Foreign protein antigens must be broken down within endosomes or lysosomes to generate suitable peptides that will form complexes with class II major histocompatibility complex molecules for presentation to T cells. However, it is not known which proteases are required for antigen processing. To investigate this, we exposed a domain of the microbial tetanus toxin antigen (TTCF) to disrupted lysosomes that had been purified from a human B-cell line. Here we show that the dominant processing activity is not one of the known lysosomal cathepsins, which are generally believed to be the principal enzymes involved in antigen processing, but is instead an asparagine-specific cysteine endopeptidase. This enzyme seems similar or identical to a mammalian homologue of the legumain/haemoglobinase asparaginyl endopeptidases found originally in plants and parasites. We designed competitive peptide inhibitors of B-cell asparaginyl endopeptidase (AEP) that specifically block its proteolytic activity and inhibit processing of TTCF in vitro. In vivo, these inhibitors slow TTCF presentation to T cells, whereas preprocessing of TTCF with AEP accelerates its presentation, indicating that this enzyme performs a key step in TTCF processing. We also show that N-glycosylation of asparagine residues blocks AEP action in vitro. This indicates that N-glycosylation could eliminate sites of processing by AEP in mammalian proteins, allowing preferential processing of microbial antigens.

Intracellular Toll-like receptor 3 (TLR3), TLR7, and TLR9 localize in endosomes and recognize single-stranded RNA and nucleotides from viruses and bacteria. This interaction induces their conformational changes resulting in the production of proinflammatory cytokines and upregulation of cell surface molecules. TLR9 requires a proteolytic cleavage for its signaling. Here, we report that myeloid and plasmacytoid dendritic cells (DCs) deficient for the asparagine endopeptidase (AEP), a cysteine lysosomal protease, showed a decrease in the secretion of proinflammatory cytokines in response to TLR9 stimulation in vitro and in vivo. Upon stimulation, full-length TLR9 was cleaved into a 72 kDa fragment and this processing was strongly reduced in DCs lacking AEP. Processed TLR9 coeluted with the adaptor molecule MyD88 and AEP after size exclusion chromatography. When expressed in AEP-deficient DCs, the 72 kDa proteolytic fragment restored TLR9 signaling. Thus, our results identify an endocytic protease playing a critical role in TLR processing and signaling in DCs.

The glycine-alanine repeat domain (GAr) of Epstein-Barr virus-encoded nuclear antigen 1 (EBNA1) prevents major histocompatibility complex (MHC) class I-restricted presentation of EBNA1 epitopes to cytotoxic T cells. This effect has previously been attributed to the ability of GAr to inhibit its own proteasomal degradation. Here we show, both in vitro and in vivo, that GAr also inhibits messenger RNA translation of EBNA1 in cis and that this effect can be distinguished from its effect on proteasomal degradation. Hence, inhibition of messenger RNA translation, but not protein degradation, is essential to prevent antigen presentation on MHC class I molecules. Thus, by minimizing translation of the EBNA1 transcript, cells expressing EBNA1 avoid cytotoxic T cell recognition. At the same time, blocking degradation maintains the EBNA1 expression level.

Little is known about the processing of putative human autoantigens and why tolerance is established to some T cell epitopes but not others. Here we show that a principal human HLA-DR2-restricted epitope--amino acids 85-99 of myelin basic protein, MBP(85-99)--contains a processing site for the cysteine protease asparagine endopeptidase (AEP). Presentation of this epitope by human antigen-presenting cells is inversely proportional to the amount of cellular AEP activity: inhibition of AEP in living cells greatly enhances presentation of the MBP(85-99) epitope, whereas overexpression of AEP diminishes presentation. These results indicate that central tolerance to this encephalitogenic MBP epitope may not be established because destructive processing limits its display in the thymus. Consistent with this hypothesis, AEP is expressed abundantly in thymic antigen-presenting cells.