The interaction between MHC:peptide complexes and the T cell receptor directs the specificity of the adaptive immune response, governing both clearance of infection and maintenance of self tolerance. MHC class II (MHC-II):peptide complexes are displayed by professional antigen presenting cells, which are able to direct the quality of the immune response through various cues, such as co-stimulation and cytokine release. It is therefore of critical importance that MHC-II:peptide complexes remain stable during this process, so that the peptides presented reflect the environment in which they were acquired. In order to maintain the fidelity of the peptide repertoire, MHC:peptide complexes tend to have extraordinarily long half-lives (sometimes in the order of weeks) and MHC-II molecules that lose peptide tend to quickly collapse to ensure opportunistic peptide binding cannot occur. These properties of MHC-II molecules must be overcome during peptide loading, when a placeholder peptide termed CLIP is exchanged for antigenic peptides by an MHC-like molecule called HLA-DM. In an endocytic compartment, HLA-DM causes peptide dissociation from MHC-II, edits the peptide repertoire so that strongly binding peptides are favoured and prevents collapse of empty MHC molecules. One HLA-DM molecule can promote peptide exchange of many MHC-II molecules and it is for this property that HLA-DM has been likened to an enzyme. The crystal structure of HLA-DM in complex with HLA-DR1 reveals the mechanism of peptide editing by HLA-DM and captures an intermediate in this complex kinetic process¹. Here I describe the protein engineering ‘tricks’ that were used to stabilise the complex and finally allow atomic resolution structure determination of an MHC-II molecule caught in the act of peptide-exchange.