3 loops using chemical cross-linking agents or use chemical treatments to inactivate the enzymatic activity of the toxin. Many conjugate vaccines use foreign proteins as carriers and adjuvants, but these are rarely directed to bind specific surface receptors. The favorable immune responses to PE-derived chimeric proteins may in part result from the targeting of V3 loops to cells expressing LRP/₂- macroglobulin receptor on their surface. By deleting the toxin's binding domain and replacing it with other cell-binding ligands, future versions of these chimeras may be targeted to selected cell populations.

Envelope proteins of HIV-1, such as gp120, are being evaluated as subunit vaccines with the expectation that antibodies to the V3 loop region of gp120 will provide protection through virus neutralization (12, 13, 47, 48). Conformational epitopes are believed to be required for optimal protective immunity (47). However, the injection of inactive virus or even the envelope protein itself has the potential to produce a mixture of neutralizing and so called "enhancing" antibodies (49-52). Because ntPE-V3 loop chimeras are constructed exclusively with amino acids derived from the tip of the V3 loop, we hope to generate a neutralizing response without the risk of generating enhancing antibodies.