Native Epitopes Research Articles

T-cell epitope determination

Synthetic approaches to T-cell epitope determination have recently been developed that complement the search for natural T-cell epitopes and the investigation of the preferences of the different MHC alleles for particular motifs in cognate peptide sequences. The combination of these different strategies opens new possibilities for basic, as well as for applied, immunology. The outlines of the strategies for determination of natural T-cell epitopes are well established. These strategies have contributed substantially to our understanding of the nature of T-cell epitopes and of many diseases. Positional scanning approaches with random synthetic peptide libraries allow comprehensive surveys of the sequence requirements for peptide selection by MHC molecules and for induction of T-cell responses. Synthetic T-cell epitopes can be determined independently of the knowledge of the natural T-cell antigen. This opens new perspectives for the development of synthetic vaccines, TCR antagonists and MHC blockers.

An HLA‐B35‐restricted epitope modified at an anchor residue results in an antagonist peptide

Peptides associated with HLA-B35 commonly have a proline or occasionally a serine residue in the P2 anchor position of the peptide, with a tyrosine at the C terminus. Based on this motif, we identified an octamer epitope from influenza A matrix protein which is presented by HLA-B35. The requirements for MHC binding and T cell receptor contact have been analyzed using analogs of this peptide with substitutions at positions 1, 2, 4, 7 and 8. The natural epitope contains a serine residue at P2 of the peptide. Substitution of this residue with proline (the favored amino acid in this position in B35-associated peptides) considerably enhances binding to HLA-B35 in the T2-B35 cell line, but the peptide is not recognized by the majority of CTL clones and can antagonize recognition of the index peptide. This suggests that a conservative substitution at the P2 anchor position results in a conformational change in the peptide-MHC surface exposed to the T cell receptor.

Use of Intrinsic and Extrinsic Helper Epitopes for In Vivo Induction of Anti-Hepatitis C Virus Cytotoxic T Lymphocytes (CTL) with CTL Epitope Peptide Vaccines

The induction of virus-specific cytotoxic T lymphocytes (CTL) is an important part of vaccine strategy. CTL induction in vivo by two hepatitis C virus (HCV) peptides containing CTL epitopes, one from the NS5 region (P17) and one from the core (C7), was compared. P17 required covalent attachment of a helper peptide (PCLUS3 containing a cluster of epitopes from the human immunodeficiency virus envelope protein), whereas C7 did not. However, the minimal decapeptide of C7, C7A10, alone did not induce CTL. The helper cells induced by PCLUS3-17 or by C7 were shown to be CD4+ and to produce interleukin-2 (IL-2). Thus, help can be supplied by a natural helper epitope intrinsic to the CTL peptide, as in C7, or by attaching a helper epitope from another protein, as in the case of P17. The cluster peptides may be useful promiscuous helper peptides for a variety of CTL epitopes from diverse pathogens.

Mapping epitopes on antigens by immunodiffusion in gel.

The mapping of B-cell epitopes on antigen molecules is best done with monoclonal antibodies (MAbs). Most methods require physical chemical procedures for purifying, binding, or labeling the antigen or the MAb. These procedures require some technical proficiency, may present methodological problems, and are based on much trial and error, but, worst of all, these manipulations, including the simple adsorption of molecules to surfaces, may induce the unfolding of the protein antigen, and therefore the loss of native epitopes with the exposure of new determinants (“unfoldons”) (). Thus, when a mapping technique is successfully set up, there is often the nagging question: “Is the antigen still in its native conformation?” Our immunodiffusion method avoids many of these problems since in most instances, both the antigen and the antibody can be used without any manipulation. The method is based on the fact that a single MAb generates soluble immune complexes with an antigen, but two MAbs to different epitopes, when tested in appropriate conditions, generate insoluble complexes that are detected as a precipitin line.

Effects of major-histocompatibility-complex-encoded subunits on the peptidase and proteolytic activities of human 20S proteasomes. Cleavage of proteins and antigenic peptides.

The proteasome is responsible for the non-lysosomal degradation of misfolded, transient, or ubiquitin-tagged proteins. This fact and the identification of two major-histocompatibility-complex-(MHC)-encoded proteasomal subunits, LMP2/7, suggest an important role of the proteasome in antigen processing. Using purified 20S proteasomes from a wild-type and a LMP2/7-deletion T lymphoblastoid cell line, we analyzed the effect of LMP2/7 on the peptidase and proteolytic activities of the complex in the context of various purification and activation methods. The incorporation of LMP2/7 alters the peptidase activity against fluorogenic substrates, but these effects are not reflected in the time-dependent degradation pattern of oxidized insulin B chain or of peptide epitopes of an antigenic protein. No effect of LMP2/7 on the degradation pattern of these substrates was observed by either reverse-phase chromatography, pool sequencing, or mass spectrometry. The 20S proteasome can cleave insulin B chain at nearly every position, showing that the P1 position alone does not determine the cleavage sites. The maximum of the length distribution of the end products, makes these ideal candidates for MHC display; yet we find that a natural epitope derived from human histone H3 is further degraded by 20S proteasomes. Alanine scans and substitutions with related amino acids of this epitope indicate that, as in insulin B chain, the cleavage sites are not determined by the P1 position alone.

Epitope mapping of anti-HIV and anti-HCV monoclonal antibodies and characterization of epitope mimics using a filamentous phage peptide library

A large filamentous phage library (1× 10 9 clones) displaying random 30-amino-acid (aa) sequences on the N terminus of the pIII coat protein was constructed and characterized. Clones in the library were affinity selected for binding to monoclonal antibodies (mAb) against two viral antigens, the HIV gpl20 protein and the HCV core protein. The obtained aa sequences precisely identified the epitopes recognized by the mAb. Binding of peptide-carrying phages to the Ab was demonstrated by ELISA, Western blot and the surface plasmon resonance (SPR) method. The mAb-specific peptides were transferred via genetic techniques onto the N terminus of Escherichia coli alkaline phosphatase (AP). When fused to the enzyme, the peptides maintained their ability to bind their respective mAb, indicating that the peptides contained the necessary contact residues for binding. The affinity of the peptides was estimated to be 100 nM by SPR. A comparison is presented of the relative affinities of phage-derived peptides to the native viral epitopes also displayed on the AP scaffold. The approach of transferring epitopes from phage to AP for further evaluation should be applicable to many other mAb or receptors

Cytotoxic T lymphocyte lysis inhibited by viable HIV mutants.

Immune evasion by the human immunodeficiency virus (HIV) is unexplained but may involve the mutation of viral antigens. When cytotoxic T lymphocytes engaged CD4-positive cells that were acutely infected with HIV bearing natural variant epitopes in reverse transcriptase, substantial inhibition of specific antiviral lysis was observed. Mutant viruses capable of these transactive effects could facilitate the persistence of a broad range of HIV variants in the face of an active and specific immune response.

The use of dedicated peptide libraries permits the discovery of high affinity binding peptides

The motif for peptide binding to monoclonal antibody mAb A16, which is known to be directed against glycoprotein D of Herpes simplex virus type 1, was determined using two dedicated peptide libraries. As a starting point for this study we used an A-16 binding lead sequence, which had previously been obtained from a phage display peptide library (Schellekens et al., 1994). Binding studies with different length variants of this peptide identified a 12mer as a suitable lead compound for our library study. Two incomplete dedicated resin-bound synthetic peptide libraries were generated. Both consisted of 2 × 10 6 12mers, in which positions were alternately fixed (amino acids identical to the lead sequence) and random. The libraries were screened with mAb A16 and beads with binding peptides were sequenced using Edman degradation. This resulted in a unique peptide binding motif, essentially comprising a 7mer core sequence. Comparison of the sequence of the natural epitope with the binding motif revealed that its sequence was identical to the motif except for one position. Substitution of a methionine in the natural epitope by a tyrosine or a phenylalanine at that position, as dictated by the motif, resulted in a peptide with an affinity for binding to mAb A16 about 50 times higher than that of the natural epitope. Thus, if a lead sequence is available, the use of incomplete, dedicated synthetic peptide libraries provides a fast and powerful tool for the detection of high affinity peptides.

Transgenic mice that express different forms of the influenza virus hemagglutinin as a neo-self-antigen.

We have generated transgenic mouse lineages that express the influenza virus hemagglutinin in different physical forms. One kind expresses the full-length hemagglutinin molecule as a cell surface glycoprotein and can be recognized by hemagglutinin-specific B and T cells. The other expresses a truncated polypeptide corresponding to the N-terminal third of the hemagglutinin molecule. This polypeptide encodes known hemagglutinin-specific T-cell determinants; however, it contains no native B-cell epitopes, since these depend on the conformation of the fully folded protein. In each case, the hemagglutinin transgenic mice display ubiquitous expression of transgenic messenger RNA and induce T-cell tolerance to the transgene-encoded T-cell determinant site 1. Thus, the hemagglutinin is a neo-self-antigen in both kinds of hemagglutinin transgenic mice and should provide a useful system for understanding the factors and mechanisms that govern tolerance and autoimmunity to self-antigens.

Changes in an HER-2 peptide upregulating HLA-A2 expression affect both conformational epitopes and CTL recognition: implications for optimization of antigen presentation and tumor-specific CTL induction.

The HER-2/neu protooncogene (HER-2) is overexpressed in a significant number of breast and ovarian tumors. Peptides of HER-2 sequence were recently found to reconstitute recognition of cytotoxic T lymphocytes (CTLs) from tumor-associated (TALs) and tumor-infiltrating (TILs) lymphocytes, indicating that they reconstitute natural epitopes recognized by CTLs on HLA-A2+ tumors. Because HER-2 is an important antigen (Ag) for tumor-specific CTL induction and the immunogenicity of peptides for CTL induction is dependent on their presentation as stable complexes with HLA-A2, we identified peptides of high and low stabilizing activity from the sequence of HER-2 and the folate-binding protein (FBP). Distinct sequence patterns in the region positions (P)3-P5 and P1 were found for peptides with high (HSA) and low (LSA) stabilizing ability. A low-HLA-A2-affinity HER-2 peptide, P1 of the CTL epitope, was found to be permissive to substitutions that enhanced HLA-A2-stabilizing ability and conserved CTL recognition. In contrast, the region P3-P5 was not permissive to sequence changes. We conclude that the selective permissivity of P1 and P9 in the tumor epitope sequence may have important implications for optimization of tumor Ag presentation, and "neoantigenicity" of self-antigens, aiming toward induction of tumor-reactive CTLs of defined affinity and specificity for target Ags.

Identification of K k-restricted T-cell epitope within influenza virus nucleoprotein

In order to determine the epitope structure in peptide NP50-63, which has been reported to be the only K k-restricted T-cell antigen within the influenza virus (A/PR/8/34) nucleoprotein, a series of 13 peptides truncated from C- and N-termini of NP50-63 were synthesized and their sensitizing activities against K k-restricted nucleoprotein-specific cytotoxic T lymphocytes (CTLs) were examined. One of the 13 peptides, NP50-57, sensitized L929 cells at the nM level, which was 100–1000 times lower in concentration than that at which the other peptides sensitized these cells. The presence of NP50-57 in A/ PR/ 8/34-infected L929 cells was also investigated. Acid extracts of virus-infected cells were separated on a reverse-phase HPLC column and then anion-exchange column. By both separations, only one peak of sensitizing activity against nucleoprotein-specific CTLs was observed. The position of the peak coincided with that of the elution of NP50-57. These results strongly suggest that NP50-57 is the natural epitope in the antigenic structure, NP50-63.

Amino-terminal alteration of the HLA-A*0201-restricted human immunodeficiency virus pol peptide increases complex stability and in vitro immunogenicity.

Initial studies suggested that major histocompatibility complex class I-restricted viral epitopes could be predicted by the presence of particular residues termed anchors. However, recent studies showed that nonanchor positions of the epitopes are also significant for class I binding and recognition by cytotoxic T lymphocytes (CTLs). We investigated if changing nonanchor amino acids could increase class I affinity, complex stability, and T-cell recognition of a natural viral epitope. This concept was tested by using the HLA-A 0201-restricted human immunodeficiency virus type 1 epitope from reverse transcriptase (pol). Position 1 (P1) amino acid substitutions were emphasized because P1 alterations may not alter the T-cell receptor interaction. The peptide with the P1 substitution of tyrosine for isoleucine (I1Y) showed a binding affinity for HLA-A 0201 similar to that of the wild-type pol peptide in a cell lysate assembly assay. Surprisingly, I1Y significantly increased the HLA-A 0201-peptide complex stability at the cell surface. I1Y sensitized HLA-A 0201-expressing target cells for wild-type pol-specific CTL lysis as well as wild-type pol. Peripheral blood lymphocytes from three HLA-A2 HIV-seropositive individuals were stimulated in vitro with I1Y and wild-type pol. I1Y stimulated a higher wild-type pol-specific CTL response than wild-type pol in all three donors. Thus, I1Y may be an "improved" epitope for use as a CTL-based human immunodeficiency virus vaccine component. The design of improved epitopes has important ramifications for prophylaxis and therapeutic vaccine development.

Engineering a peptide epitope display system on filamentous bacteriophage

The genome of bacteriophage fd has been engineered to allow foreign amino acid sequences to be displayed in the exposed N-terminal segment of the major coat protein in the virus particle: small peptides can be encoded directly; larger peptides are encoded in hybrid virions, in which wild-type coat protein subunits are interspersed with coat proteins displaying the foreign peptides. Biophysical techniques, such as X-ray diffraction, indicate that the inclusion of the peptides can be achieved without significant disturbance to the helical parameters that define the protein—protein interactions in the assembled virion and the exposure of the peptides can be verified by analysing the susceptibility to attack by proteolytic enzymes. Peptide sequences from the V3 loop of the surface glycoprotein gp120 of HIV-1 strain MN (HIV-1 MN) displayed in this way are remarkably effective structural mimics of the natural epitope. They are recognised by human HIV antisera and evoke high titres of virus-neutralizing antibodies in mice. Antibody production is stimulated by simultaneous inoculation with T cell epitopes similarly displayed on filamentous bacteriophage. The bacteriophage display system offers a powerful means of studying the immunological recognition of proteins. The specificity of the immune response, the ability to recruit helper T cells, the lack of need for external adjuvants and the structural mimicry of defined peptide epitopes, suggest that it will also be an inexpensive and simple route to the production of effective vaccines.

Engineering a peptide epitope display system on filamentous bacteriophage.

The genome of bacteriophage fd has been engineered to allow foreign amino acid sequences to be displayed in the exposed N-terminal segment of the major coat protein in the virus particle: small peptides can be encoded directly; larger peptides are encoded in hybrid virions, in which wild-type coat protein subunits are interspersed with coat proteins displaying the foreign peptides. Biophysical techniques, such as X-ray diffraction, indicate that the inclusion of the peptides can be achieved without significant disturbance to the helical parameters that define the protein-protein interactions in the assembled virion and the exposure of the peptides can be verified by analysing the susceptibility to attack by proteolytic enzymes. Peptide sequences from the V3 loop of the surface glycoprotein gp120 of HIV-1 strain MN (HIV-1MN) displayed in this way are remarkably effective structural mimics of the natural epitope. They are recognised by human HIV antisera and evoke high titres of virus-neutralizing antibodies in mice. Antibody production is stimulated by simultaneous inoculation with T cell epitopes similarly displayed on filamentous bacteriophage. The bacteriophage display system offers a powerful means of studying the immunological recognition of proteins. The specificity of the immune response, the ability to recruit helper T cells, the lack of need for external adjuvants and the structural mimicry of defined peptide epitopes, suggest that it will also be an inexpensive and simple route to the production of effective vaccines.

Characterization of TAP‐Independent and Brefeldin A‐Resistant Presentation of Sendai Virus Antigen to CD8+ Cytotoxic T Lymphocytes

H-2Kb-transfected T2 cells, which lack both TAP1/2 and LMP2/7 genes, are able to efficiently process and present Sendai virus Antigen to Kb-restricted Sendai virus-specific CTL. This presentation is not inhibited by Brefeldin A (BFA). Here we extend our analysis of this novel antigen presentation pathway. We show that presentation of Sendai virus antigen was not due to sensitization of T2Kb cells by peptides in the virus preparation or peptides released from virus infected cells. Also, the ability to present Sendai virus in a BFA resistant fashion was specific for cells of the T2 lineage. Re-expression of TAP1/2 genes in T2Kb cells did not alter the capability to present antigen in a BFA resistant fashion, i.e. the presence of a functional TAP transporter complex did not relocate (all) peptides to the classical pathway for antigen processing and presentation. We found that co-infection of T2Kb cells with either Sendai virus plus influenza virus or Sendai virus plus VSV did not relocate presentation of influenza or VSV antigen to the TAP independent BFA resistant antigen presentation pathway. Peptide elution experiments and studies with peptide-specific CTL firmly demonstrated that the antigen presented by T2Kb cells after infection with Sendai virus was the natural Sendai virus epitope NP324-332. The same epitope, when expressed as a minigene in vaccinia virus, could be presented also by T2Kb cells but this presentation could be blocked by BFA. Thus, the TAP independent BFA resistant presentation of antigen seem cell (T2 lineage) and virus (Sendai virus) specific, but not epitope specific. The ability of T2Kb cells to present Sendai virus antigen in a TAP independent BFA resistant fashion was only partially blocked by lysosomal inhibitors such as methylamine, ammonium chloride and chloroquine. These findings demonstrate that TAP1/2-independent and BFA-resistant class I processing is only expressed in certain cell types, in parallel with classical MHC class I processing, and that Sendai virus selectively can enter this pathway. Hypothetical models for the TAP-independent class I processing are discussed.

Selection of peptide ligands for the antimucin core antibody C595 using phage display technology: definition of candidate epitopes for a cancer vaccine

Aims-To further define the specificity of the antimucin core antibody C595 by fitting it with a family of hexapeptide ligands by immunoselection of filamentous bacteriophage from a gene III display library of approximately 6.4 x 10(7) random hexapeptides.Methods-Three rounds of immuno-selection were used to enrich for C595 binding phage. DNA sequencing revealed the hexapeptides expressed. Bacteriophage and corresponding synthetic hexapeptides were used in ELISA assay to determine binding affinities.Results-Twenty nine clones from this selected population were analysed. Seven contained the natural epitope RPAP, encoded by two different DNA sequences; 17/29 contained the motif RLPP. In all, 28/29 clones contained the motif RXXP and one clone (RVRPAP) contained the motif RXXP in two peptidic registers; 24/28 clones (6/8 DNA sequences) contained a hydrophobic residue (V or I) at position 1 relative to the RXXP motif. In addition the proximity of RXXP to glycine (position 5) suggests that this contributes in the natural epitope to antibody/antigen binding, which was not detected by chemical synthetic methods. One clone, KSKAGV, bears no obvious relationship to the natural epitope and therefore qualifies as a weakly binding mimotope.Conclusions-This approach has rapidly defined the specificity of this antibody in unprecedented detail, and provides a more comprehensive molecular basis for exploring the immune recognition of the MUC1 mucin by the C595 antibody. Importantly, the novel but related epitopes seen provide peptide specificities and a strategy which may prove useful in generating cancer vaccine candidates.

TheNeisseria meningitidisouter membrane protein P1 produced inBacillus subtilisand reconstituted into phospholipid vesicles elicits antibodies to native P1 epitopes

Class 1 outer membrane protein (P1) ofNeisseria meningitidisgroup B is considered a promising vaccine candidate because P1 subtype–specific antibodies have been shown to be protective in an animal model. We have previously described the production of P1 in the Gram–positiveBacillus subtilisas intracellular inclusion bodies, from which the protein (BacP1) is easily purified (Nurminenet al., Mol. Microbiol., 1992, 2499–2506). We show here that the purified BacP1 can be reconstituted into phospholipid vesicles with the formation of the native immunodominant surface epitopes. The detergent–solubilized, completely denatured BacP1 was fused with phospholipid–detergent micelles during detergent removal by dialysis or gel filtration to yield protein–lipid vesicles (liposomes). When mice were immunized with these liposomes, they produced high titers of antibodies reacting in a P1 subtype–specific manner with meningococcal cells indicating the presence of conformation–dependent P1–specific epitopes in the liposomes. The results suggest that a vaccine candidate for meningococcal disease could be developed from the BacP1–liposomes. They furthermore demonstrate the feasibility of refolding a denatured outer membrane protein, which has never been exposed to lipopolysaccharide, into a native–like conformation.

D b-binding peptides from influenza virus: Effect of non-anchor residues on stability and immunodominance

Relative affinities were determined for the interaction of H-2D b with all the peptides from the A/PR/8/34 strain of influenza virus that contained the D b-binding motif. The results indicated that, even though 23 peptides with the appropriate motif were identified and analysed, binding of only five of them could be detected at peptide concentrations lower than 10 −7 M. Of these five, only one, TGICNQNII, bound with better affinity than the nucleoprotein-derived natural epitope, ASNENMETM. The origin of the higher binding peptide was the influenza neuraminidase, a protein for which little cytosolic processing would be expected since it is a surface glycoprotein. To establish why many of the influenza-derived peptides did not bind, the role of non-anchor residues on D b-peptide interactions was analysed, using a scheme where QDIENEEKI, a non-binding peptide from the influenza virus polymerase 1, was sequentially converted to ASNENMETI, which binds to D b with an affinity similar to that of ASNENMETM. Although all positions examined influenced peptide binding, peptide residue no. 2 (P2) was of particular importance. Therefore, each of the 20 naturally occurring amino acids were inserted at this position to investigate their effects on peptide-MHC interaction. The results indicated that amino acids having side chains with charged or ring structures were deleterious, while non-polar and polar residues were either neutral or facilitated binding to different degrees. Our data also indicated that every residue of the peptide contributes to the stability of the MHC-peptide complex, and the final affinity is dependent on the nature of the amino acids at each position, not just on those at a small number of anchor positions. The results also suggested that increased stability, as indicated by the half-life of the peptide-MHC class I complex, might play an important role in selecting the immunodominant epitope.

Induction of CTL in vivo by major histocompatibility complex class I‐peptide complexes covalently associated on the cell surface

The identification of endogenously produced antigenic peptides presented by MHC class I molecules has opened the way to peptide-based strategies for CTL induction in vivo. Here we demonstrate that the induction in vivo of CTL directed against naturally processed antigens can be triggered by injection of syngeneic cells expressing covalent major histocompatibility complex class I-peptide complexes. In the model system used, the induction of HLA-Cw3 specific cytotoxic T lymphocytes (CTL) in mice by cell surface-associated, covalent H-2Kd (Kd)-Cw3 peptide complexes was investigated. The Kd-restricted Cw3 peptide 170-179 (RYLKNGKETL), which mimics the major natural epitope recognized by Cw3-specific CTL in H-2d mice, was converted to a photoreactive derivative by replacing Arg-170 with N-beta-(4-azidosalicyloyl)-L-2,3-diaminopropionic acid. This peptide derivative was equivalent to the parental Cw3 peptide in terms of binding to Kd molecules and recognition by Cw3-specific CTL clones and could be cross-linked efficiently and selectively to Kd molecules on the surface of Con A-stimulated spleen cells from H-2d mice. Photocross-linking prevented the rapid dissociation of Kd-peptide derivative complexes that takes place under physiological conditions. Cultures of spleen cells or peritoneal exudate cells from mice inoculated i.p. with peptide-pulsed and photocross-linked cells developed a strong CTL response following antigenic stimulation in vitro. The cultured cells efficiently lysed not only target cells sensitized with the Cw3 170-179 peptide but also target cells transfected with the Cw3 gene. Moreover, their TCR preferentially expressed V beta 10 and J alpha pHDS58 segments as well as conserved junctional sequences, as has been observed previously in Cw3-specific CTL responses. In contrast, no Cw3-specific CTL response could be obtained in cultures derived from mice injected with Con A-stimulated spleen cells pulsed with the peptide derivative without photocross-linking.

Characterization of neutralization epitopes in the V2 region of human immunodeficiency virus type 1 gp120: role of glycosylation in the correct folding of the V1/V2 domain

A number of monoclonal antibodies (MAbs) with various levels of neutralizing activity that recognize epitopes in the V1/V2 domain of LAI-related gp120s have been described. These include rodent antibodies directed against linear and conformational epitopes and a chimpanzee MAb, C108G, with extremely potent neutralizing activity directed against a glycan-dependent epitope. A fusion glycoprotein expression system that expressed the isolated V1/V2 domain of gp120 in native form was used to analyze the structural characteristics of these epitopes. A number of MAbs (C108G, G3-4, 684-238, SC258, 11/68b, 38/66a, 38/66c, 38/62c, and CRA3) that did not bind with high affinity to peptides immunoprecipitated a fusion glycoprotein expressing the V1/V2 domain of HXB2 gp120 in the absence of other human immunodeficiency virus sequences, establishing that their epitopes were fully specified within this region. Biochemical analyses indicated that in the majority of V1/V2 fusion molecules only five of the six glycosylation signals in the V1/V2 domain were utilized, and the glycoforms were found to be differentially recognized by particular MAbs. Both C108G and MAbs directed against conformational epitopes reacted with large fractions of the fully glycosylated molecules but with only small fractions of the incompletely glycosylated molecules. Mutational analysis of the V1 and V2 glycosylation signals indicated that in most cases the unutilized site was located either at position 156 or at position 160, suggesting the occurrence of competition for glycan addition at these neighboring positions. Mutation of glycosylation site 160 destroyed the C108G epitope but increased the fraction of the molecules that presented the conformational epitopes, while mutation of the highly conserved glycosylation site at position 156 greatly diminished the expression of the conformational epitopes and increased expression of the C108G epitope. Similar heterogeneity in glycosylation was also observed when the HXB2 V1/V2 fusion glycoprotein was expressed without most of the gp70 carrier protein, and thus, this appeared to be an intrinsic property of the V1/V2 domain. Heterogeneity in expression of conformational and glycan-dependent epitopes was also observed for the natural viral env precursor, gPr160, but not for gp120. These results suggested that the closely spaced glycosylation sites 156 and 160 are often alternatively utilized and that the pattern of glycosylation at these positions affects the formation of the conformational structures needed for both expression of native epitopes in this region and processing of gPr160 to mature env products.