Regulation of CD1d expression and function by a herpesvirus infection

Abstract

Little is known about the role of CD1d-restricted T cells in antiviral immune responses. Here we show that the lytic replication cycle of the Kaposi sarcoma-associated herpesvirus (KSHV) promotes downregulation of cell-surface CD1d. This is caused by expression of the 2 modulator of immune recognition (MIR) proteins of the virus, each of which promotes the loss of surface CD1d expression following transfection into uninfected cells. Inhibition of CD1d surface expression is due to ubiquitination of the CD1d alpha-chain on a unique lysine residue in its cytoplasmic tail, which triggers endocytosis. Unlike MIR-mediated MHC class I downregulation, however, CD1d downregulation does not appear to include accelerated lysosomal degradation. MIR2-induced downregulation of CD1d results in reduced activation of CD1d-restricted T cells in vitro. KSHV modulation of CD1d expression represents a strategy for viral evasion of innate host immune responses and implicates CD1d-restricted T cells as regulators of this viral infection.

Figure 1

KSHV lytic replication decreases CD1d levels. (A) The levels of CD1d on uninduced BCBL-1 cells, which are latently infected with KSHV. The background level of antibody staining of BCBL-1 cells is shown (2° alone). (B) BCBL-1 cells were induced to enter the lytic replication cycle and allowed to proceed to 4 days after induction, after which the cells in the late portion of the lytic cycle could be monitored by K8.1 expression. The cells were stained with a rabbit primary antibody against K8.1 and an anti-rabbit secondary antibody, and with a Zenon-1–APC–labeled antibody against CD1d and a PE-conjugated antibody against transferrin receptor (TfnR). The levels of CD1d for both the K8.1-positive [K8.1(+)] and K8.1-negative [K8.1(–)] populations were determined. To ensure that induction of lytic replication does not cause a general downregulation, the normal level of transferrin receptor was determined by analyzing a bulk population of K8.1-positive cells and creating a gate into which 80% of the cells fell. The K8.1-positive and K8.1-negative populations both showed similar levels of transferrin receptor; thus even in a nongated population, most cells had normal levels of cell-surface receptors.

Figure 2

MIR1 and MIR2 induce CD1d downregulation. (A) BJAB cells were transiently transfected by electroporation with expression vectors for EGFP, MIR1-EGFP, or MIR2-EGFP. Thirty-six hours after transfection, the cells were stained with Zenon-1–APC–labeled mouse mAbs against human CD1d. Levels of HLA-A, HLA-B, and HLA-C (MHC class I); HLA-DR (MHC class II); CD1d; and Fas are shown. For reference, the nonstained control population is also shown. (B) HepG2 cells, which express a high level of endogenous CD1d, were stably transduced with retroviral vectors encoding either lacZ or Flag-tagged MIR2. A mixed population of stable cells was selected, stained with Zenon-1–APC–labeled mAbs, and analyzed by flow cytometry. As a reference, the histogram for the nonstained control is also shown.

Figure 3

MIR2 downregulation of CD1d lowers CD1d-restricted T cell activation. (A) BJAB cells stably transduced with either lacZ (gray bars) as a control or MIR2 (white bars) were cocultured with CD1d-restricted T cells. This coculture was done in the presence of α-GalCer to provide an activating ligand for the T cells. Each coculture was performed in triplicate. After 18 hours of coculture, each supernatant was assayed for IFN-γ release by ELISA. (B) As in A, coculture experiments were done with CD1d-restricted T cells and BJAB cells, this time either in the presence or absence of α-GalCer. Cocultures were done with α-GalCer lacZ (light gray bars), α-GalCer MIR2 (white bars), lacZ without α-GalCer (black bars), and MIR2 without α-GalCer (dark gray bars). After 18 hours, the supernatants were recovered, and the IFN-γ in each supernatant was assayed by ELISA. (C) As in A, coculture experiments were done with CD1d-restricted T cells and BJAB cells, either in the presence of mouse anti–MHC class I as a control (lacZ, dark gray bars; MIR2, white bars) or mouse anti-CD1d (lacZ, black bars; MIR2, light gray bars). After 18 hours, the supernatants were recovered, and the IFN-γ in each supernatant was assayed by ELISA. Results are shown as means ± SD.

Figure 4

MIR2 expression leads to the coimmunoprecipitation of ubiquitinated proteins with CD1d. HepG2 cells stably transduced with either lacZ or MIR2 were lysed, and human CD1d (hCD1d) was immunoprecipitated from the lysates. The immunoprecipitate was then separated by SDS-PAGE and transferred to nitrocellulose, and the presence of ubiquitinated proteins was determined by Western blotting with an HRP-conjugated anti-ubiquitin antibody (Anti–Ub-HRP). As shown, only in the presence of MIR2 did ubiquitinated proteins coimmunoprecipitate with CD1d.

Figure 5

CD1d cytoplasmic lysines are required for MIR2-mediated downregulation and ubiquitination of CD1d. (A) Alignment and organization of the transmembrane and cytoplasmic tails of wild-type human CD1d and the lysine-to-arginine mutant. (B) Several chimeric proteins were created to confirm the requirement of cytoplasmic lysines for CD1d downregulation by MIR proteins. This schematic shows the theoretical domain organization of chimeric CD1d molecules that were constructed and the mutations that were introduced. (C) BJAB cells were cotransfected with either of the 2 chimeric CD1d molecules (wild type or lysine-to-arginine mutant) with an expression vector for EGFP (left panel) or MIR2 fused to EGFP (right panel). The levels of chimeric CD1d were determined by staining with anti-mouse CD1d antibodies, and the cells were analyzed by flow cytometry, gating upon EGFP-positive cells. (D) BJAB cells were stably transduced with either lacZ (as a control) or MIR2 and then stably transduced with either a wild-type chimeric CD1d or a lysine-to-arginine mutation. The cells were then lysed, chimeric mouse CD1d (mCD1d) was immunoprecipitated, and the immunoprecipitate was blotted for the presence of ubiquitin.

Figure 6

MIR2 downregulation of CD1d is due to endocytosis but not enhanced degradation. (A) BJAB cells were transiently transfected with wild-type chimeric CD1d. Both panels show the level of chimeric CD1d when cotransfected with wild-type dynamin (solid black lines) as well as the level of CD1d antibody background cross-reactivity (dotted black lines). The gray lines represent chimeric CD1d levels in the presence of MIR2 and either wild-type (left panel) or dominant-negative dynamin (K44E, right panel). (B) The half-life of CD1d in BJAB cells expressing lacZ or MIR2 was analyzed by pulse chase. As shown, the presence of the MIR2 gene does not change the half-life of the CD1d chains. Chloroquine (C) was added to block lysosomal degradation and showed no effect. (C) The half-life of CD1d in HepG2 cells expressing lacZ or MIR2 were analyzed by pulse chase. As shown, the presence of the MIR2 gene does not change the half-life of the CD1d chains (left blots), but does increase the levels of degradation of MHC class I (right blots). Addition of chloroquine stopped the MHC class I degradation. (D) Total cellular levels of proteins in BJAB cells that were stable for expression of MIR2 or lacZ was determined by flow cytometry. The presence of the MIR2 gene did not change the total levels of CD1d, even though the surface levels of CD1d were decreased. This is in contrast to the decreased total and surface levels of B7.2 and MHC class I.