Antibodies Generated in Cats by a Lipopeptide Reproducing the Membrane-Proximal External Region of the Feline Immunodeficiency Virus Transmembrane Enhance Virus Infectivity

Abstract

The immunogenicity of a lipoylated peptide (lipo-P59) reproducing the membrane-proximal external region (MPER) of the transmembrane glycoprotein of feline immunodeficiency virus (FIV) was investigated with cats. In the attempt to mimic the context in which MPER is located within intact virions, lipo-P59 was administered in association with membrane-like micelles. Analyses showed that in this milieu, lipo-P59 had a remarkable propensity to be positioned at the membrane interface, displayed a large number of ordered structures folded in turn helices, and was as active as lipo-P59 alone at inhibiting FIV infectivity in vitro. The antibodies developed differed from the ones previously obtained by immunizing cats with the nonlipoylated version of the peptide (G. Freer, S. Giannecchini, A. Tissot, M. F. Bachmann, P. Rovero, P. F. Serres, and M. Bendinelli, Virology 322:360-369, 2004) in epitope specificity and in the fact that they bound FIV virions. However, they too lacked virus-neutralizing activity and actually enhanced FIV infectivity for lymphoid cell cultures. It is concluded that the use of MPER-reproducing oligopeptides is not a viable approach for vaccinating against FIV.

Entry of lentiviruses into host cells is a multistep process mediated by trimers of the viral envelope glycoproteins SU (surface) and TM (transmembrane) (6, 13, 42). The highly conserved stretch of the TM immediately adjacent to the external monolayer of the viral envelope, termed the membrane-proximal external region (MPER), plays a crucial role in this process, effecting one of the events that occur after apposition and initial merging of the viral and cell membrane bilayers, most likely fusion pore expansion at least in the human immunodeficiency virus type 1 (HIV-1) (for a recent review, see reference 57). This, coupled with the circumstance that some of the few monoclonal antibodies with broad HIV-1-neutralizing activity recognize epitopes in this region (5, 36, 40, 58), has led others to propose MPER as an interesting target on which to model candidate protective immunogens and antivirals (57). In recent studies, we and others have demonstrated that the MPER of feline immunodeficiency virus (FIV), a naturally occurring lentivirus (41) intensively studied as a model system with which criteria for antiviral vaccines and drugs development can be tested (10, 55), shares important structural and functional features with the corresponding domain of HIV-1 (7, 14, 16, 35, 38, 47, 49, 50, 52). In addition, short synthetic peptides reproducing the MPER of FIV have been shown to inhibit the infectivity of this virus both in vitro and in vivo as a result of blocking cell entry (15, 17). However, an attempt to induce the production of FIV-neutralizing antibodies by immunizing cats with a 20-mer peptide reproducing the FIV MPER (767L to G786), designated P59, was unsuccessful: the antibodies elicited bound synthetic MPER-containing peptides and precisely a 3-amino-acid conformational epitope in the tryptophan-rich motif (TrpM) present in this region but failed to bind and neutralize FIV (12).

Recent biochemical and molecular investigations have pointed out that the epitopes recognized by two of the most potent HIV-1-neutralizing monoclonal antibodies, namely, 2F5 and 4E10, might be membrane context as well as sequence dependent (2, 19, 23, 27, 34, 37, 48). Therefore, we considered it of interest to test whether failure of P59 to induce FIV-reactive antibodies in the above study might be due to the incapacity of this peptide to maintain a functionally correct conformation unless incorporated into a membrane environment. To this end, we synthesized and characterized a lipoylated analogue of peptide P59 (lipo-P59). Because the in vitro properties of lipo-P59 examined here and in a previous report (11) were satisfactory, we immunized cats with it in association with membrane-like micelles (MLM), which might mimic the membrane context in which the MPER of FIV is located within intact virions. The antibodies thus produced differed from the ones elicited by P59 in epitope specificity and bound FIV virions effectively. However, even these antibodies were devoid of FIV-neutralizing activity. On the contrary, they exerted the paradoxical effect of enhancing FIV infectivity for cultured lymphoid cells.

MATERIALS AND METHODS

Animals, cells, and viruses.

Specific-pathogen-free (SPF) female cats were purchased from Iffa Credo (L'Arbresle, France) and immunized when 7 to 12 month old. They were housed individually in our climate-controlled animal facility in accordance with European Community guidelines. MBM cells are an interleukin 2-dependent line of T lymphocytes originally established from the peripheral blood mononuclear cells of an FIV- and feline leukemia virus-negative cat (33). They are routinely grown in RPMI 1640 medium supplemented with 10% fetal bovine serum, 5 μg of concanavalin A, and 20 U of interleukin 2 per ml. The viruses used were FIV Petaluma (FIV-Pet) of clade A and FIV Pisa M2 (FIV-M2) of clade B. The former was a tissue culture-adapted strain and was obtained from chronically infected FL4 cells (kindly provided by J. Yamamoto, Gainesville, FL), while the latter was a recent isolate and was grown in MBM cells.

Peptide synthesis and MLM.

The oligopeptides used in the study (Fig. 1A), including lipo-P59, which contained 2-amino-octadecanoic acid (Aod) at the C terminus, were manually synthesized on a solid phase, using standard Fmoc/tBu chemistry as previously described (11). Crude peptides were purified to homogenicity by semipreparative reverse-phase high-pressure liquid chromatography, with purity greater than 95%, and lyophilized. Multilamellar lipid vesicles mimicking MLM for molecular investigations were prepared using 1,2-dipalmitoyl-sn-glycero-3-phosphatidylcholine (DPPC) or 1,2-dipalmitoyl-sn-glycero-3-phosphatidylglycerol (DPPG) dissolved in chloroform-methanol (2/1 [vol/vol]). The lipid solution was evaporated and flash-dried with nitrogen for 3 to 5 min. The lipid film was hydrated with phosphate buffer (pH 7) overnight. The MLM used for other purposes contained 1,2-dipalmitoyl-sn-glycero-phosphocoline, sphingomyelin, and cholesterol (molar ratios, 3:1:1).

FIG. 1.

(A) Schematic representation of the C-terminal region of the TM of FIV and location of the synthetic peptides used in the study. Superscript numbers indicate amino acid positions according to the sequence of prototype FIV-Pet, clone 34TF10 (53). Domains are indicated as previously described (46): CHR, C-terminal α-helix region; MPER, membrane-proximal external region; TrpM, Trp-rich motif; MSR, membrane-spanning region; Cyto, cytoplasmic region. The amino acids forming TrpM are in bold, with the three residues participating in the epitope recognized by anti-P59 antibody (12) underlined. Map-C8 is a four-branched tetrameric analogue of peptide C8 used previously for assaying anti-TrpM antibody (12). Aod, 2-amino-octadecanoic acid. (B) Three-dimensional model of lipo-P59 associated with MLM. The structure was calculated using DYANA software on the basis of NOE restraints. The backbone, the side chains of Trp residues, and Aod of lipo-P59 are displayed.

CD and nuclear magnetic resonance (NMR) spectroscopies.

Peptide samples (150 μM) for circular dichroism (CD) spectra were prepared in a 25 mM phosphate buffer solution (PBS) (pH 7) with or without DPPC or DPPG MLM. MLM peptides were prepared, with peptides added at a concentration of 40 μM (lipid-to-peptide molar ratio, 1:50) and incubation of the solutions for at least 30 min. CD spectra were determined on an 810-Jasco spectropolarimeter at room temperature, using a quartz cuvette with a path length of 1 mm. The spectra were the average of 10 accumulations from 190 to 260 nm, recorded with a bandwidth of 1 nm at a scanning speed of 50 nm/min. All the spectra were analyzed, subtracted by blanks, and finally corrected by smoothing. Estimation of secondary structure composition was carried out using the algorithm K2D by Dicroweb (54).

Samples for NMR spectroscopy were prepared by dissolving the appropriate amount of peptide in water-dodecylphosphocoline (DPC) solution (pH of ≈5) to obtain a concentration of 1 mM peptide and 100 mM DPC. NMR spectra were recorded on a Bruker DRX-600 spectrometer. One-dimensional NMR spectra were recorded in the Fourier mode with quadrature detection, and the water signal was suppressed by low-power selective irradiation in the homogated mode. Double quantum filtered correlation spectroscopy (43), total correlation spectroscopy (TOCSY) (3, 4), and nuclear Overhauser effect spectroscopy (NOESY) (22, 31) experiments were carried out in the phase-sensitive mode using quadrature detection in ω₁ by time-proportional phase incrementation of the initial pulse (26). A mixing time of 70 ms was used for the TOCSY experiments. NOESY experiments were run at 300 K with mixing times in the range of 100 to 250 ms. Qualitative and quantitative analyses of DQF-COSY, TOCSY, and NOESY spectra were obtained using the SPARKY interactive program package (T. D. Goddard and D. G. Kneller, University of California, San Francisco).

Structure calculation.

Peak volumes were translated into upper distance bounds with the CALIBA routine of the DYANA software (20). The necessary pseudoatom corrections were applied for nonstereospecifically assigned protons at prochiral centers and for the methyl group. After redundant and duplicated constraints were discarded, the final list of constraints was used to generate an ensemble of 200 structures by following the standard protocol of simulated annealing in torsion angle space implemented in DYANA. No dihedral angle restraints and no hydrogen bond restraints were applied. Refinement of the structures was performed by in vacuo minimization using the Discover module of MSI INSIGHTII 2000 software (MSI Molecular Symulations, San Diego, CA). The structures were relaxed, constrained, and then unconstrained by using a combination of steepest descent and conjugate gradient minimization algorithms until the maximum root-mean-square derivative was less than 0.01 kcal/Å. Computations were performed using an SGI Octane computer. The model of peptide-MLM interaction was obtained using the equilibrated DPC and sodium dodecyl sulfate (SDS) bilayered PDB structure available at http://www.softsimu.org.

FIV inhibition assay.

Peptides were tested for the ability to inhibit FIV in 96-well flat-bottomed microplates against 10 50% tissue culture infectious doses of virus, using MBM cells as the substrate and measurement of the capsid protein p25 in the culture fluids at day 8 postinoculation as the readout. FIV was appropriately diluted and mixed with an equal volume of the test peptides, diluted from 50 to 0.005 μg per ml (final concentrations), alone or adsorbed onto MLM. Controls with MLM or the solvent alone were set up in parallel. The mixtures were then immediately inoculated into MBM cells and washed out after 5 h at 37°C. Supernatants were tested for p25 antigen content by enzyme-linked immunosorbent assay (ELISA) as described previously (17). This procedure was previously shown to provide an accurate assessment of entry inhibitor activity. Peptide inhibition of virus growth was calculated with the formula (mean p25 concentration in wells inoculated with FIV ± peptides/mean p25 concentration in wells inoculated with FIV alone) × 100. Fifty percent inhibitory concentrations (IC₅₀s) were calculated as previously described (17). Mean IC₅₀ ± standard deviations (SD) were calculated with all replicates. All experiments were repeated at least twice.

Cat immunization schedule.

Four SPF cats were inoculated intramuscularly with 200 μg of lipo-P59 adsorbed onto MLM in a total volume of 500 μl. Four doses were administered at 2-week intervals, and serum was obtained from each cat before the first inoculum and 3, 6, 9, and 10 weeks thereafter. All procedures were carried out with the cats being under light sedation.

Binding-antibody assays.

Antibody reactivity to synthetic peptides was tested by ELISA using microwells coated overnight with 100 μl of 10 μg/ml of individual peptides in carbonate buffer, pH 9.6. Antibodies binding intact FIV virions were measured by ELISA, as previously described (12, 18). Briefly, whole FIV was captured in microwells by means of Galanthus nivalis lectin (GN), which has the ability to capture FIV through the Env glycoproteins. GN-coated wells were obtained by incubation with 5 μg of GN (Sigma, St. Louis, MO) in PBS at room temperature overnight, followed by a saturation step with bovine serum albumin. GN-coated wells were then incubated twice with freshly prepared FIV for a total of 12 h to saturate all the lectin. After a postcoating step with skim milk, serially diluted sera were added to the plates in duplicate. In both assays, bound immunoglobulin G (IgG) was revealed with a mouse anti-cat IgG antibody followed by a goat anti-mouse peroxidase conjugate. Antibody titers were expressed as the reciprocal of the highest serum dilutions that gave optical density (OD) readings at least threefold higher than the average values obtained with 10 naive SPF cat sera plus three times the SD.

Neutralizing antibody assay.

FIV-neutralizing antibodies were measured against 10 50% tissue culture infectious doses of the virus using MBM cells as the indicator and 50% inhibition of reverse transcriptase (RT) production as the readout (18). Unless otherwise stated, the test was carried out by incubating the virus-serum mixtures at 4°C for 1 h. Then, the mixtures were inoculated into MBM cells, washed out after 5 h at 37°C, and replaced with fresh complete medium. Neutralizing titers were defined as the reciprocal of the highest serum dilution which reduced by ≥50% the levels of RT activity produced by virus exposed to the same dilution of serum pooled from 10 normal cats and calculated by the method of Reed and Münch (44). RT production was chosen as a readout in these assays rather than p25 production because the high titers of anti-p25 antibodies present in the sera of FIV-infected cats used in some experiments would have blurred the results.

IgG purification and Fab production.

IgG was purified by loading 5 ml of serum onto 1-ml HiTrap Protein A columns (Amersham Biosciences, Uppsala, Sweden). The columns were washed with 20 mM phosphate buffer (pH 7), and bound IgG was eluted with 0.1 M glycine buffer (pH 2.7). The pH was immediately brought back to neutrality, and fractions were filtered through a 0.45-μm filter (Costar, Acton, MA). Efficacy and reproducibility of purification were monitored by IgG subtype-specific ELISA. Antigen-binding fragment (Fab) was obtained by papain digestion. Briefly, papain at 1 mg/ml was activated by incubation in 0.05 M sodium phosphate buffer containing 0.01 M cystein, 0.02 M EDTA, and 0.13 M NaCl at 37°C for 1 h. Then, 0.1 mg/ml activated papain was added to each IgG fraction in a buffer containing 0.01 M cysteine, 0.002 M EDTA, 0.05 M sodium phosphate, and 0.1 M KCl. After 12 h at 37°C, the reaction was stopped by adding 0.02 M iodoacetamide. Digestion products corresponding to Fab and Fc fragments were separated using 1-ml HiTrap protein A columns as described for IgG purification. Before use, Fab was dialyzed against PBS at room temperature for 2 h. Efficiency of digestion and purification was checked by SDS-polyacrylamide gel electrophoresis and Coomassie blue staining. IgG and Fab concentrations were measured by the Bradford method.

RESULTS

Structural investigations.

To evaluate the impact of the lipophilic tail at the C terminus on peptide positioning on membranous surfaces, a comparative conformational analysis of P59 and lipo-P59 was performed with membrane-mimicking environments by means of CD and NMR spectroscopies. As membrane mimetics, DPPC or DPPG MLM and DPC micellar solutions were used for the CD and NMR studies, respectively. In agreement with previous observations (11), CD spectra showed that the lipophilic tail confers on lipo-P59 a preference for forming ordered structures (data not shown). NMR conformational analysis was consistent with the CD results. NOESY spectra demonstrated the presence of a higher number of nuclear Overhauser effect (NOE) connectivities in lipo-P59 than in P59. This suggests that lipo-P59 has a preferential tendency to adopt ordered conformations relative to P59. Indeed, structure calculation procedures, based on the NMR data, revealed in lipo-P59 the prevalence of a long segment of regularly structured α-helix, including the residues Trp770 to Lys785 (Fig. 1B); on the other hand, P59 was found to adopt a less regular turn-helical conformation, encompassing a shorter sequence including Trp770 to Ile780. Also, the data suggest that the lipophilic tail, which amplifies the hydrophobic interactions in lipo-P59, has the final effect of promoting peptide anchoring to the membrane surface. Thus, the structural model obtained (Fig. 1B) confirmed the remarkable degree of conformational stability of lipo-P59 in MLM and the propensity of this peptide to be positioned at the membrane interface.

Characterization of lipo-P59 in vitro.

Previous in vitro inhibition studies had shown that lipo-P59 inhibited FIV at IC₅₀s similar to those exhibited by P59 and, likely as the result of an increased stability, had a longer-lasting effect (11). To evaluate whether lipo-P59 in MLM (lipo-P59-MLM), i.e., in a micellar condition that reproduced as much as possible the lipid raft where the MPER is located in the virion surface, was still functionally active, we determined if lipo-P59-MLM conserved this activity. Table 1 shows the results of experiments in which lipo-P59 and P59, alone or mixed with MLM, were compared for FIV inhibition. As shown, lipo-P59-MLM and lipo-P59 exhibited similar IC₅₀s for both FIV-M2 and FIV-Pet. Contrariwise, P59 lost much of its antiviral activity when mixed with MLM, possibly due to partial loss of availability.

TABLE 1.

Inhibition of FIV by peptides lipo-P59 and P59 alone or mixed with MLM

Virus	Mean IC50 ± SD (μg/ml) ofa:
	Lipo-P59	Lipo-P59 + MLM	P59	P59 + MLM	MLM
FIV-M2	0.08 ± 0.03	0.06 ± 0.01	0.08 ± 0.02	3.36 ± 1.01	>50
FIV-Pet	0.07 ± 0.02	0.11 ± 0.10	0.06 ± 0.01	1.19 ± 0.13	>50

^aExperiment repeated three times with comparable results.

Immunogenicity of lipo-P59-MLM in cats.

Four SPF cats were administered four 200-μg doses of lipo-P59-MLM intramuscularly and then analyzed by ELISA at selected times for the antibody responses mounted. Antibodies were analyzed using microplates coated with lipo-P59-MLM, lipo-P59, or for a control, peptide E29K (Fig. 1) alone or mixed with MLM. As depicted in Fig. 2, all four cats responded to immunization with the production of antibodies which reacted with lipo-P59 and, to a slightly lesser extent, with lipo-P59-MLM but not with the unrelated epitope. As shown, lipo-P59 binding antibodies started to develop 1 week after the second immunizing dose and peaked 3 weeks after the last dose, declining rapidly thereafter. Subsequent studies were carried out with the sera taken at the time of peak antibody response.

FIG. 2.

ELISA antibody development in four SPF cats immunized with lipo-P59-MLM. Sera obtained at the indicated times were diluted 1:100 and analyzed for reactivity to lipo-P59 (⧫), lipo-P59-MLM (•), the unrelated peptide E29K alone (□) or mixed with MLM (○), and MLM alone (▵). Values are means ± SDs of the optical densities (OD) obtained for individual sera. Arrowheads indicate immunizing doses.

Location of epitope(s) recognized.

In a previous study (12), antisera were raised by immunizing cats with unmodified P59 covalently coupled to phage Qβ virus-like particles (anti-P59), and these antisera were found to recognize one small epitope contained in the TrpM of the FIV MPER (Fig. 1A). To investigate whether the antisera generated by lipo-P59 recognized this same epitope or a distinct one(s), we compared the binding properties of lipo-P59 antisera to those of the previously produced anti-P59. The study consisted in reacting the sera with a panel of synthetic MPER peptides that had helped in the identification of the TrpM epitope recognized by the anti-P59. The panel included several peptides encompassing the TrpM, namely, P59 and lipo-P59, used as such or mixed with MLM, the 36-mer peptide L36I, which contains the TrpM at its C terminus, and Map-C8, a four-branched tetrameric analogue of the TrpM. As a control, the 29-mer peptide E29K, completely external to the TrpM but adjacent to P59, was also used (Fig. 1A). As shown in Table 2, lipo-P59 antisera failed to react with Map-C8 or with E29K and reacted only sporadically with the other TrpM-containing peptides (P59 alone or mixed with MLM and L36I). This binding pattern contrasted with that of anti-P59, which, as expected based on our previous findings (12), consistently bound Map-C8 and the other TrpM-containing peptides, with the only exception being P59, in which most likely the appropriate conformation-dependent epitope is unavailable (12, 23).

TABLE 2.

Binding of anti lipo-P59 and P59 cat sera to a panel of synthetic peptides containing or not containing the TrpM

Test peptide	Reactivity ofa:
	Presera	Anti-lipo-P59 sera	Anti-P59 sera
Lipo-P59	0/8	4/4 (800-1,600)	4/4 (100)
Lipo-P59-MLMb	0/8	4/4 (800-1,600)	4/4 (100)
P59	0/8	1/4 (100)	0/4
P59-MBMb	0/8	1/4 (100)	0/4
L36I	0/8	1/4 (100)	4/4 (100)
Map-C8	0/8	0/4 (<100)	4/4 (100)
E29K	0/8	0/4 (<100)	0/4 (<100)

^aNumber of positive sera/number of sera tested (range of titers).

^bNo serum reacted with MLM alone.

Binding of lipo-P59 antibodies to intact FIV virions.

We next examined whether the epitope(s) in lipo-P59 is also expressed on the surfaces of intact FIV virions. To this end, lipo-P59 antisera and control presera were reacted with whole FIV bound to ELISA microwells through GN lectin (18). Table 3 shows that all four lipo-P59 sera but not the respective presera bound intact FIV at relatively high titers.

TABLE 3.

Ability of lipo-P59 sera to bind whole FIV virions and neutralize FIV infectivity

Sera	Binding activitya	Neutralizing activitya
Lipo-P59 antisera	4/4 (200-400)	0/4
Preimmune sera	0/4	0/4
Positive controlb	5/5 (400-1,200)	5/5 (64-128)

^aNo. of positive sera/no. of sera examined (range of titers).

^bSera of cats chronically infected with FIV-Pet.

Recognition of lipo-P59 by FIV infected cat sera.

Having shown that the epitope(s) in lipo-P59 is expressed on the FIV surface, it was important to check whether the same epitope(s) was capable of evoking a humoral immune response in the course of FIV infections. We therefore examined, against the same panel of peptides used above, the sera of 13 SPF cats who had been infected with FIV-Pet or FIV-M2 for 20 to 37 months. As shown in Table 4, uninfected cat sera used as controls were uniformly negative to all test peptides. In contrast, four FIV-Pet-infected cat sera and two FIV-M2-infected cat sera bound lipo-P59 regardless of whether this was mixed with MLM but not P59. On the other hand, the reactivities of the infected cat sera to the other peptides in the panel were in agreement with the fact that the FIV TrpM is totally silent in the course of infection (12, 32): in particular, confirming previous findings (12), these sera bound the peptides L36I (13/13 cats) and E29K (5/13 cats) but were completely unreactive to Map-C8.

TABLE 4.

Binding of FIV-infected cat sera to lipo-P59

Test peptide	Reactivity ofa:
	FIV-infected sera	Uninfected sera
Lipo-P59	6/13 (100)	0/12
Lipo-P59-MLMb	6/13 (100-400)	0/12
P59	0/13 (<100)	0/12
P59-MLMb	0/13 (<100)	0/12
L36I	13/13 (400-800)	0/12
Map-C8	0/13 (<100)	0/12
E29K	5/13 (200-400)	0/12

^aNo. of positive sera/no. of sera tested (range of titers).

^bNo serum reacted with MLM alone.

Enhancement of FIV infection but no neutralization with lipo-P59 antibodies.

We finally checked whether the antibodies generated in cats by lipo-P59 could inhibit the infectivity of FIV in a sensitive in vitro neutralization test using lymphoid cells as the substrate (18). As shown by Fig. 3A, while a positive control serum from a chronically infected cat inhibited FIV replication up to a dilution of 1:32, the four lipo-P59 sera showed no evidence of such an activity. On the contrary, all of the latter sera exerted a virus-enhancing effect that was clearly evident up to the highest dilution tested (1:128). As shown by Fig. 3B, the enhancing effect was also produced by purified IgG and Fab fragments obtained from the lipo-P59 antisera. We then examined whether this enhancement might be blocked by adsorbing with lipo-P59. The above IgG and Fab were incubated at 1 μg/ml for 1 h at 4°C in microwells that had been coated overnight with 1 μg of lipo-P59, and the fluids thus treated were then centrifuged, filtered, and tested for their effects on FIV replication. As shown by Fig. 4, following adsorption to the lipo-P59-coated wells, both the IgG and the Fab derived from the lipo-P599 sera lost all or most of their FIV-enhancing activity relative to similar control fluids that had been incubated in uncoated microwells.

FIG. 3.

Effect of lipo-P59 antisera on FIV replication in vitro. (A) MBM cell cultures were inoculated with FIV-Pet which had been incubated for 1 h at 37°C with the indicated dilutions of the four antisera (▪, ▴, •, and ⧫) or one positive-control, long-term FIV-infected cat serum (□) and assayed for RT activity in the supernatant fluids 8 days later. (B) The test was repeated with the indicated concentrations of purified IgG (continuous lines) or Fab (dashed lines) obtained from the same sera used for panel A. Dashed bold lines represent the 50% neutralization level. Asterisks indicate significant difference (P < 0.01; Student's t test) relative to the values obtained with the 0.1-μg/ml concentration.

FIG. 4.

Blocking of the FIV-enhancing effect of lipo-P59 antisera by lipo-P59. One hundred microliters of 1 μg/ml purified IgG and Fab from the four lipo-P59 sera, one FIV-infected cat serum (positive control), and one naive cat serum (negative control) were incubated for 1 h at 4°C in microwells that had been coated overnight with 1 μg of lipo-P59 (empty columns) or unrelated peptides (solid columns). The fluids were then centrifuged, filtered, and finally tested for the effects on FIV replication as for Fig. 3. Dashed lines represent the 50% neutralization level. Asterisks indicate significant difference (P < 0.05; Student's t test) relative to the values obtained after adsorption with the unrelated peptide.

DISCUSSION

In a previous report, the nonlipoylated version of lipo-P59, injected into cats coupled with virus-like particles, had elicited antibodies that bound the TrpM portion of the peptide but lacked FIV binding and neutralizing activity, thus showing that the epitope recognized was either irrelevant or functionally unavailable in virions (12). The present study was prompted by hopes that presenting P59 in a membrane-mimetic environment might permit a conformation of the oligopeptide that more faithfully reproduced one of the conformations the MPER of the TM presents in virions while these are resting or in the process of entering susceptible cells. Thus, we examined whether immunization of cats with lipo-P59 incorporated in MLM elicited antibodies capable of interacting with intact FIV virions.

The in vitro properties of lipo-P59, examined in this and in a previous study (11), were as expected, since (i) in amphiphilic systems mimicking biological membranes, lipo-P59 displayed a greater conformational stability and a greater propensity to be positioned on the membrane interface than P59; (ii) CD and NMR spectroscopy analysis in the presence of MLM revealed the presence of a larger number of ordered structures folded in turn helices in lipo-P59 than in P59, most likely due to enhanced peptide anchoring on the surface of the vesicles via the lipophilic tail at the C terminus; (iii) structure calculations based on NMR and NOE data demonstrated a longer folded region in lipo-P59 (residues Trp770 to Lys785) than in P59 (residues Trp770 to Ile780) as a result of a conformation-stabilizing effect conferred by the hydrophobic tail; and (iv) infectivity inhibition studies showed that lipo-P59 inhibited FIV at an IC₅₀ similar to that of P59, and in addition, it had a longer-lasting effect and remained fully active even when mixed with MLM.

All four cats that were immunized with lipo-P59 in MLM developed antibodies that bound the immunizing peptide, regardless of whether this was alone or mixed with MLM, but were virtually unreactive to P59, regardless of whether this peptide was mixed with MLM or not, probably due to the fact that relative to lipo-P59, it has a less-regular conformation or a greater tendency to remain occluded within or to be distorted by the MLM. In fact, as determined with the panel of peptides used for characterization, the four lipo-P59 sera had epitope specificities different from those of the antibodies generated by P59, probably reflecting the fact that in the lipo-P59-MLM used for immunization, the epitope recognized by the anti-P59 sera (11, 57) lay within the MLM, resulting in its immunological silencing. Most importantly, however, unlike the P59 sera, all four lipo-P59 sera bound intact FIV particles, thus showing that the immunizing strategy used had indeed led to an increased similarity between the immunizing peptide and the MPER conformation as present in the native TM found in intact FIV virions.

Upon virus interaction with susceptible cells, the MPER of FIV is believed to undergo a series of conformational changes similar to the ones, described in much more detail (9, 14, 16, 35), that are known to occur during cell entry in the corresponding region of HIV-1. Thus, theoretically, antibodies reactive against this region may interact not only with epitopes present in resting virions but also with epitopes expressed more or less transiently in the course of cell entry. The fact that the lipo-P59 sera bound intact cell-free FIV strongly suggests that the epitope(s) recognized is accessible on resting virions. Precise amino acid mapping of the epitope(s) involved will probably shed additional light on this aspect. In any case, the observation that only approximately half of the FIV-infected cats examined had detectable lipo-P59 binding antibodies indicates that during in vivo infection, the epitope(s) defined by the lipo-P59 sera is weakly or only sporadically immunogenic. It should also be noted that these antibodies are unlikely to impact significantly the therapeutic potential of MPER-derived peptides, as suggested by recent findings showing that a retroinverso derivative of P59 exerted a powerful antiviral activity in FIV-infected cats (15).

The most interesting, albeit disappointing, finding of this study, however, is that the lipo-P59 sera not only lacked FIV-inhibitory activity in a sensitive neutralizing assay but in fact had the paradoxical ability to enhance its infectivity in vitro. Although we have no evidence that the effect observed is relevant in vivo, an enhancing or accelerating effect on FIV replication has been described for several situations, including vaccine experiments using immunogens of varied complexity, in which the factors responsible have most often eluded a clear identification (18, 21, 24, 28, 45, 51). The enhancing effect described here was instead clearly characterized as mediated by the antibodies generated by lipo-P59, since it was reproduced by IgG and Fab purified from the sera and was ablated by preadsorption with lipo-P59. Although the mechanism responsible for the effect was not investigated, it is plausible that by interacting with the MPER of FIV virions, lipo-P59 antibodies stabilize these molecular structures in the appropriate location, thus facilitating virus fusion (29, 30, 56). Indeed, according to current views, during cell entry by lentivirus the MPER domain is believed to undergo a sequence of conformational transitions, plunging some Trp residues into the viral membrane and others into the cell membrane to accomplish the final step of viral fusion (9, 13). Whatever the mechanism, the finding has revealed for the first time that the MPER of the FIV TM contains at least one epitope potentially capable of inducing virus-enhancing antibodies and that this epitope is also expressed by a properly presented MPER-reproducing oligopeptide. In contrast, putative neutralizing epitopes present in the MPER of FIV have to date proven difficult to mimic using the oligopeptide approach (1, 8, 12, 32, 39, 46). It is noteworthy that the MPER in the TM of HIV-1 also is poorly immunogenic during virus infection and that although this structure contains well-conserved epitopes recognized by a few neutralizing human monoclonal antibodies (57), all efforts to generate neutralizing antibodies with a variety of linear or structurally constrained MPER peptides, used as such or within different protein scaffolds or virus-like particles, have so far failed (25, 27). This may reflect the fact that the immunogens used fail to adopt conformations sufficiently similar to that of the MPER as present in intact virions and even the fact that the native epitope(s) involved may be partially occluded and difficult for antibodies to interact with.