Analysis of physical interactions between peptides and HLA molecules and application to the detection of human immunodeficiency virus 1 antigenic peptides.

The physical association of 40 antigenic peptides and purified HLA class I and class II molecules was monitored using a direct peptide binding assay (PBA) in solid phase and an inhibition peptide binding assay (IPBA) in which the competing peptide was present in a soluble phase. We also examined the ability of different peptides to inhibit the lytic activity of human antiviral cytolytic T cells towards cells incubated with the corresponding target peptide. Our results showed that: (a) Binding of a given human T cell-recognized peptide to several HLA class I and class II molecules occurred frequently. Nevertheless, preferential binding of peptides to their respective restriction molecules was also observed. (b) Binding of HLA molecules to peptides recognized by murine T cells occurred less frequently. (c) 11 of 24 (46%) randomly selected HIV-1 peptides contained agretopic residues allowing their binding to HLA molecules. (d) The kinetics of HLA/peptide association depended on the peptide tested and were faster than or similar to those reported for Ia molecules. Dissociation of these complexes was very low. (e) Peptide/HLA molecule binding was dependent on length, number of positive charges, and presence of hydrophobic residue in the peptide. (f) A correlation was demonstrated between a peptide inhibitory effect in the IPBA and its blocking effect in the cytolytic test. Our data indicated that the restriction phenomenon observed in T cell responses was not strictly related to either an elective HLA/peptide association, or a high binding capacity of a peptide to HLA molecules. These data also showed that the PBA and IPBA are appropriate for the detection of agretopic residues within HIV-1 proteins.

Role of N-linked glycans in the interaction between the envelope glycoprotein of human immunodeficiency virus and its CD4 cellular receptor. Structural enzymatic analysis.

gp120 and CD4 are two glycoproteins that are considered to interact together to allow the binding of HIV to CD4+ cells. We have utilized enzymatic digestion by endoglycosidases in order to analyze N-linked carbohydrate chains of these proteins and their possible role in the interaction of gp120 or gp160 with CD4. SDS denaturation was not necessary to obtain optimal deglycosylation of either molecule, but deglycosylation of CD4, nonetheless, depended on the presence of 1% Triton X-100. Endo H and Endo F that cleave high mannose type and biantennary glycans diminish the molecular mass of the glycoproteins from 120 or 160 Kd to 90 or 130 Kd, respectively; but these enzymes had no action on CD4 glycans. Endo F N-glycanase mixture, which acts on all glycan species, including triantennary chains, led to complete deglycosylation of gp120/160 and of CD4. Therefore, probably half of the glycan moieties of gp120/160 are composed of high mannose and biantennary chains, the other half being triantennary species. The carbohydrate structures of CD4 seems to be triantennary chains. To analyze the binding of gp120/160 to CD4, we used a molecular assay in which an mAb (110-4) coupled to Sepharose CL4B allowed the attachment of soluble gp120/160 to the beads; 125I-sCD4 was then added to measure the binding of CD4 to different amounts of gp120/160. Binding to gp160 was not modified when using completely deglycosylated 125I-sCD4, while deglycosylation of gp120 or of gp160 resulted in the decrease of the binding to native CD4 by two- and fivefold, respectively. Native and deglycosylated gp120/160 bound to CD4+ cells with comparable affinities. In addition, deglycosylated gp120 displaced 125I-gp160 binding to CD4+ cells and inhibited fusion of fresh Molt-T4 cells with CEM HIV1- or HIV2-infected cells to the same extent. Taken together, these results indicate that carbohydrates of CD4 and of gp120/160 do not play a significant role in the in vitro interaction between these two molecules.

Linear and cyclic peptides mimicking the disulfide loops in HIV-2 envelope glycoprotein induced antibodies with different specificity.

The aim of this study was to compare the immunogenicity and antigenicity of cyclic and linear peptides that mimic the disulfide loops in HIV-2ROD gp125. Based on the hypothetical assignment of intrachain disulfide bonds in HIV-2 envelope glycoprotein, peptides expected to mimic all 11 disulfide-bonded domains were synthesized, oxidized or cysteine-alkylated; they were then purified and characterized. Rabbits were immunized with either linear cysteine-alkylated peptides (L1-L11) or cyclic oxidized peptides (C1-C11). All peptides except 7L elicited antibodies with titers between 10(3) and 5 x 10(6). Anti-peptide C (2, 3, 4, 7, 8, 9, 11) and anti-peptide L (2, 3, 8, 9, 11) antibodies recognized the native HIV-2 gp 125. Moreover, we found that cyclization of the peptides significantly increased the level of anti-peptide antibodies reacting with the intact antigen protein. Deglycosylation increased the level of protein reactivity of anti-peptide antibodies and rendered the epitopes in peptides 5, 6, 10 accessible, which were masked in the native protein. Peptide 1 induced antibodies reacting only with the denatured reduced gp125 HIV-2. In addition, while anti-peptide L antibodies reacted better with L peptide (called "linear" structural specificity), anti-peptide C antibodies reacted similarly with L and C peptides (called "broad" structural specificity). Interestingly, the "broad" structural specificity of antibodies correlated with reactivity against native gp125. Although none of these anti-peptide antisera displayed neutralizing activity against HIV-2ROD, these results support the hypothesis that the structural restriction of peptides have a major influence upon the generation of more specific antibodies for recognizing the intact protein.

Specificity of antipeptide antibodies produced against V2 and V3 regions of the external envelope of human immunodeficiency virus type 2.

The V2 region of simian immunodeficiency virus (SIV) and V3 region of human immunodeficiency virus type 1 (HIV-1) have been reported to be neutralization epitopes. We analysed the corresponding regions in HIV-2. Synthetic peptides modeling the V2 (aa 149-168) and V3 (CV3: aa 298-315 and NV3: aa 306-324) regions of the HIV-2 external envelope glycoprotein were coupled to KLH and used as immunogens in rabbits. We characterized the resulting antiV2 and antiV3 antibodies for their ability to recognize native and deglycosylated HIV-2 envelope glycoprotein, to block gp-CD4 interaction and to inhibit syncytium formation in vitro. The three synthetic peptides induced antibodies able to recognize specifically the native HIV-2 envelope glycoprotein with a significant avidity (K0.5 between 6 x 10(-7) and 8 x 10(-9) M). Interestingly, the reactivity of antibodies produced against the V2 peptide, which contains two potential sites of N-glycosylation, was higher against the fully deglycosylated than glycosylated HIV-2 external envelope glycoprotein (gp105). The antipeptide antibodies were used to investigate the topography of these regions in the preformed gp-CD4 complex in indirect immunofluorescence assays. The V2 and V3 regions in the complex remained accessible to their respective antibodies. Moreover, preincubation of gp105 with anti V2 or anti V3 antibodies did not prevent gp-CD4 interaction. Thus the V2 and V3 regions are not directly involved in the gp105 binding site for the CD4 receptor. Finally, in contrast with results obtained with antibodies produced against the V3 region of HIV-1 gp120 and monoclonal antibodies produced against the V3 of SIV, antibodies produced against V2 and V3 of HIV-2 were unable to inhibit syncytium formation induced by HIV-2 in vitro.

Immunochemistry of scorpion alpha-toxins: purification and characterization of two functionally independent IgG populations raised against toxin II of Androctonus australis Hector.

We report the isolation and characterization of two IgG populations specific to two synthetic peptides corresponding to two antigenic sites of toxin II of the North African scorpion Androctonus australis Hector. Firstly, thanks to the use of: (1) antigenic homology studies between toxin II of A. australis Hector and toxin III of Buthus occitanus tunetanus, (2) chemical modification of toxin II of A. australis Hector, and (3) prediction of the localization of the four major antigenic sites of scorpion alpha-toxins by the method developed by Hopp and Woods [Proc. natn. Acad. Sci. U.S.A. 78, 3824-3828 (1981)], we have established that the region around the disulfide bridge between cysteines 12 and 63 as well as the stretch of residues 50-59 probably each enclosed an antigenic site. Secondly, the synthetic replicates of these regions linked to Sepharose allowed us to isolate, by immunoaffinity chromatography, two IgG populations from the whole anti-toxin II of A. australis Hector IgGs. Finally, each of these two IgG populations was shown to be specific to one antigenic site as evidenced by the multideterminant effect on the slopes of binding curves developed by Berzofsky et al. [Biochemistry 15, 2113-2121 (1976)]. Furthermore, these two IgG populations were found to be functionally independent and this could be related to the fact that the two regions carrying the two antigenic sites are not close to each other in space and that there is neither steric hindrance nor cooperative effects between them. The association constant of these site-specific IgG populations was calculated and found to be equal to 1.18-5.14 X 10(9) l/mole for IgG anti-site 1 and 1.16-5.62 X 10(9) l/mole for IgG anti-site 2 respectively by Sips [J. chem. Phys. 16, 490-495 (1948)], Scatchard [Am. N.Y. Acad. Sci. 51, 660-772 (1949)] and Steward and Petty [Immunology 23, 881-887 (1972)] representations. The index of heterogeneity of 0.9 for anti-P1 and anti-P2 indicates the purification of essentially homogeneous affinity IgG populations.

Specificity of anti-P25 antibodies produced against whole HIV-1 particles or soluble forms of the protein.

Specificity of anti-p25 antibodies produced against either whole Human Immunodeficiency Virus type 1 (HIV-1) particles in humans and chimpanzees, or against soluble forms of the protein in chimpanzees and rabbits was analyzed by ELISA using a panel of 37 long (> or = 30 residues) or shorter (9-21 residues) overlapping peptides covering the entire p25 sequence. Antibodies elicited by intact virions presented similar reactivity patterns in HIV-1-infected humans and in HIV-1-infected or immunized chimpanzees and recognized only a limited region mostly the C-terminus of the molecule. Moreover, 8 of the human sera (36%), which nonetheless reacted with high titers and avidity with native p25, did not bind to any long or short peptide. These results suggest that the majority of antibodies elicited by viral particles are presumably directed to conformational epitopes. In contrast, antibodies raised against soluble forms of p25 could react against all long peptides but one (residues 211-245) and against some short peptides, indicating that most of p25 sequence may be immunogenic under these conditions. These results suggest that the reactivity spectrum of anti p25 antibodies is rather different if they are produced against intact HIV-1 particles or the soluble protein. They also indicate that it may be possible to manipulate the specificity of the humoral immune response by using either intact virions or purified proteins.

The antigenic structure of a scorpion toxin.

Scorpion toxins constitute a family of homologous proteins that exert potent pharmacological effects on ion channels. These proteins are immunogenic and constitute a good model for investigation of the molecular basis of antigenicity. In the first part of this article we summarize the results we have obtained in recent years concerning the location of the main antigenic regions of a model toxin, toxin II of the North African scorpion Androctonus australis Hector. Then, thanks to the recently available atomic coordinates of this toxin, we analyzed the relationships between the structural features of the protein and the location of the antigenic regions: we found that antigenic regions are located at exposed parts of the molecular surface, i.e. in reverse turns and the alpha-helix. These surface parts also correspond to segments of the polypeptide chain which are most accessible to a large spherical probe modelizing an antibody molecule. Finally, we obtained a general idea of what could be the main discontinuous antigenic determinants by looking for the neighboring relationships between the most exposed residues of the protein.

Evaluation of structure-antigenicity relationship of peptides from human immunodeficiency virus type 1 (HIV-1) p18 protein by circular dichroism.

The antigenicity of Human Immunodeficiency Virus type 1 (HIV-1) matrix p18 protein was evaluated by analyzing the specificity of anti-p18 antibodies elicited either in HIV-1 infected humans, or in HIV-1 infected or immunized chimpanzees, against a panel of long and short overlapping synthetic peptides [from 12 to 46 amino acid (aa) residues] covering the entire sequence of p18. The relationship between peptide structure and antigenicity was further investigated by probing the secondary structures of the peptides by circular dichroism. The results obtained clearly showed the immunodominance of the N-terminal region mimicked by peptide P1 (aa 2-45), which reacted with 52 and 100% of human and chimpanzee anti-p18 sera, respectively. In contrast smaller 15 aa long peptides C1, C2, C3, C4 and P3 which cover the entire sequence of immunodominant peptide P1, showed only weak or no reactivity. In contrast to widely accepted hypotheses, circular dichroism analysis of both small and large peptides secondary structures did not show any obvious correlation between antigenicity and the ability of peptides to adopt an ordered conformation.

Immunochemistry of scorpion alpha-toxins: study with synthetic peptides of the antigenicity of four regions of toxin II of Androctonus australis Hector.

Sequences 19-29 and 28-39 of toxin II of the North African scorpion Androctonus australis Hector have been synthesized. These two peptides correspond to the highest peaks in the hydrophilicity profile of toxin II and were thus believed to account for a significant proportion of toxin antigenicity. Affinity chromatography of solid-phase-bonded peptides allowed us to purify two sub-populations from the total IgGs raised against the native toxin. They both still bound to 125I-toxin II and showed a restricted heterogeneity in their specificity. Solid-phase immunoassays confirmed the antigenicity of these synthetic peptides and also that of two other previously described synthetic replicates of the antigenic regions of toxin II: sequences (5-14) S-S (60-64) and 50-59. The location of the four antigenic regions relative to the postulated location of the receptor-binding site of the toxin is discussed.

Identification of antigenic residues on apamin recognized by polyclonal antibodies.

The structural requirements for antigenic recognition of apamin--an 18-amino acid, disulfide-bridged peptide--by rabbit antibodies were defined using a set of 18 apamin analogs in a competition liquid-phase radioimmunoassay. Some residues contribute considerably to antigenic recognition, e.g. Ala10, Arg13, and others to a lesser extent, e.g. Arg14, Glu7 and Thr8. The N- and C-terminal moieties of apamin are less antigenically important. These findings suggest that a good part of antibody specificities are directed to the central tightly folded part of the molecule. They are consistent with the observation that in saturating conditions, labeled apamin can, on average, bind one specific Fab fragment.

HIV infection of monocytic cells: rôle of antibody-mediated virus binding to Fc-gamma receptors.

We investigated whether human immunoglobulin G (IgG) directed to gp110 may serve as an attachment system to Fc-gamma receptors (Fc-gamma R), allowing eventual infection of cells of the macrophage lineage. An anti-HIV IgG preparation that prevented viral particles and soluble recombinant radiolabelled envelope precursor gp160 from binding to CD4 on CEM lymphoid cells, and that strongly inhibited infection of these cells by HIV, was selected. In contrast, anti-HIV IgG, whether or not previously complexed to viral particles, bound to monocytic U937 cells that express both high Fc-gamma RI and low affinity Fc-gamma RII receptors. Precoating these cells with anti-HIV IgG or complexing the antibodies with soluble 125I-gp160 resulted in increased fixation of gp160 to the cells, which was inhibited by aggregated human normal IgG. These data indicate that anti-HIV IgG-dependent attachment of gp160 to monocytic cells occurs through both types of Fc-gamma R. In addition, this method of attachment resulted in productive infection of U937 cells that, since it was blocked in the presence of Leu3a, still appeared to involve gp110-CD4 interaction. Only slight enhancement of infectivity, such as described for other enveloped viruses, was noted, even when antibody concentration was titrated down. This mechanism may be one of the explanations why the humoral response to HIV is not usually protective.

Production and characterization of human monoclonal antibodies against core protein p25 and transmembrane glycoprotein gp41 of HIV-1.

With a view to obtaining human monoclonal antibodies against HIV-1 antigens we used the Epstein-Barr virus immortalization technique to induce lymphoblastoid cell lines from peripheral blood lymphocytes of 10 people who were seropositive for HIV-1 and had no clinical symptoms. A number of polyclonal lines were obtained which synthesized antibodies against most of the major proteins and glycoproteins of HIV-1. Three stable clones were characterized for class, secretion characteristics and specificity. Two of these clones produce antibodies which react with gp41, and the third reacts with p25. One of the anti-gp41 antibodies was found to have neutralizing activity.

Immunogenicity and antigenicity of conserved peptides from the envelope of HIV-1 expressed at the surface of recombinant bacteria.

We expressed peptides from the HIV-1 envelope protein at the surface of Escherichia coli by genetic insertions into an exposed loop of the outer membrane protein LamB. Recombinant bacteria expressing eight peptides from gp110 (pep1-pep8), conserved between HIV-1 and HIV-2, were used as live immunogens in rabbits by the intravenous route. The eight constructions elicited anti-LamB antibodies, showing that the hybrid proteins were immunogenic. One of them, LamB-pep8, gave rise to antibodies able to react with gp160 and to neutralize HIV-1 in vitro. We also show that this type of recombinant E. coli can provide a convenient reagent to monitor and characterize specific antibodies. Recombinant clones were used to test sera of seropositive individuals, as well as to narrow down the monoclonal antibody 110-1 recognition site to a cluster of eight residues at the carboxy-terminal end of gp110.

Use of synthetic peptides for the detection of antibodies against the nef regulating protein in sera of HIV-infected patients.

Human sera were tested for the presence of anti-nef antibodies by radioimmunoassay (RIA), with recombinant radiolabelled nef expressed in E. coli. Of the 300 HIV-positive sera tested by RIA, 70 +/- 5.3% were found to be anti-nef positive. Anti-nef antibodies bound to nef with a high affinity (K 0.5 = 2.2 x 10(-9) M). In 31 of the sera, the specificity of anti-nef antibodies was further analysed by enzyme-linked immunosorbent assay (ELISA) with large synthetic peptides ranging from 31 to 66 amino acid residues and spanning the total sequence of nef from HIV-1. The results obtained showed that the immunodominant antigenic sites of nef were located close to the N- and C-terminal regions of the molecule.

A molecular mechanism of inhibition of HIV-1 binding to CD4+ cells by monoclonal antibodies to gp110.

We have investigated the possible involvement in the interaction between HIV gp110 and its CD4 receptor of epitopes different from the currently known binding site(s) of the molecule. Four monoclonal antibodies (MAbs) to gp110 were used (Genetic Systems Corporation, Seattle, Washington, USA): one (110-1) recognized a peptide corresponding to the C-terminal part of gp110 (494-517); the other three (110-3, 110-4, 110-5) recognized the same peptide located at position 308-328. HIV or purified gp110 obtained from a vaccinia recombinant (Transgene S.A., Strasbourg, France) were pre-incubated with the MAb prior to addition to CD4+ cells. Specific binding of viral particles or of the soluble molecule was then determined by flow cytometer analysis, compared with that of control preparations where the MAb was added after HIV or gp 110 had been allowed to bind CD4+ cells. Significant inhibition of HIV binding was noted with the three MAbs to peptide (308-328), but not with 110-1. At the molecular level, these same MAbs decreased the affinity of interaction between CD4 and soluble gp110, although they could still label the latter molecule after it had bound to CD4+ cells. Therefore, steric hindrance may account for neutralization of HIV binding by antibodies that are actually directed to epitopes topographically distinct from the site of binding of gp110 to CD4.

Effects of calcium ions on proteolytic processing of HIV-1 gp160 precursor and on cell fusion.

Complete activation of human immunodeficiency virus type 1 (HIV-1) requires the endoproteolytic cleavage by cellular protease of the envelope glycoprotein precursor (gp160) into the external glycoprotein gp120, and the transmembrane glycoprotein gp41. We report here the effect of depletion of cellular calcium ions on maturation of precursor gp160 and its concomitant effect on syncytium formation. We show that the cellular endoprotease activity responsible for gp160 maturation and the capacity for HIV-1 to induce syncytium formation are calcium-dependent. In addition, we show that endoproteolytic maturation is a key step in syncytium formation induced by HIV-1.

Primary structure of scorpion anti-insect toxins isolated from the venom of Leiurus quinquestriatus quinquestriatus.

The amino acid sequences of insect-selective scorpion toxins, purified from the venom of Leiurus quinquestriatus quinquestriatus, have been determined by automatic phenyl isothiocyanate degradation of the S-carboxymethylated proteins and derived proteolytic peptides. The excitatory toxin Lqq IT1 and Lqq IT1' (70 residues) show the shift of one half-cystine from an external position, which is characteristic of anti-mammal toxins, to an internal sequence position. Lqq IT2 (61 residues) displays the half-cystine residue in position 12, common to the sequence of all known anti-mammal toxins; it induces flaccid paralysis on insects but is non-toxic for the mouse. Lqq IT2 structurally defines a new type of anti-insect toxins from scorpion venoms. CD spectra and immunological data are in agreement with this finding.

Inhibition of HIV-2(ROD) replication in a lymphoblastoid cell line by the alpha1-antitrypsin Portland variant (alpha1-PDX) and the decRVKRcmk peptide: comparison with HIV-1(LAI).

We investigated the effects of alpha1-antitrypsine Portland variant (alpha1-PDX) and decanoylRVKRchloromethylketone (decRVKRcmk) on HIV-2(ROD) replication in the Jurkat lymphoblastoid cell line. To this end, cells were stably transfected with the alpha1-PDX (J-PDX) and used as targets for HIV-2(ROD) infection. Controls were prepared with the empty vector (J-pcDNA3). HIV-2(ROD) and HIV-1(LAI) replications were significantly inhibited and delayed in the presence of the alpha1-PDX protein. When decRVKRcmk was used at 35 microM, inhibition rates were 70-80% for HIV-2(ROD) and HIV-1(LAI), while total inhibition occurred at 70 microM. Control peptides consisting of decanoylRVKR and acetylYVADcmk had no effect. In the presence of the alpha1-PDX or the decRVKRcmk at 35 microM, the infectivity of HIV-2(ROD) and HIV-1(LAI) produced was 3-4-fold lower. Both molecules inhibited syncytium formation by HIV-2(ROD) and HIV-1(LAI) to a considerable extent. Finally, the inhibition of viral replication was correlated with the ability of the decRVKRcmk at 35 and 70 microM and of the alpha1-PDX, to inhibit the processing of envelope glycoprotein precursors. The alpha1-PDX protein and the decRVKRcmk peptide at 35 microM inhibited HIV-2 and HIV-1 to a similar level suggesting that identical or closely related endoproteases are involved in the maturation of their envelope glycoprotein precursors into surface and transmembrane glycoproteins. The significant inhibition observed with alpha1-PDX indicates that furin or furin-like endoproteases appear to play a major role in the maturation process.

T4-lymphocyte endoprotease responsible for the proteolytic processing of HIV-1 gp160, like Kex2p endoprotease, is a calcium-dependent enzyme.

In the present study we show that precursor gp160 is cleaved in the HIV-1 infected CEM (CD4+) cell line preferentially in the presence of calcium ions demonstrating that the responsible cellular endoprotease is a calcium-dependent enzyme. Taking into account this similarity, a synthetic peptide modelling the cleavage site of HIV-1 envelope glycoprotein precursor was used as substrate for Kex2p. Results obtained clearly showed that the processing enzyme Kex2p (EC 3.4.21.61), a subtilisin-like serine protease that is encoded by the KEX2 gene of yeast Saccharomyces cerevisiae is able to cleave correctly this peptide at the potential cleavage site.

Production and characterization of monoclonal antibodies to simian immunodeficiency virus envelope glycoproteins.

Twelve monoclonal antibodies (MAbs), TB1 to TB12, were produced against a soluble vaccinia recombinant envelope glycoprotein (gp140) from simian immunodeficiency virus SIVmac251. These MAbs recognized SIV gp140 with a relatively high affinity (K0.5 from 6.7 x 10(-8) to 4 x 10(-9) M). All the MAbs except TB9, TB11, and TB12 cross-reacted with HIV-2 envelope glycoproteins, but none of the 12 MAbs recognized those from HIV-1. Using a panel of 87 overlapping synthetic peptides containing 20 amino acid residues, with an overlap of 10 amino acids and spanning the entire primary sequence of gp140, 3 linear epitopes were identified. The first mapped with a neutralizing MAb, TB12, which recognized a linear sequence around amino acids 28-31 within the N-terminal end of the external envelope glycoprotein. The two other new nonneutralizing MAbs recognized linear epitopes around amino acid sequence 380-381 by MAbs TB1, TB2, and TB3, and at the transmembrane glycoprotein amino acids 581-600 by MAb TB6. Seven of the 12 MAbs, TB4, TB5, TB7-9, TB10, and TB11, failed to bind the linear synthetic peptides in ELISA. Moreover, among these seven MAbs only MAbs TB4, TB5, TB9, and TB10 failed to recognize SIV envelope glycoproteins in Western blot (WB) or ELISA after reduction of disulfide bridges by dithiothreitol (DTT), suggesting that they are directed against conformational or discontinuous epitopes. It is of interest to note that MAb TB10 can block the binding of gp140 to the CD4 receptor when the MAb is previously incubated with gp140. Consistent with this result, MAb TB10 cannot bind to gp140 that has been previously complexed with the CD4 receptor. All these results suggest that MAb TB10 recognizes a conformational or discontinuous epitope overlapping or close to the CD4-binding site. These properties are probably implicated in the neutralizing activity observed with this MAb.