CD83 suppresses endogenous March-I-dependent MHC class II ubiquitination, endocytosis, and degradation

Abstract

MHC class II glycoproteins (MHC-II) bind peptides derived from exogenous antigens and dendritic cells (DCs) present these peptide MHC-II (pMHC-II) complexes to antigen-specific CD4 T cells during immune responses. The turnover of surface pMHC-II on antigen-presenting cells (APCs) is controlled by ubiquitin-mediated degradation of pMHC-II by the E3 ubiquitin ligase March-I. To study March-I protein expression, we have generated a mouse in which a V5 epitope-tag was knocked-in to the endogenous March-I gene, thereby allowing us to follow the fate of March-I using high-affinity anti-V5 antibodies. Quantitative analysis revealed that resting spleen DCs and B cells express only ~500 and 125 March-I molecules/cell, respectively. Endogenous March-I protein has a very short half-life in DCs and March-I mRNA, March-I protein, and MHC-II ubiquitination are rapidly terminated upon activation of both DCs and B cells. Like March-I, CD83 is a known regulator of MHC-II expression in APCs and we also show that CD83 suppresses endogenous March-I-dependent MHC-II ubiquitination, endocytosis, and degradation in mouse spleen DCs. Thus, our study reveals molecular mechanisms for both March-I- and CD83-dependent regulation of MHC-II expression in APCs.

Keywords: MHC class II; antigen presentation; ubiquitination.

Fig. 1.

Spleen DCs contain only 500 March-I molecules/cell. (A) March-I-V5 mice were generated by CRISPR/Cas9 modification of March-I exon 10. V5-epitope tag was inserted immediately prior to the stop codon of March-I (indicated by asterisk). (B) Representation of March-I protein orientation in cell membranes with the location of the carboxy-terminal V5-epitope tag indicated and the GST-March-I C-terminal cytosolic domain-V5 fusion protein. (C and D) GST-March-I-V5 was purified, and various amounts of fusion protein and samples from different March-I-V5 knock-in mice were analyzed by SDS page and immunoblotting with anti-V5 mAb. The amount of March-I-V5 in each anti-V5 immunoprecipitate (panel C) or cell lysate (panel D) was determined by comparison to a standard curve of GST-March-I-V5 band intensity. We estimate that there are 417 ± 83 or 561 ± 86 molecules of March-I-V5 per spleen DC when analyzing data from cell lysates or anti-V5 immunoprecipitates, respectively.

Fig. 2.

Activation of spleen DCs with LPS results in rapid downregulation of March-I-V5 protein expression and MHC-II ubiquitination. Spleen DCs were isolated from control mice (V5⁻) or March-I-V5 KI mice (V5⁺). The cells were incubated ex vivo with LPS for various times prior to harvesting and analysis as indicated. (A) mRNA was isolated from cells and the amount of March-I present at each time point was calculated by RT-PCR. Data shown are the mean ± SD obtained from four independent experiments; ***P < 0.001 (B) The amount of March-I present in V5⁻ and V5⁺ DCs was determined by immunoprecipitation and immunoblotting. The results shown are the average of five independent experiments and the errors bars are ± S.D. (C) Cells were lysed and subjected to immunoprecipitation with MHC-II mAb and immunoblotting. Blots were probed with antibodies recognizing ubiquitin (Upper panel) or total MHC-II β-chain (Lower panel). The results shown are the average of five independent experiments and the errors bars are ± S.D. (D) The relative amounts of March-I protein and MHC-II ubiquitination obtained after treatment of spleen DCs with LPS were expressed relative to amount of each present in unstimulated cells. (E) Bone marrow–derived DCs were incubated for various times in medium alone (control), medium containing LPS (+LPS), or medium containing cycloheximide (+CHX). The amount of March-I present in each sample was determined by immunoblotting of cell lysates. The results shown are the average of three independent control/+CHX experiments and two control/+LPS experiments.

Fig. 3.

Activation of spleen B cells results in rapid downregulation of March-I-V5 protein expression and MHC-II ubiquitination. Spleen B cells were isolated from control mice (V5⁻) or March-I-V5 KI mice (V5⁺). The cells were incubated ex vivo with either LPS (A–C) or by BCR cross-linking (D and E) for various times prior to harvesting and analysis as indicated. (A) mRNA was isolated from cells and the amount of March-I present at each time point in V5⁻ mice or V5⁺ mice was calculated by RT-PCR. The data shown are the average obtained from more than three independent experiments. (B and D) Cells were lysed and subjected to immunoprecipitation and immunoblotting with anti-V5 antibodies. The results shown are the average of three independent experiments, and the errors bars are ± S.D. (C and E) Cells were lysed and subjected to immunoprecipitation with MHC-II mAb and immunoblotting. Blots were probed with antibodies recognizing V5, ubiquitin, or total MHC-II β-chain as indicated. The results shown are the average of three independent experiments.

Fig. 4.

CD83 suppresses March-I-mediated MHC-II ubiquitination and degradation. (A) Spleen cells from the indicated mice were isolated and stained with antibodies recognizing pMHC-II as well as markers for either B cells or DCs. Surface pMHC-II expression was determined by FACS analysis. (B–E) Spleen DCs were isolated from March-I-V5 KI mice (WT) or CD83-deficient March-I-V5 KI mice (CD83 KO). (B) DC cell lysates were subjected to immunoprecipitation with MHC-II mAb and immunoblotting with antibodies recognizing ubiquitin or total MHC-II β-chain. The results shown are the average of three independent experiments, and the errors bars are ± S.D., *P < 0.05. (C) Spleen DCs were incubated ex vivo with LPS for various times prior to immunoprecipitation with MHC-II mAb and immunoblotting. The results shown are the average of three independent experiments, and the errors bars are ± S.D. (D) Cells were lysed and analyzed by immunoprecipitation and immunoblotting using anti-V5 antibodies. The results shown are the average of three independent experiments, and the errors bars are ± S.D. (E) The cells were incubated ex vivo with LPS for various times. Cells were lysed and analyzed by immunoprecipitation and immunoblotting using anti-V5 antibodies. The results shown are the average of three independent experiments and the errors bars are ± S.D., *P < 0.05; ***P < 0.001; ns, not significant. (F) Bone marrow–derived DCs were generated from the indicated mice, the cells were biotinylated on ice, and the survival of labeled pMHC-II from lysates of freshly labeled cells (t = 0) and after a 4 h chase period at 37 °C (4 h) determined. The results shown are the average of three independent experiments, and the errors bars are ± S.D., *P < 0.05; ns, not significant. (G) DCs were isolated from the indicated mice, the cells were incubated with modified pMHC-II mAb on ice, and the internalization rate of surface pMHC-II was measured using the endocytosis assay described in SI Appendix. The results shown are representative of two independent experiments.

Fig. 5.

CD83 suppresses March-I binding to MHC-II in vivo. (A) Spleen DCs were isolated from control mice (V5⁻) or March-I-V5 mice (V5⁺), plated on angiogenesis slides, and analyzed by PLA by confocal microscopy using mouse mAb recognizing pMHC-II and a rabbit anti-V5 mAb. (B) Spleen DCs cells were isolated from March-I-V5 mice (WT), and cell lysates were subjected to immunoprecipitation using isotype control mAb (control IP) and pMHC-II mAb (pMHC-II IP). The indicated number of cell equivalents of cell lysate or immunoprecipitated were analyzed by SDS-PAGE and immunoblotting with anti-V5 mAb. (C) Spleen DCs cells were isolated from control March-I-V5 mice (WT) or CD83-deficient March-I-V5 mice (CD83 KO). DCs from each strain were lysed and subjected to immunoprecipitation with MHC-II mAb and immunoblotting using anti-V5 or anti-MHC-II β-chain antibodies. The results shown are the average of three independent experiments, and the errors bars are ± S.D., *P < 0.05.