Finally, to associate the appearance of the MHC class I dimers described herein with alterations in the redox potential of cells undergoing hydrogen peroxide, Saracatinib nmr thimerosal and anti-CD95 treatments,

we directly measured redox activities using two methods. First we used the water-soluble tetrazolium salt (WST-8) to determine general dehydrogenase activity in the cells, and second we used monochlorobimane, which gives a direct fluorescent readout of intracellular GSH content.22 With both assay systems treatment of cells with hydrogen peroxide and thimerosal resulted in a profound reduction in signal (Fig. 4c,d). Treatment with anti-CD95 resulted in less significant loss of signal, which is a broad agreement with the immunoblotting results of Figs 2–4, where anti-CD95 induces fewer MHC class I dimers. In our previous work, we established that fully folded (i.e. recognized by conformation-specific monoclonal antibodies) MHC class I dimers exist on secretory exosome vesicles, and that these form by disulphide linkage between available cysteine residues in the cytoplasmic tail of many HLA-A and HLA-B molecules.15 In this study we extend

these observations and show that similar MHC class I dimers can be detected on cells in which the redox environment has been significantly altered, either by chemical oxidation with diamide, or chemically induced apoptosis with hydrogen peroxide and thimerosal, or by cross-linking of FasR/CD95. Control of dimer formation was likewise localized to the cytoplasmic tail domain cysteine located at residue 325, found in many HLA-B alleles. This is somewhat in contrast to previous observations wherein HLA-B27 dimer structures Tanespimycin nmr were observed even after removal of the cysteine at position 325,10,23 but this

may potentially be accounted for by the use of different cell lines and overall expression levels of the HLA-B27 heavy chain in different systems. For example, it is notable that in our CEM transfectants there was very little HLA-B27 dimer present in cell lysates in the absence of oxidative stress, as shown in Fig. 2, whereas the Jesthom cell line, which expresses higher levels of cell surface HLA-B27 than the CEM lines, displays dimers under MycoClean Mycoplasma Removal Kit normal conditions. Similarly, we have previously noted that HLA-B27 dimers tend to form in dendritic cells only after activation and significant up-regulation of MHC class I expression.24 Therefore, MHC class I expression levels and the redox status of cells may both contribute to dimer formation. In this current study we also generated a mutant form of HLA-B27 called S42C that mimics the dimer formed by the non-classical HLA-G molecule. None of the treatments applied in this current report significantly increased the dimer population over that already formed in the absence of treatment (Fig. 2a and data not shown), and indeed even the strong oxidant diamide failed to induce the formation of a 100% dimer population in all our studies.