Decoding NK cell receptor specificity : Functional and structural studies of MHC class 1 subcomponents

Abstract: Natural Killer (NK) cells are an important part of the innate immune system. They can lyse target cells and secrete immunoregulatory cytokines early during immune responses. They mediate protection against viruses and tumours, and can reject MHC class I mismatched bone marrow grafts. NK cell function is regulated through activating and inhibitory receptors, many of which recognize MHC class I molecules on the surface of surrounding cells. The overall aim of this thesis has been to elucidate the MHC class I specificity of NK cell receptors. We have investigated the role of the heavy chain, the beta2m subunit and the peptide of MHC class I complexes in the interaction with NK cell receptors, primarily Ly49 receptors in the mouse and CD94/NKG2 receptors in the human. The MHC class I specificity of NK cell receptors was studied by direct visualisation using soluble, fluorescently labelled MHC class I tetramers, by X-ray crystallography and in functional assays. After having established the novel technique of soluble, fluorescently labelled tetrameric MHC class I molecules in our laboratory, we assessed the binding of such MHC class I tetramers (H-2D , 14- 2Kb and H-2D) to different Ly49 receptors. We could demonstrate that they bind in an allelespecific manner to Ly49 receptors and concluded that the MHC class I associated glycosylation was not required for the interaction between H-2Dd , H-2Db and Ly49A, nor for the interaction between H-2Kb, H-2Db and Ly49C. Furthermore we could show that the interaction between Ly49C and H-2Kb was peptide-selective, confirming and extending earlier functional data from our laboratory in a direct binding assay. Finally we identified H-2D b as a new ligand for both Ly49A and Ly49C. The role of the MHC class I bound beta2M subunit in recognition by Ly49 receptors was investigated using MHC class I tetramers of H-2Dd , H-2Kb and H-2Db refolded with either mouse or human beta2m. The change from mouse to human beta2m resulted in a loss of binding of all three MHC class I tetramers to Ly49A and Ly49C, indicating that both these receptors bind to MHC class I in a similar manner, partly involving the beta2m-subunit. To investigate the influence of beta2m on the structure of MHC class 1, in particular in relation to recognition by Ly49 receptors, we crystallized and solved the structures of H-2Db in complex with an LCMV derived peptide and either human or mouse beta2m. This allowed us for the first time to make a comparative analysis of two MHC class I crystal structures where only the 02M subunit differed. Despite over 30% difference in sequence between mouse and human beta2m, the conformation of the peptide and the peptide-binding groove were strikingly similar between the structures. The structures implicate that the lack of Ly49 recognition conferred by human beta2m is a direct consequence of the change from a glutamine to a glycine at position 29 in human beta2m, rather than distal conformational changes. Finally, we identified two heat shock protein 60 (hsp60) derived peptides that readily bind to HLA-E. We demonstrated that HLA-E in complex with these peptides abrogated recognition by the inhibitory CD94/NKG2A receptor, both in binding assays and functional assays. Our results suggest that there is an increased proportion of HLA-E molecules in complex with non-protective hsp60-derived peptides during cellular stress, and consequently a reduced inhibition of CD94/NKG2A+ NK cells. This phenomenon, which we term stress-induced peptide interference (SPI), provides a novel mechanism for NK cells to detect stressed cells in a peptide-dependent manner. In conclusion, our data supports the notion that both human and mouse NK cells have evolved a repertoire of MHC class I specific receptors which are sensitive to perturbations in all subcomponents of the MHC class I molecule