The Epstein-Barr Virus-Encoded BILF1 Protein Modulates Immune Recognition of Endogenously Processed Antigen by Targeting Major Histocompatibility Complex Class I Molecules Trafficking on both the Exocytic and Endocytic Pathways ^{▿ † ‡}

Plasmids, retroviral expression vectors, and small interfering RNA (siRNA).

The N-terminal hemagglutinin (HA)-tagged wild-type (wt) BILF1 and a K122A mutant BILF1 have been described previously (41). The C-terminal deletion mutant was generated by PCR amplification of the BILF1 gene with an introduced stop codon mutation. HA-tagged wt BILF1 and all BILF1 mutants were cloned into the Retro-XTM universal packaging system (Clontech Laboratories, Inc.) to prepare retroviruses. This system is based on the PQCXIH vector, which contains a hygromycin resistance gene. The expression plasmid pCDNA3-HABILF1-GFP and control empty vector pCDNA3-IRES-nlsGFP have been described previously (41), as have the pCEP4-SM plasmid, containing the EBV BSLF2/BMLF1 spliced gene, and p509, containing the EBV BZLF1 gene (42).

Mutant recombinant EBV construction and generation of a virus producer cell line.

Wild-type (clone 2089) and BZLF1-negative (ΔBZLF1) recombinant EBV bacterial artificial chromosomes (BACs) have been previously described (12). The EBV BILF1-negative (ΔBILF1) mutant was constructed (see Fig. S1 in the supplemental material) by replacing the complete BILF1 gene (coordinates 151706 to 152641 of EBV strain B95.8; accession number NC_007605) with the kanamycin resistance gene by homologous recombination with a linear PCR fragment as described previously (8, 21). The kanamycin resistance gene from pCP15 was amplified using primers BILF1-Kan1, CAGGCCTGTGTGTCAGTTTGCAGGGCCATCCTCGCACTCAACCAGTCACGACGTTGTAAAACGAC, and BILF1-Kan2, TTTGCTGCAGACACCACCCAGTCTGGCTCTGACCAGCAAGAACAGCTATGACCATGATTACGCC, resulting in a linear fragment containing the kanamycin resistance gene flanked by 40-bp stretches of homology (underlined) to sequences next to the BILF1 gene. This fragment was transformed into electrocompetent Escherichia coli DH10B cells carrying EBV BAC p2089 as described previously (9, 21). Recombinant clones were selected with kanamycin and analyzed for BAC DNA integrity by restriction enzyme cleavage. Plasmid DNA from a successfully recombined clone was prepared (Nucleobond; Machery-Nagel) and transfected into HEK293 cells by lipofection (Metafectene; Biontex). Cells were kept under hygromycin selection (100 μg/ml) for 3 weeks, and outgrowing green fluoresceint protein (GFP)-positive cell clones were tested for virus production after transfection of BZLF1 and gp110 (BALF4) expression plasmids.

Recombinant EBV strains and generation of LCLs.

Stable 293 cell clones carrying the EBV BACs were selected by hygromycin (100 μg/ml) and induced to produce virus by transfection with BZLF1 and gp110 plasmids by using Lipofectamine 20000 (Invitrogen). Virus supernatants were harvested 3 days posttransfection, filtered through a 0.8-μm-pore-size filter, and stored at 4°C until used to infect transformed laboratory donor peripheral blood B lymphocytes to generate LCLs. LCLs were generated using wt EBV (clone 2089) and ΔBZLF1 and ΔBILF1 recombinant EBVs.

Cells and transfections.

LCLs were maintained in RPMI 1640 supplemented with10% fetal calf serum (FCS). EBV-specific CD8⁺ cytotoxic T cells were grown in 10% FCS in RPMI 1640 medium supplemented with 30% supernatant from the interleukin-2 (IL-2)-producing MLA 144 cell line (26) and 50 U/ml recombinant IL-2, as described elsewhere (25).

The HEK293 epithelial cell line (American Type Culture Collection) and derived cell lines were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FCS and penicillin-streptomycin antibiotics. HEK293 lines transduced by PQCXIH-based retroviruses were selected with 400 μg/ml hygromycin. Transient transfection of HEK293 with plasmid DNA was routinely performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions.

Retroviral infection.

For the production of replication-defective recombinant retrovirus, the packaging cell line (GP2-293) was transfected with retroviral vectors and PVSV-G plasmid. Supernatants containing the retrovirus particles were harvested 72 h after transfection and filtered through a 0.45-μm-pore-size, low-protein-binding filter. To generate stable cell lines by transduction with retrovirus, HEK293 cells were infected with 1 ml retrovirus supernatants with Polybrene added to a final concentration of 4 μg/ml. The stable cell lines were selected with 400 μg/ml hygromycin.

Antibodies.

The murine monoclonal antibodies (MAbs) used to detect human MHC class I were W6/32 (3), which recognizes native β₂ microglobulin-associated MHC-I complexes (HLA-A, -B, and -C alleles), and HC10 (35), which recognizes free HLA class I heavy chains. For flow cytometry experiments, phycoerythrin (PE)-conjugated MAb to HLA-A, -B, and -C (MCA81PE; clone W6/32) was purchased from AbD Serotec. Goat antibodies to calregulin (sc6467) were purchased from Santa Cruz Biotechnology. The 3F10 rat MAb directed against the influenza virus-derived HA tag was purchased from Roche Diagnostics. The rabbit serum with anti-BMLF1 (EB2) antibodies (6) was a kind gift of A. Sergeant, Lyon, France. The BZ.1 murine MAb specific for the EBV BZLF1-encoded protein was generated by our investigators.

Flow cytometry analysis of cell surface MHC molecules.

Cell surface expression of MHC-I on viable cells was determined by staining with PE-labeled W6/32 antibodies or PE-labeled isotype control MAb and detection on a Beckman Coulter XL flow cytometer. The data were analyzed using FlowJo software (Tree Star).

To assay the kinetics of internalization of surface MHC-I, HEK293 cells were incubated for 60 min on ice with saturating amounts of W6/32 MAb, then washed three times in phosphate-buffered saline (PBS) and replaced in warm culture medium to incubate at 37°C for the length of time indicated below in Results. To terminate MHC/antibody complex internalization, cells were rapidly cooled to 0°C. Finally, the MAb-bound surface MHC-I molecules were stained at 0°C with PE-conjugated goat anti-mouse IgG2a antibody (AbD Serotec), and cells were analyzed by flow cytometry.

To assay the kinetics of appearance of surface MHC-I molecules, HEK293 cells were incubated with saturating amounts of W6/32 MAb, then incubated at 37°C for different lengths of time as described for the internalization assay (see above). After cooling to 0°C to prevent further appearance of unblocked surface MHC-I molecules, the cells were stained with PE-conjugated W6/32 MAb and analyzed by flow cytometry.

To assay the kinetics of recycling of surface MHC-I molecules, HEK293 cells were incubated for 60 min on ice with saturating amounts of W6/32 MAb, then washed three times in PBS, replaced in warm culture medium, and incubated at 37°C for 30 min. The cells were then washed with PBS, and remaining surface-bound W6/32 MAb was stripped by gently resuspending the cells in pH 3.1 citrate/phosphate buffer (0.131 M citric acid, 0.066 M Na₂HPO₄) for 1 min on ice before neutralization by addition of excess standard medium. Stripped target cells were then washed twice with PBS, resuspended in warm medium, and incubated at 37°C for the length of time indicated below in Results. To terminate further recycling of MHC-I molecules, cells were rapidly cooled to 0°C. Finally, recycled MHC-I molecules were stained with PE-conjugated goat anti-mouse IgG2a antibody (Serotec), and the cells were analyzed by flow cytometry.

Flow cytometric analysis of intracellular staining for lytic cycle antigens.

The percentages of cells spontaneously reactivating into the lytic cycle in LCLs were measured by intracellular staining for BZLF1. Cells were first fixed using 100 μl of Ebiosciences intracellular (IC) fixative (catalog number 00-8222-49) for 1 h on ice, followed by permeabilization through the addition of 100 μl Triton X-100 (final concentration, 0.2%) and a further 30 min of incubation on ice. After extensive washing with PBS, cells were incubated with 1 μg/ml of either MAb BZ.1 (anti-BZLF1) or with an IgG1 isotype control MAb for 1 h at 37°C. Cells were washed twice in PBS and then incubated with 1:50-diluted R-phycoerythrin-conjugated goat anti-mouse IgG1 antibody (catalog number STAR132PE; AbD Serotec) for 1 h at 37°C. Following further washes, cells were analyzed on a Beckman Coulter XL flow cytometer, and the data were processed using FlowJo software (Tree Star).

For comparison of surface MHC-I expression in lytic and latent 293/EBV cells, the 293/wt EBV cells and 293/ΔBILF1 cells were induced to the EBV lytic cycle by transfecting BZLF1 plasmids. At 24 h postinduction, the surface HLA class I and intracellular EBV gp110 antigens were detected simultaneously. First, viable 293 cells were stained with 1:15-diluted allophycocyanin-conjugated anti-human HLA-A, -B, or -C anitbody (catalog number 311410; Biolegend) for 30 min on ice. Cells were then washed extensively in PBS and fixed and permeabilized as described above, followed by incubation for 1 h at 37°C with 1 μg/ml of L2 MAb (late antigen gp110) or with isotype control MAb. After washing in PBS, cells were incubated for 1 h with 1:20-diluted peridinin chlorophyll protein-conjugated goat anti-mouse IgG1 antibody. Cells were analyzed by flow cytometry as described above.

Western blotting.

For Western blotting, total cell lysates were denatured in reducing sample buffer (final concentrations, 2% SDS, 72.5 mM Tris-HCl [pH 6.8], 10% glycerol, 0.2 M sodium 2-mercaptoethane sulfonate, 0.002% bromophenol blue) and then sonicated and heated to 100°C for 5 min. Solubilized proteins equivalent to 1 × 10⁵ cells/20-μl sample were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) on 4-to-12% acrylamide gradient bis-Tris NuPage minigels with morpholinepropanesulfonic acid running buffer (Invitrogen).

Following electroblotting to polyvinylidene difluoride membranes (Invitrogen) and blocking with I-Block (Tropix; Applied Biosystems) in PBS with 0.1% Tween 20 detergent, specific proteins were detected by incubating the membranes with primary antibodies at 4°C overnight. The purified mouse MAb HC10 and the goat antibody to calregulin were used at 1 μg/ml, the rat anti-HA MAb was used at 50 ng/ml, and the anti-BMLF1 rabbit serum was used at 1/6,000. Primary antibodies specifically bound to blotted proteins were detected by incubation for 30 min with appropriate alkaline phosphatase-conjugated secondary antibodies, developed using a CDP-Star detection kit (Tropix; Applied Biosystems), and exposed to autoradiographic film.

T cell function assays.

The effector T cell clone GLC, specific for BMLF1 and restricted through HLA-A2, was generated as described elsewhere (25). Targets for the GLC clone were generated by cotransfection of HLA-A2-type HEK293 cells with a BMLF1 expression plasmid and wt BILF1 or different BILF1 mutants' expression plasmids. At 24 h posttransfection, recognition of target cells by the effector T cells was determined by enzyme-linked immunosorbent assay (ELISA) of gamma interferon (IFN-γ) release using a standard protocol described elsewhere (19). Briefly, 10⁴ effector T cells were incubated for 18 h at 37°C in V-bottom microtest plate wells with 10⁵ target cells, then the supernatants were harvested for quantitation of IFN-γ by ELISA (Endogen) in accordance with the manufacturer's recommended protocol. Specificity control targets included HLA-matched and -mismatched EBV-transformed LCLs, empty vector-transfected HEK293 cells, and empty vector-transfected HEK293 cells pulsed with the GLCTLVAML (GLC) synthetic peptide.

For the peptide disappearance assay, HEK293 cells were pulsed with synthetic GLC peptide on ice for 1 h, then washed extensively with PBS, replaced in warm culture medium, and incubated at 37°C for different periods of time. At each time point, the cells were washed twice with PBS and fixed with 1% paraformaldehyde (PFA) in PBS at room temperature for 10 min. The PFA was subsequently quenched with 0.2 M glycine in PBS. The fixed, quenched, and washed peptide-pulsed cells were cocultured with the CD8⁺ effector GLC T cell clone for a further 18 h, and the supernatants were tested for the release of IFN-γ as a measure of T cell recognition. All results are expressed in terms of IFN-γ release (in pg/ml), and error bars indicate standard deviations of triplicate cultures.

For the EBV lytic antigen CD8⁺ T cell against LCLs recognition assay, 10⁴ effector T cells were incubated for 18 h at 37°C in V-bottom microtest plate wells with 2 × 10⁵ target LCLs, and then the supernatants were harvested for quantitation of IFN-γ by ELISA (Endogen) in accordance with the manufacturer's recommended protocol.

NF-κB reporter assays.

HEK293 cells were seeded at 2 × 10⁵ cells/well in 24-well plates at 24 h prior to transfection with a constitutively expressed Renilla luciferase reporter construct (phRL-TK; Promega) for normalizing transfection efficiency, and 3 enh-ConA (an NF-κB-dependent luciferase reporter construct in which transcription of the firefly luciferase gene is driven by three NF-κB binding sites [1]), together with vectors for wt BILF or BILF1 mutants or empty vector DNA. Transfection was performed with Lipofectamine 2000 (Invitrogen) and DNA mix in Opti-MEM, prepared and used according to the supplier's instructions. Cell extracts were generated after 48 h using cell culture lysis buffer (Promega), and extracts were assayed for firefly luciferase and Renilla luciferase activities by using the dual-luciferase reporter assay system (Promega).

BILF1 contributes to the downregulation of surface MHC-I in the EBV lytic cycle.

BILF1 as an immune evasion gene causing downregulation of surface MHC-I expression was originally demonstrated using a reductionist approach by expressing the BILF1 gene in isolation. We therefore sought first to confirm the contribution of BILF1 in the context of the whole viral genome. To this end, a recombinant EBV lacking BILF1 was generated (EBV ΔBILF1), and stable 293 cell clones carrying the EBV wt BAC and EBV ΔBILF1 BAC were generated by transfection and hygromycin selection. The EBV lytic cycle was then synchronously induced in these cells by transient transfection of a BZLF1-expressing plasmid. These induced cultures were analyzed by two-color immunoflourescence staining for EBV gp110 (BALF4) late lytic cycle antigen and cell surface MHC-I. As shown in Fig. 1, surface MHC-I levels in the 293/EBV wt cells were reduced by about 80% in the lytically infected subpopulation expressing gp110. In contrast, in the 293/ΔBILF1 cells, the surface MHC-I expression in the lytically infected subpopulation was reduced by less than 50%. This result is consistent with the hypothesis that BILF1 cooperates with other EBV immune evasion genes to reduce the expression of MHC-I at the cell surface.

The EKT signaling motif and a C-terminal domain of BILF1 cooperate for degradation of MHC-I.

To identify the functional domain(s) of BILF1 responsible for the diversion of MHC-I for degradation and the impairment of T cell recognition, a series of BILF1 mutants were made. Three mutants proved to be informative for the current study: K122A mutant BILF1, in which the lysine at codon 122 is replaced with an alanine, disrupting a DRY-like EKT motif (31, 38) necessary for NF-κB signaling (41); ΔC mutant BILF1, from which the last 21 C-terminal amino acids are deleted; the K122A/ΔC double mutant BILF1 (see Fig. 2A, below). An N-terminal HA tag sequence was engineered into wt BILF1 and each of the mutants, which were then cloned into the PQCXIH retrovirus vector to generate the recombinant retroviruses used to transduce HEK293 cells. Stable expressing cell lines were selected with hygromycin.

Total cell lysates of these transduced cell lines showed comparable expression levels of wt and mutant BILF1 proteins when analyzed by SDS-PAGE and immunoblotting, while the levels of MHC-I were affected by the expression of BILF1, as shown by the representative immunoblotting experiments with whole-cell lysates in Fig. 2 B. Densitometry was performed on immunoblots from three separate experiments to quantitate expression of total MHC-I; consistent with previous observations (41), wt BILF1 caused a 52% reduction in the expression of total MHC-I relative to control (PQC) transduced HEK293 cells (Fig. 2C). In addition, expression of K122A mutant BILF1 similarly reduced the level of total MHC-I by 46%, and expression of the ΔC mutant BILF1 was associated with a 40% reduction (Fig. 2C). In marked contrast, cells transduced with virus expressing the K122A/ΔC double mutant BILF1 showed levels of total MHC-I that did not differ significantly from the levels seen in control (PQC) transduced cells (Fig. 2C). The results thus far described (Fig. 2) suggest that the C-terminal domain of BILF1 contains a determinant that cooperates with the EKT signaling function of BILF1 to target MHC class I for degradation.

However, when the effects on levels of cell surface MHC-I expression were examined by flow cytometry of W6/32 MAb-stained cells, a different pattern of results was obtained. While cells transduced with the wt BILF1, K122A mutant, or ΔC mutant BILF1 showed a reduction in cell surface expression that reflected the reduction in total MHC-I expression, the KI22A/ΔC double mutant BILF1 showed an approximately 25% reduction in cell surface MHC-I expression, although there was no significant change in total cellular MHC-I expression (cf. Fig. 2C and D). However, we did see a pronounced accumulation of intracellular MHC-I in K122A/ΔC mutant BILF1-expressing cells, as revealed by immunofluorescence staining of fixed and permeabilized cells (Fig. 2E).

The EKT signaling motif, but not the C terminus of BILF1, is necessary for enhanced endocytosis of surface MHC-I.

The disconnection between the levels of total and cell surface MHC-I expression with the K122A/ΔC mutant BILF1 hinted at a greater mechanistic complexity for the effect of BILF1 on MHC-I antigen presentation than was previously envisaged. We therefore included this panel of mutants in a reexamination of key known phenotypes of BILF1.

First we examined the signaling functions of the mutants in a reporter assay by cotransfecting an NF-κB reporter plasmid with or without wt or mutant BILF1 expression vectors (Fig. 3 A). The wt BILF1 constitutively activated NF-κB signaling by approximately 12-fold above background, and disruption of the EKT motif in the K122A mutant BILF1 reduced this activation by almost 70%. In addition, these experiments showed for the first time an even greater impairment of NF-κB activation in the ΔC and K122A/ΔC mutants.

Next, we examined the ability of these mutants to mediate the previously reported ability of wt BILF1 to enhance endocytosis of surface MHC-I. As before, we used an internalization assay in which cell surface MHC-I was first labeled with W6/32 MAb at 0°C, and the amount remaining on the cell surface after 20 min of incubation at 37°C was determined by detection with PE-conjugated anti-mouse IgG and analyzed by flow cytometry. The pooled results from three independent experiments are shown in Fig. 3B. While wt BILF1 caused an enhanced rate of MHC-I internalization (16.4 ± 1.4% versus the control, 10.9 ± 1.5%, in 20 min), the K122A mutant almost completely abrogated this effect. The rate of MHC-I internalization in K122A mutant BILF1-expressing cells (12.4 ± 1.0% in 20 min) was not significantly different from that in the control transduced cells. In contrast, the ΔC mutant BILF1 retained the ability to enhance surface MHC-I internalization (16.2 ± 1.4%). Similar to the K122A mutant, the K122A/ΔC double mutant BILF1 showed a rate of internalization of surface MHC-I (11.9 ± 0.3%) that was indistinguishable from control cells.

The data in Fig. 3 show that the EKT signaling motif of BILF1 is essential for the ability of BILF1 to enhance internalization of surface MHC-I, while the C terminus is dispensable. This cannot be explained simply in terms of the NK-κB signaling properties of BILF1.

The K122A/ΔC double mutant BILF1 selectively reduces the rate of appearance of surface MHC-I.

Since the K122A/ΔC double mutant BILF1 is completely defective with regard to enhancing surface MHC-I internalization (Fig. 3B), we questioned how it achieves the small but reproducible reduction in surface MHC-I levels shown in Fig. 2D. We therefore examined the phenotype of the K122A/ΔC mutant in more detail, by comparing the rate of internalization and the rate of appearance of MHC-I at the cell surface. Figure 4 A shows a representative internalization assay in which the percentage of MHC-I remaining on the cell surface was measured over a 60-min time course. As expected from the data shown in Fig. 3B, cells expressing wt BILF1 showed a higher rate of internalization than control cells, while K122A/ΔC double mutant BILF1 expressing cells showed the same rate of internalization as the control cells (Fig. 4A). In contrast, when we used a modification of the assay that measured the rate of appearance of surface MHC-I, the K122A/ΔC double mutant BILF1 behaved exactly as did wt BILF1 in reducing the rate of appearance of surface MHC-I (Fig. 4B).

We previously attributed the effect of wt BILF1 on the rate of appearance of surface MHC-I to a secondary and delayed effect of the BILF1-mediated enhanced internalization (41). However, these new results with the K122A/ΔC double mutant BILF1 (Fig. 4A and B) are inconsistent with that interpretation, as K122A/ΔC double mutant BILF1 reduced the rate of appearance of surface MHC-I under conditions when this mutant did not enhance internalization. Since the assay shown in Fig. 4B measures the appearance of both recycled and newly synthesized MHC-I molecules, we therefore devised a modification of the assay to measure the amount of MHC-I recycling to the cell surface after internalization.

In this recycling assay, the surface MHC-I was first saturated with W6/32 MAb and then allowed to internalize for 30 min at 37°C before acid stripping away all MAb remaining on the surface. The reappearance of internalized MHC-I/MAb complexes was then determined at serial time points by staining the surface with PE-conjugated anti-mouse IgG antibodies. In such assays, we found that BILF1 had no effect on the rate of recycling of internalized MHC-I molecules (Fig. 4C). Notably, 15% of the internalized MHC-I was recycled to the surface within 5 min; thereafter, the rate of recycling was slower, such that the following 15% recycled portion took more than 25 min. This suggests that the initial MHC-I to arrive at the surface in the appearance assay (Fig. 4B) will be derived predominantly from the recycling pool and, thereafter, newly synthesized MHC-I molecules arriving from the endoplasmic reticulum (ER)/Golgi compartments make a greater contribution to the MHC-I molecules appearing at the cell surface.

Together, the results in Fig. 4 showed that wt BILF1 both enhances the rate of internalization of surface MHC-I molecules and reduces the rate of appearance of newly synthesized MHC-I molecules at the cell surface. The K122A/ΔC double mutant BILF1 only reduces the rate of appearance of newly synthesized MHC-I. Thus, the EKT signaling motif is necessary for enhanced internalization, and the C terminus is necessary for subsequent enhanced degradation of MHC-I (Fig. 2 and 3). Neither the EKT motif nor the C-terminal domain is required for diverting newly synthesized MHC-I on the exocytic pathway (Fig. 4B).

BILF1-mediated enhanced endocytosis of MHC-I from the cell surface is important for antigen presentation to CD8⁺ T cells.

As BILF1 can act on both the exocytic and endocytic trafficking pathways for MHC-I, we next addressed which of these functions contributed to evading recognition by immune CD8⁺ T cells.

To address this question with regard to the endocytic pathway, we first pulsed control 293, wt BILF1 293, and K122A/ΔC BILF1 293 cells with synthetic GLCTLVAML peptide before fixing the cells and using them as targets in T cell assays with the HLA-A2-restricted GLC CD8⁺ T cell clone. As expected, the results reflected the levels of MHC-I molecules on the surface of the respective cell lines (cf. Fig. 5A and 2D). Thus, T cell recognition of peptide-pulsed cells expressing wt BILF1 was reduced by 61% compared to the recognition of peptide-pulsed control cells, and recognition of peptide-pulsed cells expressing K122A/ΔC double mutant BILF1 was 31% lower than the recognition of peptide-pulsed control cells (Fig. 5A).

Next, viable peptide-pulsed cells were incubated at 37°C for up to 3 h before harvesting and fixing for use as targets in T cell assays. In these experiments, the peptide-pulsed control 293 cells rapidly internalized peptide/MHC-I complexes from the cell surface and became less potent targets for stimulating the effector T cells to produce IFN-γ; thus, about 45% of T cell recognition was lost within the first hour (Fig. 5B). The peptide-pulsed wt BILF1 293 cells showed an even higher rate of peptide/MHC-I complex internalization, showing an almost 80% loss of T cell recognition within 1 h of incubation (Fig. 5B). In contrast, the peptide-pulsed K122A/ΔC mutant BILF1 293 cells showed loss of surface peptide/MHC-I complexes with kinetics and magnitude indistinguishable from control 293 cells (Fig. 5B), consistent with the inability of this double mutant to enhance endocytosis (Fig. 4A). These results demonstrate a role for BILF1-mediated enhanced endocytosis of MHC-I complexes in modulating antigen presentation to immune effector T cells.

Diversion of MHC-I on the exocytic pathway by BILF1 affects antigen presentation to CD8⁺ T cells.

We next asked to what extent the effect of BILF1 on the MHC-I exocytic trafficking pathway might influence antigen presentation for functional CD8⁺ T cell recognition. To address this question, HEK293 cells were cotransfected with BMLF1 target antigen plasmid together with plasmid vectors for wt BILF1 or the different BILF1 mutants. These target cells were then examined for antigen presentation to effector GLC CD8⁺ T cells specific for a BMLF1-derived peptide (GLCTLVAML). Following coculture of transfected target cells with effector CD8⁺ T cells, T cell recognition was measured by ELISA for release of IFN-γ from the effector cells.

The representative results in Fig. 6A show good recognition of cells transfected with BMLF1 alone, which was inhibited by more than 90% when wt BILF1 was coexpressed. Expression of the K122A mutant BILF1 or ΔC mutant BILF1 resulted in similarly efficient inhibition of CD8⁺ T cell recognition. Notably, expression of the K122A/ΔC double mutant BILF1 resulted in an up-to-10-times-greater CD8⁺ T cell recognition than when wt BILF1 or K122A mutant BILF1 or ΔC mutant BILF1 were expressed. Nevertheless, the CD8⁺ T cell recognition in the K122A/ΔC double mutant BILF1 transfectant was typically only about 35% of that seen in the control transfectant lacking any BILF1. Immunoblotting of target cells showed that expression of transfected BMLF1 target protein was comparable in the different transfected targets (Fig. 6B). Similar results were obtained in three separate experiments. In addition, experiments using MelJuSo target cells with RAK CD8⁺ effector T cells specific for a BZLF1-derived peptide (41) gave the same pattern of results with the panel of BILF1 mutants (data not shown).

As the K122A/ΔC double mutant BILF1 selectively disrupts trafficking of newly synthesized MHC-I to the cell surface and still inhibits CD8⁺ T cell recognition by 65%, these results imply that diversion of MHC-I on the exocytic pathway by BILF1 substantially affects antigen presentation to CD8⁺ T cells.

Together, the results in Fig. 5 and 6 demonstrate that the mechanism of immune evasion of BILF1 includes both enhanced internalization of existing surface peptide/MHC-I complexes and diversion of newly synthesized peptide/MHC-I complexes. From Fig. 5 and 6, it can be estimated that the effect of BILF1 on the exocytic pathway is responsible for about 70% of the immune evasion property of wt BILF1 in this model.

Deletion of BILF1 confers an increase in recognition by EBV lytic antigen-specific CD8⁺ T cells.

Having used a reductionist approach in epithelial cells to elucidate the broad mechanisms by which BILF1 modulates MHC-I expression and thereby recognition of antigen by immune CD8⁺ effector cells, we returned to the question of whether BILF1 contributes to immune evasion in the context of the whole virus during the lytic cycle in B cells. A panel of LCLs from an HLA-A2-matched donor was derived by transformation of peripheral blood B cells with either recombinant EBV wt, EBV ΔBILF1, or EBV ΔBZLF1. As LCLs normally show variable numbers of cells spontaneously entering the lytic cycle, we generated sets of LCLs in parallel from the same donor bleeds and monitored levels of the lytic cycle by intracellular staining for the immediate-early (IE) gene product BZLF1. Lines which had similar levels of spontaneous lytic cycle were selected for T cell recognition assays. In the representative experiment shown, the EBV wt and EBV ΔBILF1 LCLs both contained around 3% BZLF1⁺ cells, while the control EBV ΔBZLF1 LCLs were defective for lytic cycle entry (Fig. 7A). These lines were used as targets in T cell assays with HLA-A2-restricted, EBV lytic antigen-specific CD8⁺ T cells: the YVL effectors specific for a peptide which comes from the IE lytic antigen BRLF1; GLC effectors specific for the early (E) lytic antigen BMLF1; TLD specific for the delayed early (DE) lytic antigen BMRF1. As shown in Fig. 7B, the ΔBILF1 LCLs were recognized more efficiently than the wild-type LCLs by each of the three effector CD8⁺ T cell clones.

Together, the results in Fig. 1 and 7 show that, in the context of whole EBV, BILF1 not only contributes to the downregulation of surface MHC-I in the EBV lytic cycle but also contributes to the impairment of CD8⁺ T cell recognition of EBV lytic cycle antigens in B cells.