Single amino acid substitution analogs of the known Mamu A*01 binding peptide gag 181-190 and libraries of naturally occurring sequences of viral or bacterial origin were used to rigorously define the peptide binding motif associated with Mamu A*01 molecules. The presence of S or T in position 2, P in position 3, and hydrophobic or aromatic residues at the C terminus is associated with optimal binding capacity. At each of these positions, additional residues are also tolerated but associated with significant decreases in binding capacity. The presence of at least two preferred and one tolerated residues at the three anchor positions is necessary for good Mamu A*01 binding; optimal ligand size is 8–9 residues. This detailed motif has been used to map potential epitopes from SIVmac239 regulatory proteins and to engineer peptides with increased binding capacity. A total of 13 wild type and 17 analog candidate epitopes were identified. Furthermore, our analysis reveals a significantly lower than expected frequency of epitopes in early regulatory proteins, suggesting a possible evolutionary- and/or immunoselection directed against variants of viral products that contain CTL epitopes.

Simian immunodeficiency virus-infected rhesus macaques represent an important disease model for the study of the pathogenicity of HIV and progression to AIDS. Several independent lines of evidence have implicated cellular immunity in general, and CTL in particular, as playing a crucial role in the control of SIV and HIV infection (for reviews, see Refs. 1, 2, 3). In this context, it is important to develop accurate methods to identify epitopes that bind and are presented to CTL by rhesus macaque class I MHC molecules.

Previous studies using live cell binding assay systems were used to identify SIV epitopes presented by various rhesus macaque class I molecules (4). General motifs crucial for high affinity binding for five different Mamu class I molecules (A*01, A*11, B*03, B*04, and B*17) were also identified (5). The utility of this approach was emphasized by a recent study in which 27 new SIV-derived A*01-restricted epitopes were revealed, based on sequence motif and Mamu A*01 in vitro binding assays using purified MHC molecules.3

In this report, the peptide binding specificity of Mamu A*01 is more rigorously characterized through the analysis of the binding capacity of single substitution analogs of a model Mamu A*01 ligand and of a large library of peptides corresponding to naturally occurring sequences. This analysis has allowed us to define the side chain specificity of both primary and secondary anchor residues.

It was previously shown that accuracy in epitope prediction can be greatly increased by developing detailed peptide binding motifs defining specificity at both primary and secondary anchor positions (6, 7, 8, 9, 10, 11, 12, 13). This knowledge allows for the rational design of optimized ligands. For example, natural sequences carrying suboptimal residues at primary and/or secondary positions can be identified. The suboptimal residues are then replaced with optimal anchors, generating epitopes with increased binding affinity (11, 14, 15). Following this modification, the wild-type peptides that were unable to elicit responses, or were poor immunogens, became highly immunogenic following analoging to increase their MHC binding affinity (14, 15, 16, 17, 18, 19). The CTL induced by such analogs were capable, in most instances, of recognizing target cells expressing wild-type Ag sequences. This phenomenon is likely to reflect less stringent epitope binding requirements for target cell recognition compared with that needed for stimulation of naive T cells to induce differentiation into effectors (20). In this respect, it has been noted that expression of as little as a single molecule of antigenic peptide complexed to class I molecules is sufficient for target cell recognition by effector CTL (21).

SIV encodes three structural retroviral genes (Gag, Pol, and Env), three early genes (Tat, Rev, and Nef), and four accessory genes (Tev, Vpu, Vpr, and Vif). The regulatory and accessory genes are transcribed early in viral replication as multiply spliced mRNAs and appear to be substantial virulence factors critical for the development of AIDS. They are important determinants of viral pathogenicity and are responsible for high titer virus replication in vivo (22, 23, 24, 25, 26, 27, 28, 29, 30). Tat regulates the high level transcription needed to maximize virus production during the short survival time of infected cells (26, 31), whereas Rev activates the export of unspliced RNA required for efficient expression of all viral genes (32). Nef has pleiotropic effects and appears to modulate the cellular membrane proteins that induce CD4 down-regulation (33, 34). Interestingly, infection with Nef-deleted strains of SIV leads to low viral loads, and simian AIDS does not develop in most monkeys. Variants of HIV isolated from long-term nonprogressors are often found with deletions in the Nef gene or defective Nef alleles (35).

Vpu also modulates cellular membrane proteins leading to down-regulation of CD4 (36). Vpu binds CD4 in the endoplasmic reticulum and targets it for proteolysis. The Vif protein must be present in the cells that produce virus, and its absence results in a block of infection soon after viral entry into target cells (37, 38). In addition, Vif may modulate virion assembly (39). HIV-1 and several strains of SIV contain the Vpr accessory gene, whereas HIV-2 and several other strains of SIV contain both Vpr and a highly homologous gene Vpx (40). Vpr appears to augment the importation of uncoated nucleoprotein complexes into the nucleus of the infected cell by interaction with cellular import factors (41). Vpr can also facilitate an increase in virus production by causing the infected cell to delay for extended periods of time in the G2 phase of the cell cycle, where the viral long terminal repeat (LTR) is more active (42). Thus, virus production is maximized in short lived T cells by reducing the time interval between the initial infection and the active production of new virions.

Recent data support the notion that immune responses directed against early SIV regulatory proteins might be extremely effective in controlling viral spread in the initial phases of infection (Ref. 43 ; see also Refs. 25, 44). Recognition of regulatory proteins might also be important to attack hidden latent reservoirs of SIV. Accordingly, these proteins are logical targets for the immune system and should be evaluated as vaccine immunogens. Based on the detailed Mamu A*01 motifs defined herein, we set out to identify natural sequences derived from SIV regulatory protein that could serve as CTL epitopes. Furthermore, we rationally designed a set of “fixed anchor” analog epitopes. Taken together, these studies should enable the design of vaccine constructs designed to elicit powerful CTL responses directed against SIV regulatory proteins.

SIVmac239 protein sequences (accession no. M33262) were analyzed using the text string search software program Motifsearch 1.4 (D. Brown, San Diego, CA) to identify potential peptide sequences containing defined motifs. Peptides were purchased as crude material from Chiron Mimotopes (San Diego, CA) or synthesized at Epimmune using standard tertiary butyloxycarbonyl or fluronylmethyloxycarbonyl solid phase methods as previously described (7). Peptides were resuspended at 4–20 mg/ml in 100% DMSO, then diluted to required concentrations in PBS.

Radiolabeled probe peptides were synthesized at Epimmune on a larger scale using standard tertiary butyloxycarbonyl or fluronylmethyloxycarbonyl solid phase methods. These peptides were subsequently purified to >95% homogeneity by reversed phase HPLC, and their composition ascertained by amino acid analysis, sequencing, and/or mass spectrometry analysis.

721.221 cells transfected with Mamu A*01 cDNA were used as the source of Mamu A*01 molecules. Cells were maintained in vitro by culture in RPMI 1640 medium (Flow Laboratories, McLean, VA) supplemented with 2 mM l-glutamine (Life Technologies, Grand Island, NY), 100 U (100 μg/ml) penicillin-streptomycin solution (Life Technologies), and 10% heat-inactivated FCS (Hazelton Biologics, Lenexa, KS), and grown for large scale cultures in roller bottle apparatuses.

Mamu A*01 was purified from cell lysates using affinity chromatography (45). Briefly, cells were lysed at a concentration of 108 cells/ml in 50 mM Tris-HCL, pH 8.5, containing 1% Nonidet P-40 (Fluka Biochemika, Buchs, Switzerland), 150 mM NaC1, 5 mM EDTA, and 2 mM PMSF. Lysates were then passaged through 0.45-μM filters and cleared of nuclei and debris by centrifugation at 10,000 × g for 20 min, and MHC molecules were purified by affinity chromatography.

For affinity purification, columns of inactivated Sepharose CL4B and protein A-Sepharose were used as precolumns. Mamu A*01 was captured by repeated passage over protein A-Sepharose beads conjugated with the anti-HLA (A, B, C) Ab W6/32 (4). After 2–4 passages the W6/32 column was washed with 10-column volumes of 10 mM Tris-HCL, pH 8.0, with 1% Nonidet P-40, 2-column volumes of PBS, and 2-column volumes of PBS containing 0.4% n-octylglucoside. Finally, Mamu A*01 molecules were eluted with 50 mM diethylamine in 0.15 M NaC1 containing 0.4% n-octylglucoside, pH 11.5. A 1/26 volume of 2.0 M Tris, pH 6.8, was added to the eluate to reduce the pH to ∼8.0. The eluate was then concentrated by centrifugation in Centriprep 30 concentrators at 2000 rpm (Amicon, Beverly, MA). Protein purity, concentration, and effectiveness of depletion steps were monitored by SDS-PAGE.

Quantitative assays for the binding of SIV peptides to soluble Mamu A*01 molecules was based on the inhibition of binding of a radiolabeled standard probe peptide. These assays were performed using the same protocol described for the measurement of peptide binding to HLA class I molecules (45). Briefly, 1–10 nM of radiolabeled probe peptide, a position 1 C→A analog of the SIV Gag 181-190 peptide (ATPYDINQML), was coincubated at room temperature with 1 μM to 1 nM of purified Mamu A*01 in the presence of 1 μM human β2-microglubulin (Scripps Laboratories, San Diego, CA) and a cocktail of protease inhibitors. Following a 2-day incubation period, the percentage of MHC-bound radioactivity was determined by size exclusion gel filtration chromatography on a TSK 2000 column.

In the case of competitive assays, the concentration of peptide yielding 50% inhibition of the binding of the radiolabeled probe peptide was calculated. Peptides were initially tested at one or two high doses. The IC50 of peptides yielding positive inhibition were then determined in subsequent experiments in which two to six further dilutions were tested. Because under the conditions to be used, where [label] < [MHC] and IC50 ≥ [MHC], the measured IC50 values are reasonable approximations of the true Kd values. Each competitor peptide was tested in two to four completely independent experiments, and all different replicate observations were contained in a 4-fold range. As a positive control, in each experiment the unlabeled version of the radiolabeled probe was tested.

For detailed analysis of the peptide binding data, and to allow comparison of data obtained in different experiments, a relative binding value was calculated for each peptide assayed by dividing the IC50 of the positive control for inhibition (SIV Gag 181-190 C1→A: 8.3 nM) by the IC50 for each tested peptide. These values can subsequently be converted back into IC50 nM values by dividing the average IC50 nM of the positive control for inhibition by the relative binding of the peptide of interest. This method of data compilation has proved to be the most accurate and consistent for comparing peptides that have been tested on different days or with different lots of purified MHC. Standardized relative binding values also allow the calculation a geometric mean, or average relative binding (ARB)4 value, for all peptides of a particular characteristic (7, 8, 9, 10, 11, 12, 46).

In analogy to the method described to determine secondary anchor effects influencing the capacity of peptide ligands to bind to HLA class I molecules (7, 8, 9, 10, 11, 12), maps of secondary interactions influencing peptide binding to Mamu A*01 were derived. All peptides of a given size (8, 9, 10, or 11 aa) and with at least one tolerated, and two preferred anchor residues, were selected for analysis. The binding capacity of peptides in each size group was further analyzed by determining the ARB values for peptides that contain specific amino acid residues in specific positions. For determination of the specificity at Mamu A*01 main anchor positions, ARB values were standardized relative to the ARB values of peptides carrying the residue associated with the best binding. For secondary anchor determinations, ARB values were standardized relative to the ARB of the whole peptide set considered. For example, an ARB value was determined for all 9-mer peptides that contain A in position 1, or F in position 7, etc. Because of the rare occurrence of certain amino acids, residues were grouped according to individual chemical similarities as previously described (7, 8, 9, 10, 11, 12, 46).

Previous studies using Edman degradation analyses of sets of pooled peptides eluted from Mamu A*01 indicated a dominant anchor specificity for proline in position 3 (4). Further analysis, using a live cell binding assay and panels of single substitution analogs, also implicated position 2 and the C terminus in determining Mamu A*01 binding capacity. In the present study, the Mamu A*01 binding capacity of the same panels of single substitution analogs was examined with an in vitro binding assay using purified Mamu A*01 molecules.

In the first panel, a nonconservative lysine (K) substitution was introduced at every position of peptide 1279.06, a C1, to A analog of the SIV Gag 181-190 epitope. The results confirmed position 2 and the C terminus, in addition to position 3, as critical for peptide binding (Table I). In each of these cases, substitution of the native peptide residue to K resulted in a 100-fold or greater decrease in binding capacity. Substitutions at other positions were not associated with significant decreases in binding capacity.

Table I.

Binding capacity of a panel of single substitution analogs of a Mamu A*01 binding peptidea

ResidueRelative Binding Capacity
12345678910
ATPYDINQML1.0
Lysine scan          0.29 
          — 
          — 
          0.52 
          0.63 
          0.21 
          0.39 
          0.18 
          0.26 
          — 
            
Substitutions at position 2          0.025 
          0.024 
          0.014 
          — 
          — 
          — 
            
Substitutions at position 3          — 
          — 
          — 
          — 
          — 
          — 
            
Substitutions at C termius          0.56 
          0.46 
          0.35 
          0.010 
          — 
          — 
ResidueRelative Binding Capacity
12345678910
ATPYDINQML1.0
Lysine scan          0.29 
          — 
          — 
          0.52 
          0.63 
          0.21 
          0.39 
          0.18 
          0.26 
          — 
            
Substitutions at position 2          0.025 
          0.024 
          0.014 
          — 
          — 
          — 
            
Substitutions at position 3          — 
          — 
          — 
          — 
          — 
          — 
            
Substitutions at C termius          0.56 
          0.46 
          0.35 
          0.010 
          — 
          — 
a

Substitutions with relative binding capacity >0.1 are considered preferred and are identified by bold type. Substitutions with relative binding capacity in the 0.1–0.01 range are defined as tolerated and are in plain type. Substitutions resulting in >100-fold reduction in binding capacity (relative binding <0.01) are defined as not tolerated and are indicated by a dash.

Based on these results, the specificity of positions 2, 3, and the C terminus was then analyzed in more detail using additional panels of single substitution analogs. In this analysis, as in previous studies of HLA class I molecules (see Ref. 11), preferred anchor residues were defined as those whose binding capacity is 0.1, or better, relative to the binding capacity of the optimal residue. Residues whose binding capacity is between 0.01 and 0.1 were defined as tolerated. Finally, residues whose binding capacity is <0.01, representing a 100-fold reduction in binding capacity, were considered as nontolerated.

It was found that at position 2, small and/or hydrophobic residues A, P, and V were tolerated, with binding capacities 40- to 75-fold lower than the parent peptide (Table I). Larger residues (Q, F, and K) bound either not at all, or at least 100-fold less than the parent peptide. On the basis of this data, and considering the chemical similarity of different amino acid residues, the position 2 specificity of Mamu A*01 was tentatively ascribed to a preference for the small polar residues T and S. Small and/or hydrophobic residues, such as A, P, G, L, I, V, and M, were defined as tolerated.

At position 3, all analogs tested were associated with at least 100-fold decreases in binding capacity (Table I). Interestingly, this was true even for relatively conserved T and A substitutions, or the semiconserved substitution to V. Thus, it was concluded that the position 3 specificity of Mamu A*01 was solely for P.

At the C terminus, analogs carrying aliphatic and/or hydrophobic residues such as I, M, or F (Table I) were associated with binding capacities within 3-fold of the parent peptide, which carries L. The small polar residue T was also tolerated, but was associated with a 100-fold decrease in binding capacity. Polar residues such as Q or K were not tolerated. In summary, and in consideration of chemical similarity, these data suggest a preference at the C-terminal anchor position for aliphatic (L, I, V, and M) and aromatic residues (F, W, and Y), with T also potentially being tolerated.

To define the Mamu A*01 motif in more detail, a library comprised of 714 peptides between 8 and 11 residues in length was tested for Mamu A*01 binding capacity. Each peptide represented a naturally occurring sequence of either viral or bacterial origin and carried residues conforming to the preliminary motif described above in at least two of the three primary anchor positions. The effect of specific amino acid residues was determined by calculating the ARB value (see Materials and Methods) associated with each residue. The percentage of peptides bearing a specific residue at each anchor position that bound Mamu A*01 with an IC50 value of 500 nM or better was also calculated.

To determine in more detail the specificity at position 2, the binding capacity of the library subset of peptides with P in position 3 and aliphatic (L, I, V, M, and T) or aromatic (F, W, and Y) residues at the C terminus was analyzed (Fig. 1). Peptides with S in position 2 were, on average, the best binding peptides. T was also preferred, with an ARB of 0.75, relative to S2 peptides. In the case of both S and T, over 50% of the peptides bearing these residues bound Mamu A*01 with affinities of 500 nM or better. Aromatic (F, W, Y), aliphatic (L, I, V, M, N), or small (A, G, P) residues were tolerated, as defined above, with ARB values in the 0.01–0.1 range. Large polar (Q) or charged residues (D, E, R, K) or C were not tolerated, with ARB values of <0.01.

FIGURE 1.

Position 2 fine specificity of Mamu A*01 ligands. A, Relative Mamu A*01 binding capacity of peptides bearing specific residues in position 2. ARB values were calculated as described in Methods and Methods, and indexed relative to the residue with the highest ARB. In the case of residues for which the number of peptides bearing specific residues was low, residues were grouped on the basis of chemical similarity as previously described (7 ). The average geometric binding capacity of the 518 peptides considered was 5533 nM. B, Graphic summary of the data shown in A.

FIGURE 1.

Position 2 fine specificity of Mamu A*01 ligands. A, Relative Mamu A*01 binding capacity of peptides bearing specific residues in position 2. ARB values were calculated as described in Methods and Methods, and indexed relative to the residue with the highest ARB. In the case of residues for which the number of peptides bearing specific residues was low, residues were grouped on the basis of chemical similarity as previously described (7 ). The average geometric binding capacity of the 518 peptides considered was 5533 nM. B, Graphic summary of the data shown in A.

Close modal

The fine specificity of Mamu A*01 for the C-terminal anchor residue was analyzed next. Based on the single substitution data, a library of peptides with S or T, L, I, V, M, A, G, or P in position 2, and P in position 3 were selected for analysis. The range of residues represented at the C terminus in the resulting set of peptides was limited to those with aromatic (F, W, Y) or aliphatic (L, I, V, M, A, or T) chemical properties, based on the results of single amino acid substitutions (see Table I). Consistent with the single substitution analysis, it was found that F, L, I, and M were preferred, with ARB values in the 0.25–1.0 range (Fig. 2). A preference for V, predicted on the basis of chemical similarity, was also confirmed. The aromatic residues Y and W were tolerated, with ARB values ∼0.070. The small residues T and A were also tolerated, although somewhat less well, with ARB values ∼0.040.

FIGURE 2.

C terminus fine specificity of Mamu A*01 ligands. A, Relative Mamu A*01 binding capacity of peptides bearing specific residues at the C terminus. ARB values were calculated as described in Materials and Methods, and indexed relative to the residue with the highest ARB. The average (geometric) binding capacity of the 520 peptides considered was 5237 nM. B, Graphic summary of the data shown in A.

FIGURE 2.

C terminus fine specificity of Mamu A*01 ligands. A, Relative Mamu A*01 binding capacity of peptides bearing specific residues at the C terminus. ARB values were calculated as described in Materials and Methods, and indexed relative to the residue with the highest ARB. The average (geometric) binding capacity of the 520 peptides considered was 5237 nM. B, Graphic summary of the data shown in A.

Close modal

Finally, to analyze whether P in position 3 is absolutely required even in the case of peptides carrying preferred residues in both position 2 and at the C terminus, a set of peptides with S or T in position 2, and F, L, I, V, or M at the C terminus, was analyzed. As expected, it was found that P in position 3 was indeed the only preferred residue in position 3 (Fig. 3). The small residues T and A were barely tolerated, with relative binding capacities in the 0.01–0.02 range. No other residue was associated with an ARB value of >0.01.

FIGURE 3.

Position 3 fine specificity of Mamu A*01 ligands. A, Relative Mamu A*01 binding capacity of peptides bearing specific residues in position 2. ARB values were calculated as described in Materials and Methods, and indexed relative to the residue with the highest ARB. The average (geometric) binding capacity of the 191 peptides considered was 2286 nM. Except for C, when the number of peptides bearing specific residues was low, residues were grouped on the basis of chemical similarity, as previously described (7 ). B, Graphic summary of the data shown in A.

FIGURE 3.

Position 3 fine specificity of Mamu A*01 ligands. A, Relative Mamu A*01 binding capacity of peptides bearing specific residues in position 2. ARB values were calculated as described in Materials and Methods, and indexed relative to the residue with the highest ARB. The average (geometric) binding capacity of the 191 peptides considered was 2286 nM. Except for C, when the number of peptides bearing specific residues was low, residues were grouped on the basis of chemical similarity, as previously described (7 ). B, Graphic summary of the data shown in A.

Close modal

Taken together, the data in this and the preceding sections indicate an S or T in position 2, P in position 3, and F, L, I, V, or M at the C terminus as preferred residues for Mamu A*01 binding. Small, aliphatic, and aromatic residues are also tolerated in position 2. Position 3 will tolerate small residues (A, T, and perhaps C, on the basis of chemical similarity). Finally, small residues (T and A) and aromatic residues (Y and W) are also tolerated at the C terminus. This motif is summarized in Table II.

Table II.

Mamu A*01 primary anchor motifa

Anchor TypePosition
23C terminus
Preferred ST FLIVM 
Tolerated FWYNAGPLIVM TA[C]b WYTA 
 [CH]   
Anchor TypePosition
23C terminus
Preferred ST FLIVM 
Tolerated FWYNAGPLIVM TA[C]b WYTA 
 [CH]   
a

A valid motif requires the presence of preferred residues in at least two positions and a tolerated residue at the third.

b

Residues in brackets indicate provisional assignments based on chemical similarity.

A library of peptide ligands carrying the Mamu A*01 motif defined in Table II was next analyzed to determine the correlation between peptide length (between 8 and 11 residues) and binding capacity. It was found that 39.4% of the 9-mer peptides and 35.2% of the 8-mer peptides bound with IC50 values of 500 nM or less (Table III). Longer peptides were also capable of binding, although somewhat less well. Specifically, 17.8% of 10-mer and 21.8% of the 11-mer peptides had affinities of 500 nM or better. In conclusion, this data has indicated that the optimal ligand size for Mamu A*01 is 9 residues. Shorter peptides of 8 residues, or longer peptides up to 11 residues in length, are also relatively well tolerated.

Table III.

Optimal ligand size for binding to Mamu A*01

Peptide LengthARBan% Binding Peptides
0.50 88 35.2 
1.0 137 39.4 
10 0.096 118 17.8 
11 0.23 78 21.8 
Total  421 29.2 
Peptide LengthARBan% Binding Peptides
0.50 88 35.2 
1.0 137 39.4 
10 0.096 118 17.8 
11 0.23 78 21.8 
Total  421 29.2 
a

ARB values are standardized relative to the 9-mer data set.

Each peptide sequence in the library was evaluated to determine the number of preferred, tolerated, and nontolerated residues present at position 2, position 3, and the C terminus. Peptides were then grouped as shown in Table IV, and binding data was evaluated as above. As expected, it was found that peptides with preferred residues in each anchor position were overall the best binders, binding more frequently (68.0%) and with higher ARB capacity than other peptides in the library. Peptides with one tolerated anchor (and two preferred anchors) bound with a frequency of (20.8%). The ARB of these peptides was ∼25-fold less than peptides with all preferred main anchor residues. Peptides with more than one tolerated anchor residue, or with at least one nontolerated residue in an anchor position, were found to be poor binders. These peptides bound with ARB values in the 0.0025–0.0050 range, representing a 200- to 400-fold reduction in binding capacity, compared with peptides with three preferred anchors. Correspondingly, only between 3.6 and 5.8% of the peptides were found to have binding affinities of 500 nM or better. In conclusion, optimal binding is obtained, in general, when all three anchor positions carry preferred residues. However, peptides with one tolerated residue and two preferred residues may also be expected to bind with reasonably high frequencies.

Table IV.

Mamu A*01 primary anchor stringency

Anchor ResiduesaARBbn% Binding Peptides
3P 1.0 75 68.0 
2P 1T 0.041 346 20.8 
2P 1N 0.0042 104 5.8 
1P 2T 0.0025 166 3.6 
Total  691 19.5 
Anchor ResiduesaARBbn% Binding Peptides
3P 1.0 75 68.0 
2P 1T 0.041 346 20.8 
2P 1N 0.0042 104 5.8 
1P 2T 0.0025 166 3.6 
Total  691 19.5 
a

Indicates the number of preferred (P), tolerated (T), or nontolerated (N) residues at the position 2, position 3, and C terminus anchors.

b

ARB values are standardized to the peptide set carrying preferred residues in all three anchor positions.

In the next series of analyses, we sought to determine whether secondary effects influencing peptide binding to Mamu A*01 could be detected at positions other than the main anchors. All peptides of a given size (8, 9, 10, or 11) with at least one tolerated and two preferred anchor residues, were selected. The binding capacity of peptides in each size group was further analyzed by determining the ARB values for peptides that contain specific amino acid residues in specific positions. The resulting relative binding values, by corresponding residue/position pairs, for 8- to 11-mer sequences are shown in Tables V-VIII. Summary maps are shown in Fig. 4, A–D.

FIGURE 4.

Summary maps of secondary effects influencing the Mamu A*01 binding capacity of 8-mer (A), 9-mer (B), 10-mer (C), and 11-mer (D) peptides, as described in Tables V-VIII. Shown are the preferred and deleterious residues associated with an ARB capacity of at least 5-fold greater, or 5-fold worse, compared with peptides of the same size carrying other residues.

FIGURE 4.

Summary maps of secondary effects influencing the Mamu A*01 binding capacity of 8-mer (A), 9-mer (B), 10-mer (C), and 11-mer (D) peptides, as described in Tables V-VIII. Shown are the preferred and deleterious residues associated with an ARB capacity of at least 5-fold greater, or 5-fold worse, compared with peptides of the same size carrying other residues.

Close modal

More specifically, it was found that most positions exerted some secondary influence, either positive or negative (Fig. 4). Of note, aromatic residues were often preferred in position 1, whereas negative charges (D, E) and G were often deleterious at the same position. Charged (K, R, H, or D, E) residues were also deleterious toward the middle of the peptide (positions 5 and 6 for 8-mers; 5, 6, and 8 for 9-mers; 4, 5, 6, and 9 for 10-mers; 6 and 9 for 11-mers). Similarly, it was noted that each particular size was associated with unique preferences. For example, the position immediately preceding the C terminus preferred aliphatic residues (L, V, I, or M) in the case of 8-mers, aromatic residues (Y, F, or W) in the case of 9-mers, G in the case of 10-mers, and positive charges (K, H, or R) in the case of 11-mers. Strikingly, P is strongly preferred in position 4 only in the case of 8-mer peptides, suggesting (because P is also an anchor in position 3) that a double P might be important to achieve a favorable conformation for binding for ligands of this size. In conclusion, detailed maps of primary anchor preferences and secondary effects have been defined for Mamu A*01 binding to peptides of 8, 9, 10, or 11 residues in length.

ARB values have been shown to represent effective coefficients for use in designing polynomial algorithms for identifying peptide sequences with a good probability of binding class I molecules with high affinity (12). The basic premise of this predictive methodology is the independent binding of peptide side chains, where the stability contributed by a given residue at a given position is independent of the nature of the residues at other positions. These algorithms take into account both extended and refined motifs (7, 12) and are essentially based on the premise that the overall affinity (ΔG) of peptide-MHC interactions can be approximated as a linear polynomial function of the type ΔG = a1i × a2i × a3i … … × ani where aij is a coefficient that represents the effect of the presence of a given amino acid (j) at a given position (i) along the sequence of a peptide of n amino acids. The crucial assumption of this method is that the effects at each position are essentially independent of each other (i.e., independent binding of individual side chains). When residue j occurs at position i in the peptide, it is assumed to contribute a constant amount ji to the free energy of binding of the peptide, irrespective of the sequence of the rest of the peptide. This assumption is justified by the studies from our laboratories that demonstrated that peptides are bound to MHC and recognized by T cells in essentially an extended conformation.

A method for the derivation of specific algorithm coefficients has been described by Gulukota et al. (Ref. 12 ; see also Refs. 10, 11, 46). Briefly, for all i positions, anchor and nonanchor alike, the geometric mean of the ARB of all peptides carrying j is calculated relative to the remainder of the group and used as the estimate of ji. To calculate the algorithm score of a given peptide in a test set, the geometric mean of all ARB values corresponding to the sequence of the peptide is calculated. If the resulting score exceeds a chosen threshold, the peptide is predicted to bind. Appropriate thresholds can be chosen as a function of the degree of stringency of prediction desired. For example, algorithm scores that allowed prediction of 90 or 75% of the binders in the data set analyzed are shown in Table IX.

Table IX.

Algorithm cut-off scores

Peptide SizeFraction of Possible Binders
90%75%
ScoreaEfficiencybScoreEfficiency
0.69 77.8 0.82 88.9 
0.63 68.1 0.73 75.9 
10 0.78 50.0 0.89 69.6 
11 1.02 72.7 1.13 76.5 
Average  67.2  77.7 
Peptide SizeFraction of Possible Binders
90%75%
ScoreaEfficiencybScoreEfficiency
0.69 77.8 0.82 88.9 
0.63 68.1 0.73 75.9 
10 0.78 50.0 0.89 69.6 
11 1.02 72.7 1.13 76.5 
Average  67.2  77.7 
a

Indicates the algorithm score that allows selection of a set of peptides that includes 90% (or 75%) of the peptides in the data set that bound Mamu A*01 with an IC50 value of 500 nM or better.

b

Indicates the percentage of peptides in the set identified by the algorithm score that bind Mamu A*01 with an IC50 value of 500 nM or better.

To identify peptide ligands derived from SIV proteins that could represent candidates for use in an SIV vaccine, we next scanned sequences of SIVmac239 early and late regulatory proteins for the presence of peptides with the broadly defined Mamu A*01 binding motif (Table II). Of the 51 wild-type sequences identified, 12 were found to have the capacity to bind Mamu A*01 with an IC50 value of 500 nM or less (Table X). Four of these peptides are conserved between SIVmac239 and SIVmac251, and eight others are unique to SIVmac239. Overall, 23.5% of the motif-positive peptides were Mamu A*01 binders. When the same SIV proteins were analyzed using the algorithms described above, nine sequences were identified that were predicted to be Mamu A*01 binders on the basis of scores greater than or equal to the 75% cut-off criteria. Seven (78%) of these peptides, representing 58% of all binders, were found to bind Mamu A*01 with affinities of 500 nM or better.

Table X.

Mamu A*01 binding capacity of peptides derived from SIV mac239 regulatory proteinsa

PeptideSequenceBinding Capacity (IC50 nM)
Vif 100–109 VTPNYADILL 2.6 
Vif 100–108 VTPNYADIL 12 
Vpx 62–72 MSPSYVKYRYL 17 
Vpx 102–111 GPPPPPPPGL 23 
Tat 28–35 STPESANL 37 
Rev 86–95 DPPTNTPEAL 39 
Vpx 39–48 HLPRELIFQV 50 
Vif 75–82 LTPEKGWL 61 
Tat 27–35 ASTPESANL 108 
Nef 41–50 QSPGGLDKGL 148 
Vif 14–22 RIPERLERW 170 
Vif 100–107 VTPNYADI 230 
Vpx 8–18 IPPGNSGEETI 241 
PeptideSequenceBinding Capacity (IC50 nM)
Vif 100–109 VTPNYADILL 2.6 
Vif 100–108 VTPNYADIL 12 
Vpx 62–72 MSPSYVKYRYL 17 
Vpx 102–111 GPPPPPPPGL 23 
Tat 28–35 STPESANL 37 
Rev 86–95 DPPTNTPEAL 39 
Vpx 39–48 HLPRELIFQV 50 
Vif 75–82 LTPEKGWL 61 
Tat 27–35 ASTPESANL 108 
Nef 41–50 QSPGGLDKGL 148 
Vif 14–22 RIPERLERW 170 
Vif 100–107 VTPNYADI 230 
Vpx 8–18 IPPGNSGEETI 241 
a

Bold indicates newly identified Mamu A*01 binder.

An additional peptide, Vif 14, which does not strictly conform to the motif definition of Table II, had previously been identified as a Mamu A*01 binder and is also shown in Table IX. However, it should be noted that Vif14 was the only peptide among a panel of 39 SIVmac239 peptides with motifs not described by Table II that bound Mamu A*01 with IC50 values of 500 nM or better. In conclusion, the motif described herein allowed the identification of 12 peptides derived from SIVmac239 regulatory proteins that have the capacity to bind Mamu A*01 with high affinity. An additional nonmotif-bearing peptide with high affinity binding activity was also identified.

In the same series of experiments, an additional 20 wild-type peptides were identified that bound weakly, in the 500–30,000 nM range. Nineteen of the weak binders were associated with suboptimal motifs, carrying a tolerated residue at one of the three main anchor positions (Table XI). For each suboptimal sequence, an analog peptide was engineered by replacing tolerated anchor residues with a preferred residue. For this study, tolerated residues in position 2, position 3, and at the C terminus were analogued to S, P, and L, respectively.

Table XI.

Fixed anchor analogs allow identification of additional peptides with the capacity to bind Mamu A*01 with high affinity

PeptideSequenceBinding Capacity (IC50 nM)
Vpr 91–101 LSAIPPSRSML 557 
Analog LSPIPPSRSML 20 
Vpr 40–47 LTALGNHI 647 
Analog LTPLGNHI 12 
Vpr 91–100 LSAIPPSRSM 725 
Analog LSPIPPSRSM 124 
Vif 144–153 QVPSLQYLAL 784 
Analog QSPSLQYLAL 349 
Vpx 62–71 MSPSYVKYRY 1378 
Analog MSPSYVKYRL 92 
Vpr 93–101 AIPPSRSML 1529 
Analog ASPPSRSML 8.0 
Vif 144–151 QVPSLQYL 2519 
Analog QSPSLQYL 170 
Vif 136–145 RFPRAHKYQV 2733 
Analog RSPRAHKYQV 5.8 
Rev 62–71 DPPTDTPLDL 5330 
Analog DSPTDTPLDL — 
Vif 113–122 YFPCFTAGEV 5359 
Analog YSPCFTAGEV 89 
Vif 110–117 HSTYFPCF 6210 
Analog HSPYFPCF 207 
Vpx 62–69 MSPSYVKY 6381 
Analog MSPSYVKL 75 
Vif 75–84 LTPEKGWLST 6827 
Analog LTPEKGWLSL 52 
Vpr 94–101 IPPSRSML 7576 
Analog ISPSRSML 93 
Vif 82–89 LSTYAVRI 8717 
Analog LSPYAVRI 37 
Vpr 93–100 AIPPSRSM 9371 
Analog ASPPSRSM 68 
Vif 75–85 LTPEKGWLSTY 13929 
Analog LTPEKGWLSTL 71 
Vif 195–203 KPPTKGANF 17340 
Analog KSPTKGANF 32 
Vif 117–126 FTAGEVRRAI 21189 
Analog FTPGEVRRAI 187 
PeptideSequenceBinding Capacity (IC50 nM)
Vpr 91–101 LSAIPPSRSML 557 
Analog LSPIPPSRSML 20 
Vpr 40–47 LTALGNHI 647 
Analog LTPLGNHI 12 
Vpr 91–100 LSAIPPSRSM 725 
Analog LSPIPPSRSM 124 
Vif 144–153 QVPSLQYLAL 784 
Analog QSPSLQYLAL 349 
Vpx 62–71 MSPSYVKYRY 1378 
Analog MSPSYVKYRL 92 
Vpr 93–101 AIPPSRSML 1529 
Analog ASPPSRSML 8.0 
Vif 144–151 QVPSLQYL 2519 
Analog QSPSLQYL 170 
Vif 136–145 RFPRAHKYQV 2733 
Analog RSPRAHKYQV 5.8 
Rev 62–71 DPPTDTPLDL 5330 
Analog DSPTDTPLDL — 
Vif 113–122 YFPCFTAGEV 5359 
Analog YSPCFTAGEV 89 
Vif 110–117 HSTYFPCF 6210 
Analog HSPYFPCF 207 
Vpx 62–69 MSPSYVKY 6381 
Analog MSPSYVKL 75 
Vif 75–84 LTPEKGWLST 6827 
Analog LTPEKGWLSL 52 
Vpr 94–101 IPPSRSML 7576 
Analog ISPSRSML 93 
Vif 82–89 LSTYAVRI 8717 
Analog LSPYAVRI 37 
Vpr 93–100 AIPPSRSM 9371 
Analog ASPPSRSM 68 
Vif 75–85 LTPEKGWLSTY 13929 
Analog LTPEKGWLSTL 71 
Vif 195–203 KPPTKGANF 17340 
Analog KSPTKGANF 32 
Vif 117–126 FTAGEVRRAI 21189 
Analog FTPGEVRRAI 187 

The resulting 19 analogs were tested for their capacity to bind Mamu A*01. Seventeen of them were found to bind with an IC50 of 500 nM or better and to have a binding capacity 5-fold greater than that of the wild-type peptide (Table XI). This corresponds to an 89% success rate in engineering peptides with significantly increased binding capacity. It was noted that substitutions at any of the three anchor positions were equally effective in generating higher affinity binding analogs.

When the protein of origin of the peptides described in the preceding sections was examined, a bias toward late regulatory proteins was observed (Table XII; see also Tables IX and X). A total of 15 Vif-, 6 Vpx-, and 6 Vpr-derived peptides were identified that bound with good (IC50 ≤ 500 nM) or weak (IC50 = 500–3000 nM) affinity. By comparison, only one Rev-, one Nef-, and two (but overlapping) Tat-derived epitopes were identified. This bias is not related to the different size of the proteins considered. As shown in Table XII, Vif carried 7.0 binders/100 residues, Vpx 5.4, and Vpr 5.9 (average of 6.1 ± 0.8). By comparison, the early regulatory proteins Tat, Rev, and Nef carried 1.5, 1.9, and 1.1 binders/100 residues (average of 1.5 ± 0.4), respectively. This difference in frequency of binders is significant beyond p = 0.001, as determined in a two-sample Z test.

Table XII.

Incidence of Mamu A*01 binders in SIV mac239-derived regulatory proteins

Regulatory Protein TypeProteinNo. of ResiduesNo. of BindersTotal Binders/100 Residuesa
GoodWeakTotal
Early Rev 107 1.9 
 Tat 130 1.5 
 Nef 92 1.1 
Late Vif 214 10 15 7.0 
 Vpx 112 5.4 
 Vpr 101 5.9 
Regulatory Protein TypeProteinNo. of ResiduesNo. of BindersTotal Binders/100 Residuesa
GoodWeakTotal
Early Rev 107 1.9 
 Tat 130 1.5 
 Nef 92 1.1 
Late Vif 214 10 15 7.0 
 Vpx 112 5.4 
 Vpr 101 5.9 
a

The difference between early and late proteins in the occurrence of binding peptides is significant beyond p = 0.001, as determined using a two-sample Z test.

The analysis presented herein has defined the molecular determinants involved in peptide binding to Mamu A*01. In accordance with previous data (4), three main anchor residues (positions 2, 3, and the C terminus) have been found to play an important role in binding. The specificity for three anchors is somewhat unique, as most class I molecules are associated with only two main anchors, typically located at position 2 and the C terminus. The only other known case of a molecule associated with three main anchor residues is the HLA molecule A*01, which is characterized by anchors with spacing identical with Mamu A*01, but using somewhat different amino acids (9).

To more rigorously define the peptide binding specificity of Mamu A*01, the single substitution analysis was refined using a large peptide library. The use of large libraries allows the analysis of the effect of single residues on binding capacity in the context of multiple peptide sequences. This avoids potential biases and limitations presented by the use of single substitution analogs, where the effect of a specific residue is determined in the context of only a single amino acid sequence. Furthermore, the approach used herein offers an effective means of simultaneously probing for effects at both primary and secondary anchor positions. This method relies on the assumption that each position contributes independently to binding affinity. The validity of this assumption is supported by previous studies that compared neural network- and polynomial matrix-based methods for their success in predicting class I MHC binding capacity (12).

The preferred peptide size for Mamu A*01 binding was found to be 8 or 9 residues, although peptides of 10 or 11 residues could also bind with appreciable frequencies. This type of size specificity is somewhat different from what was previously noted in the case of most HLA class I molecules (47, 48, 49), which appear to prefer ligands of 9–10 residues in size. Future studies will be needed to determine whether this finding can be extended as a generalization for Mamu class I alleles. Interestingly, the one and only optimal epitope thus far identified for Mamu A*11 is indeed an 8-mer (5, 50). Additional analyses are necessary to determine whether shorter (7 residues), or longer (12 residues or above) peptides also bind Mamu A*01 with appreciable frequency.

This study represents the first in-depth analysis of secondary influences on peptide binding to rhesus macaque-derived class I molecules. Typically, for class I molecules, primary anchor residues are necessary for peptide binding (7, 51). However, other residues can act as secondary anchors providing supplemental binding energy. Alternatively, the presence of certain residues in specific positions may also have a negative effect on peptide binding capacity. Herein we demonstrate that the same type of influence is also apparent in the case of Mamu A*01 molecules. The fact that different secondary effects were associated with different ligand sizes explains why secondary anchor determination by sequencing of heterogeneous pools of acid-eluted natural ligands has thus far proved to be of limited success.

Primate lentiviruses may have evolved to exploit several viral and immunologic mechanisms that escape or weaken virus-specific immune responses. Indeed, despite vigorous virus-specific immune responses, HIV is able to establish chronic infection in most cases. Effective prophylactic vaccines against SIV or HIV may need to be targeted to proteins active very early in the course of infection (43). In the case of therapeutic vaccines, it might be crucial to target late regulatory proteins, involved in immune evasion and in maintaining reservoirs of latent infection.

Because of their small size, these proteins contain relatively few binding peptides. However, by virtue of refined motifs, we have been able to identify four peptides derived from early, and nine from late, SIV regulatory proteins that bind Mamu A*01 with high affinity. In addition, we have also been able to design 17 additional potential epitopes, by replacing suboptimal anchor residues with optimal ones. The identification of these epitopes is important given the very small number of known epitopes derived from the early and late regulatory proteins. Thus, this study allows the design of specific vaccine constructs targeting early and late SIV regulatory proteins. In this light, we have recently described the engineering and testing of experimental multiepitope MiniGene vaccines in HLA-transgenic animals (52) and rhesus macaques (43).3

Finally, our exhaustive analysis revealed a striking difference in the number of Mamu A*01 binding epitopes contained in early, compared with late, regulatory SIV proteins. A possible explanation of this observation is that these different frequencies are the result of selective pressure from immune responses, resulting in viral escape and reduced numbers of class I binding peptides. These results are also in agreement with a recent study by Allen et al. that underlined the crucial role of Tat epitopes in controlling acute infection, and demonstrated immune pressure and viral escape in the case of the Mamu A*01-restricted Tat 27 epitope but not in the case of other Vif- or Vpx-derived epitopes (43).

Alternative explanations should also be considered. It is possible, for example, that this observation might reflect selection for structural reasons against proline residues in these early regulatory proteins. This explanation appears unlikely in light of the fact that no significant difference exists in the overall frequency of proline in early (7.3 ± 4.2) or late (8.2 ± 2.8) regulatory proteins or in structural proteins (5.2 ± 1.9). It should also be noted that SIVmac239 was derived after serial passage in macaques of a virus originally derived from sooty mangabeys. Thus, these results might reflect selection in sooty mangabeys with alleles binding a motif similar to Mamu A*01. In this light, it is intriguing to point out that in previous studies (53) one of the original epitopes recognized by sooty mangabeys carried a proline in position 3 and isoleucine in position 9.

These results are also reminiscent of earlier studies by Berzofsky and associates (54) that had suggested a lower number of class I motif-positive peptides as one of the structural features of HIV-1. However, actual binding capacity of the motif-containing peptides was not measured. This might explain why, in that case, differences among early and late regulatory proteins were not reported. In future studies, analysis of regulatory proteins might be expanded to other class I and class II specificities of both human and rhesus macaque origin. A similar analysis of the frequency of binding peptides in the structural proteins Gag, Pol, and Env could also be performed. Additional studies addressing viral selection have been presented by other groups (55, 56).

In conclusion, our experiments provide an in-depth look into the interactions between peptide ligands and Mamu A*01. Primary and secondary anchor maps allowed for efficient identification of peptides binding Mamu A*01 and allowed for the design of ligands with enhanced binding capacity. Using the simple primary anchor motif, ∼25% of the predicted peptides bound with high affinity, whereas with the improved algorithm ∼75% of the predicted peptides bound well. The converse study, to determine how many good binders from all possible 8- to 11-mer peptides encoded SIV regulatory genes are effectively predicted, was not performed. Previous studies analyzing HLA class I binding and human papillomavirus E6 and E7 proteins indicated that over 90% of the binders identified carry the appropriate motif. We have applied this knowledge to the identification and engineering of epitopes derived from SIV regulatory proteins, thus enabling the design of vaccine constructs aimed at focusing immune response directed against these Ags.

Table V.

Average relative binding values of different residues at nonanchor positions of 8-mer peptides

Residue (8-mer peptides)Position (ARB)a
12345678
2.1 0.048 0.013 1.2 8.0 7.9 1.3 0.041 
3.0 0.0036 0.0016 0.64 1.2 4.5 0.19 0.0010 
0.043 0.0010 0.0010 0.66 0.15 0.093 0.36 0.0010 
0.043 0.0010 0.0010 0.66 0.15 0.093 0.36 0.0010 
5.9 0.031 0.0010 0.12 1.2 3.8 1.5 1.0 
0.17 0.024 0.0010 1.8 12 0.13 0.072 0.0010 
1.3 0.00055 0.0010 0.41 0.14 0.66 1.6 0.0010 
1.6 0.014 0.0010 0.46 1.1 2.1 5.6 0.33 
1.3 0.0010 0.0010 0.41 0.14 0.66 1.6 0.0010 
1.6 0.018 0.0010 0.46 1.1 2.1 5.6 0.78 
1.6 0.030 0.0010 0.46 1.1 2.1 5.6 0.27 
0.12 0.047 0.0010 0.74 0.22 0.42 0.38 0.0010 
0.27 0.021 1.0 30 2.7 0.40 1.1 0.0010 
0.12 0.0010 0.0010 0.74 0.22 0.42 0.38 0.0010 
1.3 0.0010 0.0010 0.41 0.14 0.66 1.6 0.0010 
3.0 1.0 0.0010 0.64 1.2 4.5 0.19 0.0010 
3.0 0.75 0.018 0.64 1.2 4.5 0.19 0.040 
1.6 0.072 0.0010 0.46 1.1 2.1 5.6 0.28 
5.9 0.031 0.0010 0.12 1.2 3.8 1.5 0.067 
5.9 0.031 0.0010 0.12 1.2 3.8 1.5 0.075 
Residue (8-mer peptides)Position (ARB)a
12345678
2.1 0.048 0.013 1.2 8.0 7.9 1.3 0.041 
3.0 0.0036 0.0016 0.64 1.2 4.5 0.19 0.0010 
0.043 0.0010 0.0010 0.66 0.15 0.093 0.36 0.0010 
0.043 0.0010 0.0010 0.66 0.15 0.093 0.36 0.0010 
5.9 0.031 0.0010 0.12 1.2 3.8 1.5 1.0 
0.17 0.024 0.0010 1.8 12 0.13 0.072 0.0010 
1.3 0.00055 0.0010 0.41 0.14 0.66 1.6 0.0010 
1.6 0.014 0.0010 0.46 1.1 2.1 5.6 0.33 
1.3 0.0010 0.0010 0.41 0.14 0.66 1.6 0.0010 
1.6 0.018 0.0010 0.46 1.1 2.1 5.6 0.78 
1.6 0.030 0.0010 0.46 1.1 2.1 5.6 0.27 
0.12 0.047 0.0010 0.74 0.22 0.42 0.38 0.0010 
0.27 0.021 1.0 30 2.7 0.40 1.1 0.0010 
0.12 0.0010 0.0010 0.74 0.22 0.42 0.38 0.0010 
1.3 0.0010 0.0010 0.41 0.14 0.66 1.6 0.0010 
3.0 1.0 0.0010 0.64 1.2 4.5 0.19 0.0010 
3.0 0.75 0.018 0.64 1.2 4.5 0.19 0.040 
1.6 0.072 0.0010 0.46 1.1 2.1 5.6 0.28 
5.9 0.031 0.0010 0.12 1.2 3.8 1.5 0.067 
5.9 0.031 0.0010 0.12 1.2 3.8 1.5 0.075 
a

A panel of 88 8-mer peptides based on naturally occurring sequences from various viral, bacterial, or pathogen origin was analyzed. All peptides had, minimally, two preferred and one tolerated residue at the main anchor positions, as described in Table II. ARB values shown were calculated as described in Materials and Methods, and are based on the grouping of chemically similar residues (see e.g., 7 ). Values corresponding to a 5-fold or greater increase in binding capacity are indicated in bold. Values corresponding to a 5-fold or greater reduction in binding capacity are indicated by italics. Main anchor positions are positions 2, 3, and 8 and residues determined to be preferred or tolerated anchors are indicated in bold. ARB values at the anchor positions were derived from the analyses described in Figs. 1–3. To filter out nonmotif peptides, the values for nontolerated anchor residues have been set to 0.001, equivalent to a 1000-fold reduction in binding capacity. The average geometric binding capacity of the panel was 1858 nM.

Table VI.

Average relative binding values of different residues at nonanchor positions of 9-mer peptides

Residue (9-mer peptides)Position (ARB)a
123456789
0.21 0.048 0.013 3.6 6.9 6.0 2.2 1.2 0.041 
4.9 0.0036 0.0016 2.1 0.56 0.40 1.7 0.36 0.0010 
0.034 0.0010 0.0010 1.2 1.5 0.025 0.28 1.6 0.0010 
0.034 0.0010 0.0010 1.2 1.5 0.025 0.28 1.6 0.0010 
2.3 0.031 0.0010 1.1 2.3 9.6 2.3 6.8 1.0 
0.15 0.024 0.0010 0.71 1.2 0.83 0.22 0.16 0.0010 
1.8 0.0055 0.0010 0.86 0.16 0.39 0.79 0.20 0.0010 
2.2 0.014 0.0010 0.70 2.4 3.3 0.78 2.5 0.33 
1.8 0.0010 0.0010 0.86 0.16 0.39 0.79 0.20 0.0010 
2.2 0.018 0.0010 0.70 2.4 3.3 0.78 2.5 0.78 
2.2 0.030 0.0010 0.70 2.4 3.3 0.78 2.5 0.27 
0.27 0.047 0.0010 0.12 0.21 0.66 1.8 0.65 0.0010 
0.44 0.021 1.0 2.5 4.0 0.29 1.6 2.3 0.0010 
0.27 0.0010 0.0010 0.12 0.21 0.66 1.8 0.65 0.0010 
1.8 0.0010 0.0010 0.86 0.16 0.39 0.79 0.20 0.0010 
4.9 1.0 0.0010 2.1 0.56 0.40 1.7 0.36 0.0010 
4.9 0.75 0.018 2.1 0.56 0.40 1.7 0.36 0.040 
2.2 0.072 0.0010 0.70 2.4 3.3 0.78 2.5 0.28 
2.3 0.031 0.0010 1.1 2.3 9.6 2.3 6.8 0.067 
2.3 0.031 0.0010 1.1 2.3 9.6 2.3 6.8 0.075 
Residue (9-mer peptides)Position (ARB)a
123456789
0.21 0.048 0.013 3.6 6.9 6.0 2.2 1.2 0.041 
4.9 0.0036 0.0016 2.1 0.56 0.40 1.7 0.36 0.0010 
0.034 0.0010 0.0010 1.2 1.5 0.025 0.28 1.6 0.0010 
0.034 0.0010 0.0010 1.2 1.5 0.025 0.28 1.6 0.0010 
2.3 0.031 0.0010 1.1 2.3 9.6 2.3 6.8 1.0 
0.15 0.024 0.0010 0.71 1.2 0.83 0.22 0.16 0.0010 
1.8 0.0055 0.0010 0.86 0.16 0.39 0.79 0.20 0.0010 
2.2 0.014 0.0010 0.70 2.4 3.3 0.78 2.5 0.33 
1.8 0.0010 0.0010 0.86 0.16 0.39 0.79 0.20 0.0010 
2.2 0.018 0.0010 0.70 2.4 3.3 0.78 2.5 0.78 
2.2 0.030 0.0010 0.70 2.4 3.3 0.78 2.5 0.27 
0.27 0.047 0.0010 0.12 0.21 0.66 1.8 0.65 0.0010 
0.44 0.021 1.0 2.5 4.0 0.29 1.6 2.3 0.0010 
0.27 0.0010 0.0010 0.12 0.21 0.66 1.8 0.65 0.0010 
1.8 0.0010 0.0010 0.86 0.16 0.39 0.79 0.20 0.0010 
4.9 1.0 0.0010 2.1 0.56 0.40 1.7 0.36 0.0010 
4.9 0.75 0.018 2.1 0.56 0.40 1.7 0.36 0.040 
2.2 0.072 0.0010 0.70 2.4 3.3 0.78 2.5 0.28 
2.3 0.031 0.0010 1.1 2.3 9.6 2.3 6.8 0.067 
2.3 0.031 0.0010 1.1 2.3 9.6 2.3 6.8 0.075 
a

A panel of 137 9-mer peptides derived from naturally occurring sequences of various viral, bacterial, or pathogen origin was analyzed. All peptides had, minimally, two preferred and one tolerated residue at the main anchor positions, as described in Table II. ARB values shown were calculated as described in Materials and Methods, and are based on the grouping of chemically similar residues (see e.g., 7 ). Values corresponding to a 5-fold or greater increase in binding capacity are indicated in bold. Values corresponding to a 5-fold or greater reduction in binding capacity are indicated by italics. Main anchor positions are positions 2, 3, and 9, and residues determined to be preferred or tolerated anchors are indicated in bold. ARB values at the anchor positions were derived from the analyses described in Fig. 1–3. To filter out nonmotif peptides, the values for nontolerated anchor residues have been set to 0.001, equivalent to a 1000-fold reduction in binding capacity. The average geometric binding capacity of the panel of peptides analyzed was 1256 nM.

Table VII.

Average relative binding values of different residues at nonanchor positions of 10-mer peptides

Residue (10-mer peptides)Position (ARB)a
12345678910
2.2 0.048 0.013 0.99 5.4 3.7 0.42 0.23 4.4 0.041 
1.2 0.0036 0.0016 1.9 0.71 2.4 0.92 0.62 2.0 0.0010 
0.60 0.0010 0.0010 0.11 0.56 0.19 0.81 2.5 0.18 0.0010 
0.60 0.0010 0.0010 0.11 0.56 0.19 0.81 2.5 0.18 0.0010 
5.3 0.031 0.0010 1.9 1.2 1.4 0.48 2.0 1.1 1.0 
0.40 0.024 0.0010 4.8 0.76 0.23 1.2 3.8 7.5 0.0010 
1.9 0.0055 0.0010 0.68 0.088 0.26 0.21 1.0 2.0 0.0010 
0.89 0.014 0.0010 0.89 0.96 1.7 4.3 0.75 0.93 0.33 
1.9 0.0010 0.0010 0.68 0.088 0.26 0.21 1.0 2.0 0.0010 
0.89 0.018 0.0010 0.89 0.96 1.7 4.3 0.75 0.93 0.78 
0.89 0.030 0.0010 0.89 0.96 1.7 4.3 0.75 0.93 0.27 
0.46 0.047 0.0010 0.13 1.2 0.86 0.57 1.1 0.53 0.0010 
0.38 0.021 1.0 1.6 20 1.9 7.7 0.49 0.087 0.0010 
0.46 0.0010 0.0010 0.13 1.2 0.86 0.57 1.1 0.53 0.0010 
1.9 0.0010 0.0010 0.68 0.088 0.26 0.21 1.0 2.0 0.0010 
1.2 1.0 0.0010 1.9 0.71 2.4 0.92 0.62 2.0 0.0010 
1.2 0.75 0.018 1.9 0.71 2.4 0.92 0.62 2.0 0.040 
0.89 0.072 0.0010 0.89 0.96 1.7 4.3 0.75 0.93 0.28 
5.3 0.031 0.0010 1.9 1.2 1.4 0.48 2.0 1.1 0.067 
5.3 0.031 0.0010 1.9 1.2 1.4 0.48 2.0 1.1 0.075 
Residue (10-mer peptides)Position (ARB)a
12345678910
2.2 0.048 0.013 0.99 5.4 3.7 0.42 0.23 4.4 0.041 
1.2 0.0036 0.0016 1.9 0.71 2.4 0.92 0.62 2.0 0.0010 
0.60 0.0010 0.0010 0.11 0.56 0.19 0.81 2.5 0.18 0.0010 
0.60 0.0010 0.0010 0.11 0.56 0.19 0.81 2.5 0.18 0.0010 
5.3 0.031 0.0010 1.9 1.2 1.4 0.48 2.0 1.1 1.0 
0.40 0.024 0.0010 4.8 0.76 0.23 1.2 3.8 7.5 0.0010 
1.9 0.0055 0.0010 0.68 0.088 0.26 0.21 1.0 2.0 0.0010 
0.89 0.014 0.0010 0.89 0.96 1.7 4.3 0.75 0.93 0.33 
1.9 0.0010 0.0010 0.68 0.088 0.26 0.21 1.0 2.0 0.0010 
0.89 0.018 0.0010 0.89 0.96 1.7 4.3 0.75 0.93 0.78 
0.89 0.030 0.0010 0.89 0.96 1.7 4.3 0.75 0.93 0.27 
0.46 0.047 0.0010 0.13 1.2 0.86 0.57 1.1 0.53 0.0010 
0.38 0.021 1.0 1.6 20 1.9 7.7 0.49 0.087 0.0010 
0.46 0.0010 0.0010 0.13 1.2 0.86 0.57 1.1 0.53 0.0010 
1.9 0.0010 0.0010 0.68 0.088 0.26 0.21 1.0 2.0 0.0010 
1.2 1.0 0.0010 1.9 0.71 2.4 0.92 0.62 2.0 0.0010 
1.2 0.75 0.018 1.9 0.71 2.4 0.92 0.62 2.0 0.040 
0.89 0.072 0.0010 0.89 0.96 1.7 4.3 0.75 0.93 0.28 
5.3 0.031 0.0010 1.9 1.2 1.4 0.48 2.0 1.1 0.067 
5.3 0.031 0.0010 1.9 1.2 1.4 0.48 2.0 1.1 0.075 
a

A panel of 118 10-mer peptides derived from naturally occurring sequences of various viral, bacterial, or pathogen origin was analyzed. All peptides had, minimally, two preferred and one tolerated residue at the main anchor positions, as described in Table II. ARB values shown were calculated as described in Materials and Methods, and are based on the grouping of chemically similar residues (see e.g., 7 ). Values corresponding to a 5-fold or greater increase in binding capacity are indicated in bold. Values corresponding to a 5-fold or greater reduction in binding capacity are indicated by italics. Main anchor positions are positions 2, 3, and 10, and residues determined to be preferred or tolerated anchors are indicated in bold. ARB values at the anchor positions were derived from the analyses described in Fig. 1–3. To filter out nonmotif peptides, the values for nontolerated anchor residues have been set to 0.001, equivalent to a 1000-fold reduction in binding capacity. The average geometric binding capacity of the panel of peptides analyzed was 1858 nM.

Table VIII.

Average relative binding values of different residues at nonanchor positions of 11-mer peptides

Residue (11-mer peptides)Position (ARB)a
1234567891011
0.24 0.048 0.013 0.94 13 5.4 0.81 2.0 8.9 0.36 0.041 
2.5 0.0036 0.0016 1.1 1.9 4.7 0.34 1.1 1.7 1.3 0.0010 
0.093 0.0010 0.0010 1.3 0.28 0.12 2.1 1.4 0.12 0.64 0.0010 
0.093 0.0010 0.0010 1.3 0.28 0.12 2.1 1.4 0.12 0.64 0.0010 
49 0.031 0.0010 1.8 2.4 3.0 0.62 2.3 3.5 0.44 1.0 
0.14 0.024 0.0010 25 0.11 0.21 91 0.25 9.4 0.60 0.0010 
1.0 0.0055 0.0010 0.50 0.24 0.60 0.27 0.70 0.077 12 0.0010 
1.9 0.014 0.0010 0.58 0.58 3.4 1.2 0.41 1.2 0.97 0.33 
1.0 0.0010 0.0010 0.50 0.24 0.60 0.27 0.70 0.077 12 0.0010 
1.9 0.018 0.0010 0.58 0.58 3.4 1.2 0.41 1.2 0.97 0.78 
1.9 0.030 0.0010 0.58 0.58 3.4 1.2 0.41 1.2 0.97 0.27 
2.5 0.047 0.0010 0.52 1.4 0.89 2.2 28 0.87 0.63 0.0010 
0.16 0.021 1.0 0.22 1.5 0.18 2.3 0.12 NA 0.30 0.0010 
2.5 0.0010 0.0010 0.52 1.4 0.89 2.2 28 0.87 0.63 0.0010 
1.0 0.0010 0.0010 0.50 0.24 0.60 0.27 0.70 0.077 12 0.0010 
2.5 1.0 0.0010 1.1 1.9 4.7 0.34 1.1 1.7 1.3 0.0010 
2.5 0.75 0.018 1.1 1.9 4.7 0.34 1.1 1.7 1.3 0.040 
1.9 0.072 0.0010 0.58 0.58 3.4 1.2 0.41 1.2 0.97 0.28 
49 0.031 0.0010 1.8 2.4 3.0 0.62 2.3 3.5 0.44 0.067 
49 0.031 0.0010 1.8 2.4 3.0 0.62 2.3 3.5 0.44 0.075 
Residue (11-mer peptides)Position (ARB)a
1234567891011
0.24 0.048 0.013 0.94 13 5.4 0.81 2.0 8.9 0.36 0.041 
2.5 0.0036 0.0016 1.1 1.9 4.7 0.34 1.1 1.7 1.3 0.0010 
0.093 0.0010 0.0010 1.3 0.28 0.12 2.1 1.4 0.12 0.64 0.0010 
0.093 0.0010 0.0010 1.3 0.28 0.12 2.1 1.4 0.12 0.64 0.0010 
49 0.031 0.0010 1.8 2.4 3.0 0.62 2.3 3.5 0.44 1.0 
0.14 0.024 0.0010 25 0.11 0.21 91 0.25 9.4 0.60 0.0010 
1.0 0.0055 0.0010 0.50 0.24 0.60 0.27 0.70 0.077 12 0.0010 
1.9 0.014 0.0010 0.58 0.58 3.4 1.2 0.41 1.2 0.97 0.33 
1.0 0.0010 0.0010 0.50 0.24 0.60 0.27 0.70 0.077 12 0.0010 
1.9 0.018 0.0010 0.58 0.58 3.4 1.2 0.41 1.2 0.97 0.78 
1.9 0.030 0.0010 0.58 0.58 3.4 1.2 0.41 1.2 0.97 0.27 
2.5 0.047 0.0010 0.52 1.4 0.89 2.2 28 0.87 0.63 0.0010 
0.16 0.021 1.0 0.22 1.5 0.18 2.3 0.12 NA 0.30 0.0010 
2.5 0.0010 0.0010 0.52 1.4 0.89 2.2 28 0.87 0.63 0.0010 
1.0 0.0010 0.0010 0.50 0.24 0.60 0.27 0.70 0.077 12 0.0010 
2.5 1.0 0.0010 1.1 1.9 4.7 0.34 1.1 1.7 1.3 0.0010 
2.5 0.75 0.018 1.1 1.9 4.7 0.34 1.1 1.7 1.3 0.040 
1.9 0.072 0.0010 0.58 0.58 3.4 1.2 0.41 1.2 0.97 0.28 
49 0.031 0.0010 1.8 2.4 3.0 0.62 2.3 3.5 0.44 0.067 
49 0.031 0.0010 1.8 2.4 3.0 0.62 2.3 3.5 0.44 0.075 
a

A panel of 77 11-mer peptides derived from naturally occurring sequences of viral, bacterial, or pathogen origin was analyzed. All peptides had, minimally, two preferred and one tolerated residue at the main anchor positions, as described in Table II. ARB values shown were calculated as described in Materials and Methods, and are based on the grouping of chemically similar residues (see e.g., 7 ). Values corresponding to a 5-fold or greater increase in binding capacity are indicated in bold. Values corresponding to a 5-fold or greater reduction in binding capacity are indicated by italics. Main anchor positions are positions 2, 3, and 11, and residues determined to be preferred or tolerated anchors are indicated in bold. ARB values at the anchor positions were derived from the analyses described in Figs. 1–3. To filter out nonmotif peptides, the values for nontolerated anchor residues have been set to 0.001, equivalent to a 1000-fold reduction in binding capacity. The average geometric binding capacity of the panel of peptides analyzed was 3623 nM.

We thank Robin Delp for expert secretarial assistance. Rose Correa and Carlos Betancourt are acknowledged for their skilled and enthusiastic technical assistance.

1

This work was supported by Small Business Innovation Research, AI38081-03, Small Business Innovation Research AI38620, National Institutes of Health Contract NO1-AI-95362, and Grants RR00168 and AI43890. D.I.W. is an Elizabeth Glaser Scientist.

3

T. M. Allen, B. R. Mothe, J. Sidney, P. Jing, J. L. Dzuris, M. E. Liebl, T. U. Vogel, D. H. O’Connor, X. Wang, M. C. Wussow, et al. CD8+ lymphocytes from SIV-infected rhesus macaques recognize 27 different epitopes bound by the MHC class I molecule Mamu A*01: implications for vaccine design and testing. Submitted for publication.

4

Abbreviation used in this paper: ARB, average relative binding.

1
Letvin, N. L., J. E. Schmitz, H. L. Jordan, A. Seth, V. M. Hirsch, K. A. Reimann, M. J. Kuroda.
1999
. Cytotoxic T lymphocytes specific for the simian immunodeficiency virus.
Immunol. Rev.
170
:
127
2
Goulder, P., D. Price, M. Nowak, S. Rowland-Jones, R. Phillips, A. McMichael.
1997
. Co-evolution of human immunodeficiency virus and cytotoxic T-lymphocyte responses.
Immunol. Rev.
159
:
17
3
Brander, C., B. D. Walker.
1999
. T lymphocyte responses in HIV-1 infection: implications for vaccine development.
Curr. Opin. Immunol.
11
:
451
4
Allen, T. M., J. Sidney, M.-F. del Guercio, R. L. Glickman, G. L. Lensmeyer, D. A. Wiebe, R. DeMars, C. D. Pauza, R. P. Johnson, A. Sette, D. I. Watkins.
1998
. Characterization of the peptide binding motif of a rhesus MHC class I molecule (Mamu-A*01) that binds an immunodominant CTL epitope from simian immunodeficiency virus.
J. Immunol.
160
:
6062
5
Dzuris, J. L., J. Sidney, E. Appella, R. W. Chesnut, D. I. Watkins, A. Sette.
2000
. Conserved MHC class I peptide binding motif between humans and rhesus macaques.
J. Immunol.
164
:
283
6
Parker, K. C., M. A. Bednarke, L. K. Hull, U. Utz, B. Cunningham, H. J. Zweerink, W. E. Biddison, J. E. Cooligan.
1992
. Sequence motifs important for peptide binding to the human MHC class I molecule, HLA-A2.
J. Immunol.
149
:
3580
7
Ruppert, J., J. Sidney, E. Celis, R. T. Kubo, H. M. Grey, A. Sette.
1993
. Prominent role of secondary anchor residues in peptide binding to HLA-A2.1 molecules.
Cell
74
:
929
8
Kondo, A., J. Sidney, S. Southwood, M.-F. del Guercio, E. Appella, J. Sakamoto, E. Celis, H. M. Grey, R. W. Chesnut, R. T. Kubo, A. Sette.
1995
. Prominent roles of secondary anchor residues in peptide binding to HLA-A24 human class I molecules.
J. Immunol.
155
:
4307
9
Kondo, A., J. Sidney, S. Southwood, M.-F. del Guercio, E. Appella, H. Sakamoto, H. M. Grey, E. Celis, R. W. Chesnut, R. T. Kubo, A. Sette.
1997
. Two distinct HLA-A*0101-specific submotifs illustrate alternative peptide binding modes.
Immunogenetics
45
:
249
10
Sidney, J., H. M. Grey, S. Southwood, E. Celis, P. A. Wentworth, M.-F. del Guercio, R. T. Kubo, R. W. Chesnut, A. Sette.
1996
. Definition of an HLA-A3-like supermotif demonstrates the overlapping peptide binding repertoires of common HLA molecules.
Hum. Immunol.
45
:
79
11
Sidney, J., S. Southwood, M.-F. del Guercio, H. M. Grey, R. W. Chesnut, R. T. Kubo, A. Sette.
1996
. Specificity and degeneracy in peptide binding to HLA-B7-like class I molecules.
J. Immunol.
157
:
3480
12
Gulukota, K., J. Sidney, A. Sette, C. DeLisi.
1997
. Two complementary methods for predicting peptides binding major histocompatibility complex molecules.
J. Mol. Biol.
267
:
1258
13
Milik, M., D. Sauer, A. P. Brunmark, L. Yuan, A. Vitiello, M. R. Jackson, P. A. Peterson, J. Skolnick, C. A. Glass.
1998
. Application of an artificial neural network to predict specific class I MHC binding peptide sequences.
Nat. Biotechnol.
16
:
753
14
Pogue, R. R., J. Eron, J. A. Frelinger, M. Matsui.
1995
. Amino-terminal alteration of the HLA-A*0201-restricted human immunodeficiency virus pol peptide increases complex stability and in vitro immunogenicity.
Proc. Natl. Acad. Sci. USA
92
:
8166
15
Bakker, A. B., S. H. van der Burg, R. J. Huijbens, J. W. Drijfhout, C. J. Melief, G. J. Adema, C. G. Figdor.
1997
. Analogues of CTL epitopes with improved MHC class-I binding capacity elicit anti-melanoma CTL recognizing the wild-type epitope.
Int. J. Cancer
70
:
302
16
Parkhurst, M. R., M. L. Salgaller, S. Southwood, P. F. Robbins, A. Sette, S. A. Rosenberg, Y. Kawakami.
1996
. Improved induction of melanoma-reactive CTL with peptides from the melanoma antigen gp100 modified at HLA-A*0201-binding peptides.
J. Immunol.
157
:
2539
17
Rosenberg, S. A., J. C. Yang, D. J. Schwartzentruber, P. Hwu, F. M. Marincola, S. L. Topalian, N. P. Restifo, M. E. Dudley, S. L. Schwarz, P. J. Spiess, et al
1998
. Immunologic and therapeutic evaluation of a synthetic peptide vaccine for the treatment of patients with metastatic melanoma.
Nat. Med.
4
:
321
18
Sarobe, P., C. D. Pendleton, T. Akatsuka, D. Lau, V. H. Engelhard, S. M. Feinstone, J. A. Berzofsky.
1998
. Enhanced in vitro potency and in vivo immunogenicity of a CTL epitope from hepatitis C virus core protein following amino acid replacement at secondary HLA-A2.1 binding positions.
J. Clin. Invest.
102
:
1239
19
Ahlers, J. D., T. Takeshita, C. D. Pendleton, J. A. Berzofsky.
1997
. Enhanced immunogenicity of HIV-1 vaccine construct by modification of the native peptide sequence.
Proc. Natl. Acad. Sci. USA
94
:
10856
20
Cho, B. K., C. Wang, S. Sugawa, H. N. Eisen, J. Chen.
1999
. Functional differences between memory and naive CD8 T cells.
Proc. Natl. Acad. Sci. USA
96
:
2976
21
Sykulev, Y., M. Joo, I. Vturina, T. J. Tsomides, H. N. Eisen.
1996
. Evidence that a single peptide-MHC complex on a target cell can elicit acytolytic T cell response.
Immunity
4
:
565
22
B. N. Fields, and D. M. Knipe, and P. M. Howley, eds.
Fundamental Virology
3rd ed.
1996
Lippincott, Philadelphia, PA.
23
Cullen, B. R., E. D. Garrett.
1992
. A comparison of regulatory features in primate lentiviruses.
AIDS Res. Hum. Retroviruses
8
:
387
24
Collins, K. L., D. Baltimore.
1999
. HIV’s evasion of the cellular immune response.
Immunol. Rev.
168
:
65
25
Almond, N. M., J. L. Heeney.
1998
. AIDS vaccine development in primate models.
AIDS
12
: (Suppl. A):
S133
26
Emerman, M., M. H. Malim.
1998
. HIV-1 regulatory/accessory genes: keys to unraveling viral and host cell biology.
Science
280
:
1880
27
Roebuck, K. A., M. Saifuddin.
1999
. Regulation of HIV-1 transcription.
Gene Expression
8
:
67
28
Peter, F..
1998
. HIV nef: the mother of all evil.
Immunity
9
:
433
29
Karn, J..
1999
. Tackling tat.
J. Mol. Biol.
293
:
235
30
Frankel, A. D., J. A. T. Young.
1998
. HIV-1: fifteen proteins and an RNA.
Annu. Rev. Biochem.
67
:
1
31
Peterlin, B. M., M. Adams, A. Alonso, A. Baur, S. Ghosh, X. Lu, and L. Luo. 1993. Tat trans-activator. In Human Retroviruses. Cullen B. R., ed. Oxford:IRL Press, Oxford, U.K., p. 75.
32
Pollard, V. W., M. H. Malim.
1998
. The HIV-1 rev protein.
Annu. Rev. Microbiol.
52
:
491
33
Greenberg, M. E., S. Bronson, M. Lock, M. Neumann, G. N. Pavlakis, J. Skowronski.
1997
. Co-localization of HIV-1 Nef with the AP-2 adaptor protein complex correlates with Nef-induced CD4 down-regulation.
EMBO J.
16
:
6964
34
Piguet, V., Y. L. Chen, A. Mangasarian, M. Foti, J. L. Carpentier, D. Trono.
1998
. Mechanism of Nef-induced CD4 endocytosis: Nef connects CD4 with the μ-chain of adaptor complexes.
EMBO J.
17
:
2472
35
Salvi, R., A. R. Garbuglia, A. Di Caro, S. Pulciani, F. Montella, A. Benedetto.
1998
. Grossly defective nef gene sequences in a human immunodeficiency virus type 1-seropositive long-term nonprogressor.
J. Virol.
72
:
3646
36
Schubert, U., L. C. Anton, I. Bacik, J. H. Cox, S. Bour, J. R. Bennink, M. Orlowski, K. Strebel, J. W. Yewdell.
1998
. CD4 glycoprotein degradation induced by human immunodeficiency virus type 1 Vpu protein requires the function of proteasomes and the ubiquitin-conjugating pathway.
J. Virol.
72
:
2280
37
von Schwedler, U., J. Song, C. Aiken, D. Trono.
1993
. Vif is crucial for human immunodeficiency virus type 1 proviral DNA synthesis in infected cells.
J. Virol.
67
:
4945
38
Gabuzda, D. H., K. Lawrence, E. Langhoff, E. Terwilliger, T. Dorfman, W. A. Haseltine, J. Sodroski.
1992
. Role of vif in replication of human immunodeficiency virus type 1 in CD4+ T lymphocytes.
J. Virol.
66
:
6489
39
Borman, A. M., C. Quillent, P. Charneau, C. Dauguet, F. Clavel.
1995
. Human immunodeficiency virus type 1 Vif-mutant particles from restrictive cells: role of Vif in correct particle assembly and infectivity.
J. Virol.
69
:
2058
40
Myers, G., B. Korber, S. Wain-Hobson, K. T. Jeang, L. E. Henderson, G. Pavlakis.
1994
.
Human retroviruses and AIDS: a compilation and analysis of nucleic acid and amino acid sequences
Los Alamos National Laboratory, Los Alamos, NM.
41
Vodicka, M. A., D. M. Koepp, P. A. Silver, M. Emerman.
1998
. HIV-1 Vpr interacts with the nuclear transport pathway to promote macrophage infection.
Genes Dev.
12
:
175
42
Emerman, M..
1996
. HIV-1, Vpr and the cell cycle.
Curr. Biol.
6
:
1096
43
Allen, T. M., D. H. O’Conner, P. Jing, J. L. Dzuns, B. R. Mothe, T. U. Vogel, E. Dunphy, M. E. Liebl, C. Emerson, N. Wilson, et al
2000
. Tat-specific cytotoxic T lymphocytes select for SIV escape variants during resolution of primary viremia.
Nature
407
:
386
44
Ensoli, B., A. Cafaro.
2000
. Control of viral replication and disease onset in cynomolgus monkeys by HIV-1 TAT vaccine.
J. Biol. Regul. Homeostatic Agents
14
:
22
45
Sidney, J., S. Southwood, C. Oseroff, M.-F. del Guercio, A. , A. Sette, H. M. Grey.
1998
. The measurement of MHC/peptide interactions by gel infiltration.
Curr. Protocols Immunol.
18
:
3.1
46
Southwood, S., J. Sidney, A. Kondo, M.-F. del Guercio, E. Appella, S. Hoffman, R. T. Kubo, R. W. Chesnut, H. M. Grey, A. Sette.
1998
. Several common HLA-DR types share largely overlapping peptide binding repertoires.
J. Immunol.
160
:
3363
47
Madden, D..
1995
. The three-dimensional structure of peptide-MHC complexes.
Annu. Rev. Immunol.
13
:
587
48
Rammensee, H. G., T. Friede, S. Stevanovic.
1995
. MHC ligands and peptide motifs: first listing.
Immunogenetics
41
:
178
49
Rammensee, H. G., J. Bachmann, N. N. Emmerich, O. A. Bachor, and S. Stevanovic. 1999. SSYFPEITHI: database for MHC ligands and peptide motifs. Immunogenetics50:213. (access via http://www.uni-tuebingen.de/uni/kxi/)
50
Evans, D. T., D. H. O’Connor, P. Jing, J. L. Dzuris, J. Sidney, J. da Silva, T. M. Allen, H. Horton, J. E. Venham, R. A. Rudersdorf, et al
1999
. Virus-specific cytotoxic T-lymphocyte responses select for amino-acid variation in simian immunodeficiency virus Env and Nef.
Nat. Med.
5
:
1270
51
Kast, W. M., R. M. P. Brandt, J. Sidney, J.-W. Drijfhout, R. T. Kubo, H. M. Grey, C. J. M. Melief, A. Sette.
1994
. The role of HLA-A motifs in identification of potential CTL epitopes in human papillomavirus type 16 E6 and E7 proteins.
J. Immunol.
152
:
3904
52
Ishioka, G. Y., J. Fikes, G. Hermanson, B. Livingston, C. Crimi, M. Qin, M.-F. del Guercio, C. Oseroff, C. Dahlberg, J. Alexander, et al
1999
. Utilization of MHC class I transgenic mice for development of minigene DNA vaccines encoding multiple HLA restricted CTL epitopes.
J. Immunol.
162
:
3915
53
Kaur, A., J. Yang, D. Hempel, L. Gritz, G. P. Mazzara, H. McClure, R. P. Johnson.
2000
. Identification of multiple simian immunodeficiency virus (SIV)-specific CTL epitopes in sooty mangabeys with natural and experimentally acquired SIV infection.
J. Immunol.
164
:
934
54
Zhang, C., J. L. Cornette, J. A. Berzofsky, C. DeLisi.
1997
. The organization of human leucocyte antigen class I epitopes in HIV genome products: implications for HIV evolution and vaccine design.
Vaccine
15
:
1291
55
de Campos-Lima, P. O., R. Gavioli, Q. J. Zhang, L. E. Wallace, R. Dolcetti, M. Rowe, A. B. Rickinson, M. G. Masucci.
1993
. HLA-A11 epitope loss isolates of Epstein-Barr virus from a highly A11+ population.
Science
260
:
98
56
Shankar, P., J. A. Fabry, D. M. Fong, J. Lieberman.
1996
. Three regions of HIV-1 gp160 contain clusters of immunodominant CTL epitopes.
Immunol. Lett.
52
:
23