Portable flanking sequences modulate CTL epitope processing

Abstract

Peptide presentation is critical for immune recognition of pathogen-infected cells by CD8+ T lymphocytes. Although a limited number of immunodominant peptide epitopes are consistently observed in diseases such as HIV-1 infection, the relationship between immunodominance and antigen processing in humans is largely unknown. Here, we have demonstrated that endogenous processing and presentation of a human immunodominant HIV-1 epitope is more efficient than that of a subdominant epitope. Furthermore, we have shown that the regions flanking the immunodominant epitope constitute a portable motif that increases the production and antigenicity of otherwise subdominant epitopes. We used a novel in vitro degradation assay involving cytosolic extracts as well as endogenous intracellular processing assays to examine 2 well-characterized HIV-1 Gag overlapping epitopes presented by the same HLA class I allele, one of which is consistently immunodominant and the other subdominant in infected persons. The kinetics and products of degradation of HIV-1 Gag favored the production of peptides encompassing the immunodominant epitope and destruction of the subdominant one. Notably, cytosolic digestion experiments revealed flanking residues proximal to the immunodominant epitope that increased the production and antigenicity of otherwise subdominant epitopes. Furthermore, specific point mutations in these portable flanking sequences modulated the production and antigenicity of epitopes. Such portable epitope processing determinants provide what we believe is a novel approach to optimizing CTL responses elicited by vaccine vectors.

Figure 1. The endogenous processing of an immunodominant Gag epitope is more efficient than that of subdominant epitopes.

(A) HLA-A3 HeLa cells were pulsed with decreasing amounts of p17 Gag immunodominant RK9 (squares) or p17 Gag subdominant KK9 (triangles) and used as targets in a ⁵¹Cr release assay with RK9- or KK9-specific CTLs respectively. (B) HLA-A3 HeLa cells transfected with a CMV-driven HIV-1 LAV p17 expression vector (schemed above the graph) were used as targets in a ⁵¹Cr assay with RK9- (black bars) or KK9-specific (gray bars) CTLs. The recognition of endogenously processed RK9 and KK9 epitopes by specific CTLs (left side) was compared with that of transfected cells pulsed with 0.1 μM RK9 or KK9 peptides (right side). (C) p17 immunodominant RK9 (squares) and RT subdominant ATK9 (circles) peptide titration on HeLa-A3 cells. (D) HLA-A3 HeLa cells transfected with a CMV-driven vector expressing a HIV-1 LAV RT fragment (HXB2 residues 153–560) with a C terminal p17 sequence (5-RK9-3; according to the position of RK9) (schemed above the graph) were used as targets in a ⁵¹Cr assay with RK9- (black bars) or ATK9-specific (gray bars) CTLs. The recognition of endogenously processed RK9 and ATK9 epitopes by specific CTLs (left side) was compared with that of transfected cells pulsed with 0.1 μM RK9 or ATK9 peptides (right side). Lysis of cells without peptides or with mismatched peptides was below 3%. All data represent averages of 3 experiments.

Figure 2. The cytosolic processing of Gag fragments facilitates the generation of an immunodominant epitope and the progressive loss of the subdominant one.

8 nmol of a Gag 17-mer (A) or 32-mer (C) were incubated with 40 μg PBMC cytosol for increasing periods of time. Peptides encompassing both epitopes (black bars), RK9 only (blue bars), KK9 only (green bars), or no epitope (white bars) were identified by mass spectrometry. Optimal RK9 and KK9 are indicated with blue and green stars. Another HLA-A3 epitope, RY10 (RLRPGGKKKY), codominant with RK9 in acute infection (19), was detected (white star). (B) Digestion products were fractioned on a C18 column and identified and quantified by RP-HPLC. Each peptide is identified by a unique peak at a defined elution time and quantified by the surface under peak. The column was calibrated with defined amounts of peptides covering the corresponding HIV sequence. The identity of the peptides in the digestion mix was confirmed by mass spectrometry. Peptides containing RK9 and KK9 epitopes (RK9+/KK9+, circles), RK9 only (RK9+/KK9–, squares), or KK9 only (RK9–/KK9+, triangles) were quantified over a 3-hour digestion with PBMC cytosol. Data are the average of 3 experiments performed with extracts from 3 healthy donors. The kinetics of peptide production using both the 17-mer and the 32-mer was affected by the amount of cytosol used for the experiment, with high amounts of cytosol yielding a more rapid appearance of RK9 and KK9 (not shown).

Figure 3. The degradation products of a Gag sequence led to a strong RK9 CTL response and a weak KK9 response.

(A) ⁵¹Cr-labeled HLA-A3 B cells were pulsed with 0.25 μg/ml synthetic peptides including KK9 or RK9 with various N or C extensions and used as targets in a ⁵¹Cr release assay using RK9- (black bars) or KK9-specific CTLs (gray bars). (B) Gag 17-mer digestion products at 0.5 μg/ml were used to pulse ⁵¹Cr-labeled HLA-A3 B cells used as targets in a ⁵¹Cr release assay using RK9- (squares) or KK9-specific CTLs (triangles). Lysis percentages of peptide-pulsed or digestion product–pulsed targets were calculated after a 4-hour ⁵¹Cr release assay. Data are the average of 3 experiments performed with extracts from 3 healthy donors. Similar results were obtained with RK9 and KK9 CTL clones isolated from 2 HIV-1–infected donors (not shown).

Figure 4. Identification of portable immunodominance sequences.

(A) 8 nmol of 17-mer WT-ATK9-WT or hybrid peptides DOM-ATK9-DOM and sub-ATK9-sub (where DOM and sub designate the origin [RK9 or KK9] of the flanking sequences) were incubated with 40 μg PBMC extracts for increasing periods of time (peptide sequences in top panel). The amount of ATK9 produced at time 60 (gray bars) and 180 (back bars) was analyzed by mass spectrometry and RP-HPLC (lower left panel). 100% corresponds to the amount of ATK9 produced in 1 hour during the degradation of WT-ATK9-WT. Digestion products from WT-ATK9-WT (WT, Xs), DOM-ATK9-DOM (DD, squares), or sub-ATK9-sub (ss, triangles) were used at 0.05 μg/ml to pulse ⁵¹Cr-labeled HLA-A3 B cells used as targets in a ⁵¹Cr release assay using ATK9-specific CTLs (lower right panel). (B) Similar to A except that the 4 hybrid peptides used for the digestion experiment were DOM-ATK9-DOM, sub-ATK9-sub, DOM-ATK9-sub, and sub-ATK9-DOM. 100% corresponds to the amount of ATK9 produced in 1 hour during the degradation of DOM-ATK9-DOM. Digestion products from 5R-ATK9-3R (DD, squares), sub-ATK9-sub (ss, filled triangles), DOM-ATK9-sub (Ds, diamonds) and sub-ATK9-DOM (sD, open triangles) were used at 0.05 μg/ml to pulse ⁵¹Cr-labeled HLA-A3⁺ B cells used as targets in a ⁵¹Cr release assay using ATK9-specific CTLs (lower right panel). Data are the average of 3 experiments performed with extracts from 3 healthy donors.

Figure 5. Point mutations in the N terminal portable sequences alter epitope production.

(A) The in vitro degradation of 8 nmol of peptides DOM-ATK9-DOM, DOMe-ATK9-DOM (where flanking I is mutated to E), and sub-ATK9-sub was performed as described in the Figure 4 legend. The amount of ATK9 produced at time 60 was analyzed by mass spectrometry and RP-HPLC. 100% corresponds to the amount of ATK9 produced in 1 hour during DOM-ATK9-DOM degradation (lower left panel). The digestion products from DOM-ATK9-DOM (DD, filled squares), DOMe-ATK9-DOM (DeD, open squares), and sub-ATK9-sub (ss, triangles) were used at 0.05 μg/ml to pulse ⁵¹Cr-labeled HLA-A3⁺ B cells used as targets in a ⁵¹Cr release assay using ATK9-specific CTLs. (B) Similar to A except that the 3 peptides were sub-ATK9-sub, subI-ATK9-sub (where flanking E is replaced by I), and DOM-ATK9-sub. 100% corresponds to the amount of ATK9 produced in 1 hour during sub-ATK9-sub degradation. The digestion products from DOM-ATK9-sub (Ds, diamonds), subI-ATK9-sub (sIs, open triangles), and sub-ATK9-sub (ss, triangles) were used in a ⁵¹Cr release assay as in A. (C) Similar to A except that the 3 peptides were WT-ATK9-WT, DOM-ATK9-DOM, and DOM+1-ATK9- DOM+1 (Gag flanking sequences shifted by 1 residue). 100% corresponds to the amount of ATK9 produced in 1 hour during WT-ATK9-WT degradation. The digestion products from WT-ATK9-WT (WT, Xs), DOM-ATK9-DOM (DD, squares), and DOM+1-ATK9-DOM+1 (DD+1, circles) were used in a ⁵¹Cr release assay as in A. Data are the average of 3 experiments performed with extracts from 3 healthy donors.

Figure 6. Transient peptide transfection endogenously replicates the effect of portable flanking sequences on epitope processing observed with in vitro degradation assays.

(A) HeLa-A3 cells were transfected with Gag 17-mer by osmotic loading (black bars) or without osmotic shock (gray bars) and used as targets with RK9- or KK9-specific CTLs. Similar results were obtained with HLA-A3 B cells (not shown). (B) Cells were preincubated with serum-free medium (black bar), MG132 (10 μM; dark gray bar), butabindide (150 μM; light gray bar), bestatin (120 μM; white bar) before transfection by osmotic loading and throughout the ⁵¹Cr assay experiment with RK9-specific CTLs. (C) HeLa A3 cells were transfected by osmotic loading with WT-ATK9-WT (WT, dark gray bar), DOM-ATK9-DOM (DD, black bar), sub-ATK9-sub (ss, light gray bar), DOM-ATK9-sub (Ds, blue bar), sub-ATK9-DOM (sD, yellow bar), DOMe-ATK9-DOM (DeD, green bar), or subI-ATK9-sub (sIs, red bar) or without osmotic loading (shown for DD, white bar). Cells were used as targets in a ⁵¹Cr assay with ATK9-specific clone. (D) The lysis percentage of transfected cells was compared with that of HeLa-A3 cells pulsed with defined amounts of ATK9 (as in Figure 1C). 100% was assigned to the amount of pulsed ATK9 required to reach the same lysis as in WT-ATK9-WT–transfected cells (5.8 nM). The equivalent ATK9 in cells transfected with hybrid sequences was calculated similarly. Data are the average of 3 experiments.