The Simian Immunodeficiency Virus Envelope Glycoprotein Contains Two Epitopes Presented by the Mamu-A*01 Class I Molecule

Abstract

Cytotoxic T lymphocyte (CTL) responses against the simian immunodeficiency virus (SIV) envelope and Gag proteins were monitored in a Mamu-A*01-positive rhesus macaque infected with SIVsmE660. Peripheral blood mononuclear cells (PBMC) cultured with synthetic peptides spanning the entire gp160 and Gag coding region recognized a total of three epitopes. One located in Gag was identified as the previously described Mamu-A*01-restricted p11cC→M epitope (CTPYDINQM). The other two epitopes, designated p15m and p54m, were located in the gp160 envelope protein. Both were nine amino acids in length and were predicted to bind Mamu-A*01 because they contained proline and leucine residues at positions 3 and 9, respectively. Indeed, expression of this class I major histocompatibility complex molecule was required for target cell recognition by envelope-specific CD8⁺ T cells directed against both epitopes. These Mamu-A*01-restricted epitopes in the SIV envelope will be useful for monitoring immune responses in vaccinated or infected animals.

Rhesus macaques infected with the simian immunodeficiency virus (SIV) are a valuable model for studying cellular immune responses against medically important lentiviruses such as human immunodeficiency virus (HIV). CD8⁺ T cells are thought to be important for the control of HIV and SIV infections (2, 12–14), probably by cytotoxic activity and by the production of factors that interfere with the virus replication cycle (15, 17, 18, 20). CD8⁺ T cells recognize viral peptides associated with class I major histocompatibility complex (MHC) molecules on the surfaces of infected cells (21). Characterization of these peptide epitopes and class I MHC restriction elements is an important starting point for understanding how host immunogenetics and immune selection pressure on the viral quasispecies might influence the outcome of infection. Knowledge of class I MHC-restricted epitopes also facilitates vaccine design and monitoring of CD8⁺ T-cell responses in immune animals. Indeed, methods for determining the frequency of virus-specific CD8⁺ T-cell populations in immune individuals rely on predefined epitopes. This is especially true for fluorescein-labeled tetrameric MHC class I-peptide complexes that bind to epitope-specific CD8⁺ T cells or detection of cytokine production by individual cells by using ELISPOT or flow cytometric assays (10).

Relatively few SIV epitopes have been identified, and to date, class I restriction elements encoded by the Mamu (Macaca mulatta) histocompatibility complex have been described for only three of them (11, 16, 19). The Mamu-A and -B class I genes are highly polymorphic, but one molecule designated Mamu-A*01 is expressed by about 25% of all rhesus macaques originating from the Indian subcontinent (7). Epitopes presented by this molecule are therefore of practical value for analysis of SIV-specific CD8⁺ cytotoxic T lymphocyte (CTL) responses. Only one Mamu-A*01-restricted epitope (p11cC→M) has been described previously (1, 11). Located in the SIVmac 251 Gag protein, it has been instrumental in evaluation of immune responses elicited by infection or various candidate vaccines (reviewed in reference 9).

We have analyzed immune responses in rhesus macaques infected with SIV after vaccination with venezuelan equine encephalitis virus (VEE) replicons expressing the SIV envelope and gag proteins. A simultaneous response against three Mamu-A*01-restricted epitopes was observed in the peripheral blood of one animal that was transiently viremic after SIV challenge. These included the previously described p11cC→M gag epitope and two new epitopes in the gp160 envelope glycoprotein. We did not detect CTL activity against any epitopes other than these three, suggesting that the Mamu-A*01 allele is sometimes a dominant factor shaping the immune response against SIV. These envelope epitopes will be useful for probing CTL responses against the SIV envelope.

MATERIALS AND METHODS

Animals.

Rhesus macaques (Macaca mulatta) were vaccinated with VEE replicons expressing either gp160 or Gag of SIVsmH4 and then challenged with the SIVsmE660 strain of that virus. Details of the immunization and virus challenge are to be published elsewhere (2a). The animals were housed and maintained as proscribed by the Institutional Animal Care and Use Committee of the Children’s Research Institute (Columbus, Ohio) and the Guide for the Care and Use of Laboratory Animals (publication no. 82–83, revised 1985) published by the Department of Health and Human Services, National Institutes of Health.

PCR detection of the Mamu-A*01 gene amplified by sequence-specific primers (PCR-SSP).

The presence of the Mamu-A*01 allele was assessed as previously described (7). Briefly, genomic DNA isolated from B lymphoblastoid cell lines was amplified by using Mamu-A*01-specific primers Mamu-A*01F (5′-GACAGCGACGCCGCGAGCCAA-3′) and Mamu-A*01R (5′-GCTGCAGCGTCTCCTTCCCC-3′). Internal control primers 5′ MDRB (5′-GCCTCGAGTGTCCCCCCAGCACGTTTC-3′) and 3′ MDRB (5′-GCAAGCTTTCACCTCGCCGCTG-3′), specific for the conserved second exon of all Mamu-DRB alleles in rhesus macaques, were also included in all reactions. Fifty to one hundred fifty nanograms of genomic DNA was amplified in PCR buffer B (Invitrogen, San Diego, Calif.) containing 2 mM MgCl₂, 2.5 mM (each) of the four deoxyribonucleotide triphosphates, 1.25 U of Taq polymerase (Perkin-Elmer, Foster City, Calif.), 0.8 μM (each) Mamu-A*01-specific primer, and 0.68 μM (each) internal control primer. PCR cycling conditions were as described previously (7). Following PCR, 5 μl of the product was run on a 1% agarose gel.

Peptides.

Sets of overlapping SIV gp160 and Gag peptides (20 amino acids offset by 10 residues) were synthesized by Chiron Mimotopes (Clayton, Australia). Custom peptides used to define minimum optimal epitopes were synthesized by Research Genetics, Huntsville, Ala. All peptides were made by using Fmoc (9-fluorenylmethoxycarbonyl) chemistry and had free-acid COOH termini and free-amine NH₂ termini. SIVsmH4 p55^gag and gp160 amino acid coordinates covered by these peptides were numbered as described previously (8).

Cell lines.

Monkey B lymphoblastoid cell lines (B-LCL) were generated by infection of Ficoll-Hypaque-separated peripheral blood mononuclear cells (PBMC) with herpesvirus papio (isolate 594 X1004, kindly provided by Mark Sharp, Southwest Foundation for Biomedical Research, San Antonio, Tex.) as previously described (3).

CTL cultures.

PBMC (4 × 10⁶) separated on Ficoll-Hypaque gradients were cultured in 2 ml of RPMI medium containing 10% heat-inactivated fetal calf serum, 10 U of recombinant interleukin-2 (IL-2)/ml, and 5% (vol/vol) human T-Stim (Collaborative Research Products). Approximately 10⁶ of these cells were incubated prior to culture with pools of 9 or 10 overlapping contiguous SIV gp160 or gag peptides. The numbers of peptide pools required to span the gp160 and Gag proteins were 8 and 5, respectively. Lymphocytes were sensitized with pools that contained a 10 μM concentration of each individual peptide for 1 h at 37°C and were washed once before their addition to the cultures. After 10 to 14 days, CD8⁺ T cells were enriched from these cultures by using anti-CD8 antibody-coated paramagnetic beads (Dynal Corporation) and then tested for lysis of autologous B lymphoblastoid cell lines as described below.

Long-term CTL lines.

CD8⁺ T cells with SIV-specific CTL activity were cloned by limiting dilution at 1 or 10 cells per well in microtiter tissue culture plates in IL-2-containing RPMI medium. Each well also contained 5 × 10⁴ irradiated (3,000 rads) human PBMC as feeder cells and concanavalin A (ConA) at a concentration of 10 μg/ml. Virus-specific CTL lines were maintained by periodic restimulation with ConA and human feeder cells.

CTL assays.

Peptide-stimulated CD8⁺ CTL were tested for lytic activity against autologous B-LCL targets sensitized with SIV gp160 or Gag antigens. Briefly, B-LCL were cultured for 1 h with 50 μCi of ⁵¹Cr and SIV peptide(s) at a 10 μM concentration. In some experiments, target cells were infected overnight with recombinant vaccinia viruses expressing either the SIV gp160 or Gag proteins before labeling with ⁵¹Cr. After three washes, 5 × 10³ target cells were cocultured in duplicate in 96-well round bottom microtiter plates with various numbers of CD8⁺ T cells. Autologous cold B-LCL targets were also added to the cultures at a cold-to-hot ratio of 50:1. Minimum and maximum release of ⁵¹Cr was assessed by incubating target cells alone in culture medium or 1% NP-40 detergent, respectively. After being cultured for 4 h at 37°C, 50 μl of culture supernatant was harvested into 96-well Lumaplates (Packard) containing a solid Yterrium scintillant, dried overnight, and counted in a Wallac 1450 microbeta counter. Specific ⁵¹Cr release was calculated as follows:

RESULTS AND DISCUSSION

SIV-specific CTL responses were evaluated in a rhesus macaque (VW6) that was immunized with VEE replicons expressing the Gag and gp160 proteins of SIVsmH4 (5) and then challenged with the related SIV smE660 isolate (4, 6). This animal was transiently viremic during the acute phase of infection but was virus negative by the B-DNA assay when SIV-specific CTL activity was analyzed 30 to 45 weeks later. Results of the entire vaccine study are to be published elsewhere (2a). PBMC were assessed for gag-specific CTL after restimulation with five pools of overlapping peptides. Enriched CD8⁺ T cells failed to recognize targets pulsed with any of the five peptide pools, except number 2 (Fig. 1A). One peptide in this pool contained amino acids 181 to 189 (CTPYDINQM), previously described as the Mamu-A*01-restricted p11cC→M epitope (1). Indeed, further mapping studies with pool 2 effector cells revealed that this sequence was recognized (Fig. 1B). A similar analysis was carried out with gp160 envelope peptides. CD8⁺ T cells expanded with peptide pools 2 (amino acids 191 to 300) and 6 (amino acids 591 to 700) had envelope-specific lytic activity, indicating recognition of at least two epitopes (data not shown). There were sufficient effector cells to deconvolute both peptide pools in a two-step process. In the first step, subpools containing three to four peptides each were used to narrow the location of the epitopes to pool 2 peptides 14 to 16 (Fig. 2A) and pool 6 peptides 54 to 56 (Fig. 2C). Individual peptides were identified next. CD8⁺ CTL expanded with gp160 peptide pool 2 recognized peptide 15, representing amino acids 231 to 250 (Fig. 2B), and pool 6 effectors recognized peptide 54, representing amino acids 621 to 640 (Fig. 2D).

FIG. 1.

SIV p55 gag-specific CTL responses. (A) PBMC from animal VW6 were restimulated with one of five pools of SIV Gag peptides. Enriched CD8⁺ T cells were then tested for lysis of autologous B-LCL sensitized with the same peptide pool used for restimulation. Data shown are at an E:T ratio of 40:1 in a 4-h assay. Lysis of targets sensitized with an irrelevant peptide pool was less than 5%. (B) p55 pool 2-restimulated effector cells were tested for lysis of autologous B-LCL that were untreated (⧫) or sensitized with the Mamu-A*01-restricted p11cC→M peptide (●) or two irrelevant control peptides (▴, ■).

FIG. 2.

Identification of gp160 epitopes. PBMC restimulated with gp160 peptide pool 2 (A) and pool 6 (C) were tested for lysis of autologous targets sensitized with three peptide subpools. Individual peptides in each positive subpool were then tested for recognition by pool 2-specific (B) and pool 6-specific (D) CTL. Sequences of the individual peptides are shown in single-letter code.

Recognition of the p11cC→M Gag epitope suggested that macaque VW6 expressed Mamu-A*01 and raised the possibility that one or both of the gp160 epitopes might also be presented by this class I MHC allele. All macaques in this study were therefore typed for Mamu-A*01 by PCR-SSP (7). Two, PE9 and VW6, were positive for the allele (Fig. 3A). A panel of B-cell lines from these animals was sensitized with gp160 peptide 54 and then tested for lysis by an antigen-specific CD8⁺ CTL line. Only the Mamu-A*01-positive target cells (PE9 and VW6) were killed (Fig. 3B). The same pattern of recognition was displayed by the peptide 15-specific CTL line (data not shown), strongly suggesting that both epitopes were presented by Mamu-A*01.

FIG. 3.

Class I MHC restriction of gp160 peptide 54-specific CTL. (A) DNA extracted from B-LCL lines of the indicated macaque was amplified with PCR primers specific for Mamu-A*01 and a conserved region of the class II Mamu-DRB gene. (B) B-LCL from these animals were sensitized with peptide 54 and tested for lysis by an antigen-specific CTL line from animal VW6. The E:T ratio was 40:1 in a 4-h assay.

Peptides binding to the Mamu-A*01 class I MHC molecule are typically 9 or 10 amino acids in length and are anchored by a proline residue at position 3 and a COOH-terminal methionine or leucine (1). Peptides 15 and 54 contained sequences that matched this motif (CAPPGYALL and TVPWPNETL, respectively), and it was therefore predicted that they represented the minimum optimal epitopes. Their presentation by Mamu-A*01 was confirmed by using class I negative 721.221 cells transfected with this class I allele. A peptide 15-specific CTL line killed Mamu-A*01 but not Mamu-B*01-positive target cells pulsed with CAPPGYALL (Fig. 4A). This epitope, which spans amino acids 234 to 242 of the SIV smH4 envelope gp130 ectodomain, is designated p15m. The sequence is conserved in SIVmac251 and SIVmac239 (8) and is also present in many HIV-2 strains belonging to genotypes A and B. The peptide 54-specific CTL line killed Mamu-A*01-transfected target cells sensitized with various concentrations of peptide TVPWPNETL (Fig. 4B). This epitope designated p54m is located between amino acids 626 to 634 of the SIVsmH4 gp41 envelope protein. It is not highly conserved in other SIV strains. These effector cells appeared to be virus type-specific, as target cells sensitized with the SIVmac239 peptide TVPWPNASL were not recognized (Fig. 4B). Nevertheless, the ET→AS substitution at positions 7 and 8 of the epitope is quite conservative, and thus this sequence might also be presented by the Mamu-A*01 class I molecule.

FIG. 4.

Envelope epitopes presented by Mamu-A*01. (A) Peptide 15-specific CTL were tested for lysis of Mamu-A*01- and Mamu-B*01-transfected 721.221 cell lines sensitized with the predicted epitope CAPPGYALL. The E:T ratio was 40:1. (B) Peptide 54-specific CTL were tested for lysis of the predicted epitopes from SIVsmH4 (TVPWPNETL [■]) and mac239 (TVPWPNASL [●]) on Mamu-A*01-transfected 721.221 targets at an E:T ratio of 40:1 over a range of peptide concentrations.

Envelope-specific effector cells used in the previous experiments were expanded from the peripheral blood with peptides and assessed for killing against peptide-sensitized targets. To ensure that they also recognized virus-infected cells, CTL lines specific for p15m and p54m were tested for lysis of Mamu-A*01-transfected 721.221 cells infected with a recombinant vaccinia virus expressing the SIV smH4 gp160 protein. Both effector cells killed the targets pulsed with infected with VVgp160 but not VVgag (Fig. 5).

FIG. 5.

Lysis of target cells expressing SIV gp160. Peptide 54-specific (A) and peptide 15-specific (B) CTL were tested for lysis of Mamu-A*01-transfected 721.221 cells that were sensitized with the relevant peptide or infected with VVgp160 or VVgag.

In summary, we have defined the first Mamu-A*01-restricted epitopes in the envelope glycoprotein of SIV. Both should facilitate studies of CTL activity in this animal model of HIV-1 infection, especially since the Mamu-A*01 molecule is present in 25 to 30% of rhesus macaques. Recombinant tetrameric Mamu-A*01 class I molecules folded around the p11cC→M Gag epitope have already provided new insights into the frequency and localization of antigen-specific CD8⁺ T cells in infected animals (9). Adapting this method of CTL detection to include the two new gp160 epitopes p15m and p54m should be straightforward from a technical standpoint, and these reagents could further our understanding of the evolution of viruses and the immune system in this lentiviral infection. It is not clear why the entire CTL response against the envelope and gag proteins was restricted by Mamu-A*01. It is possible that the animal is homozygous for Mamu-A*01, but the lack of participation by B alleles was striking. The reasons for the dominance of peptides presented by Mamu-A*01 probably merit further investigation.

We cannot exclude the possibility that CTL recognizing other gag or gp160 epitopes were not detected by screening with peptide pools. Peptides in each pool could theoretically compete for binding to a given class I MHC molecule. Minimum optimal peptides of 9 to 12 amino acids in length may also be more efficient than 20 mers for CD8⁺ T-cell stimulation or detection, especially for some alleles such as Mamu-A*01. This approach nonetheless provides a minimum estimate of the number of epitopes recognized by vaccinated or infected animals, which could be an important factor in the outcome of infection.