We have exploited Drosophila melanogaster Schneider cells and compatible inducible expression vectors to produce large amounts of secreted major histocompatibility complex (MHC) class II molecules (I-Ed). A simple two-step purification protocol was developed. In the first step, recombinant molecules were enriched using a monoclonal anti-class II antibody column followed by a nickel chelate column which further purified and concentrated the recombinant protein to several mg/ml. Characterization of the purified material indicates that the molecules are correctly assembled into alpha beta heterodimers. Further analysis shows that the recombinant MHC class II molecules are devoid of endogenous peptides and, therefore, homogeneous peptide/MHC complexes could be prepared by adding exogenous I-Ed-specific peptides at slightly acidic pH. Upon peptide addition, molecules underwent a conformational change into a more compact form revealed by gel filtration analysis. In addition, the peptide/MHC complexes were biologically active. As little as 10 ng of these complexes coated on plastic from a 100 ng/ml solution were sufficient to trigger antigen-specific T cell hybridomas. These MHC class II molecules, together with various forms of soluble T cell receptor (TcR) proteins, provide valuable tools to analyze the molecular details of TcR/antigen recognition.

The striking and unique structural feature of the T cell receptor (TCR) β chain is the bulky solvent-exposed FG loop on the Cβ domain, the size of almost half an immunoglobulin domain. The location and size of this loop suggested immediately that it could be a crucial structural link between the invariant CD3 subunits and antigen-recognizing α/β chains during TCR signaling. However, functional analysis does not support the above notion, since transgene coding for TCR β chain lacking the complete FG loop supports normal α/β T cell development and function.

A striking feature of the T cell receptor (TCR) β chain structure is the large FG loop that protrudes freely into the solvent on the external face of the Cβ domain. We have already shown that a transgene-encoded Vβ8.2+ TCR β chain lacking the complete Cβ FG loop supports normal development and function of conventional α/β T cells. Thus, the FG loop is not absolutely necessary for TCR signaling. However, further analysis has revealed that a small population of α/β T cells coexpressing NK1.1 are severely depleted in these transgenic mice. The few remaining NK1.1 T cells have a normal phenotype but express very low levels of TCR. We find that the TCR Vβ8.2+ chain lacking the Cβ FG loop cannot pair efficiently with the invariant Vα14-Jα281 TCR α chain commonly expressed by this T cell family. Consequently, fewer NK1.1 T cells develop in these mice. Our results suggest that expression of the Vα14+ TCR α chain is particularly sensitive to TCR-β conformation. Development of NK1.1 T cells appears to need a TCR-β conformation dependent on the presence of the Cβ loop that is not necessarily required for assembly and function of TCRs on most α/β T cells.