Formular las estrategias para la exportación de yogur

The full length Ii trimer is known to associate with three MHC Class II hetero- dimers containingα(MHCα) andβ(MHCβ) chains. This interaction is known to be mediated by the association of the CLIP domain of Ii occupying the antigen binding site of the MHCα/βhetero-dimer. Castellino et al mapped a further site of interaction between Ii and MHC to the transmembrane segment of Ii using mutagenesis (Castellino, Han et al., 2001).

The work presented in Chapters 3 and 4 of this study reveals that the TM domains of Ii and MHCαand MHCβproteins are sites of potentially important helix-helix interactions that may further stabilise the Ii-MHC complex. Using the limited number of methods available for studying hetero-interactions between hydrophobic proteins we sought to determine if in isolation the TM domain of Ii could associate with those of MHCαand MHCβand determine the stoichiometry of these interactions. This will verify the findings of Castellino et al. and provide the basis for further study into identifying the residues involved in these interactions.

A method for studying the hetero-association of isolated TM domains in vivo is the GALLEX assay. Use of this assay suggested that the TM domain of Ii can interact with those of MHCαand MHCβin theE.coliinner membrane. This is the first time that such an interaction has been observed between these proteins. It is important to remember that GALLEX can only report on association and does not provide details on the oligomeric state of the interaction so that information has to be obtained using other techniques.

In order to confirm the findings from the GALLEX assay, and determine the stoichiometry of these interactions, model peptides derived from the TM domains of Ii, MHCαand MHCβwere produced and their association was monitored using

biophysical techniques. Chemical cross-linking of the peptides in a mild detergent hinted at the possible association of these domains since differences could be observed between the peptides cross-linked separately and when mixed. The analysis of such an experiment is complicated by to the similar size of the peptides under study and the complications involved when those peptides can self-associate as is the case for Ii, MHCαand MHCβ. Cross-linking is not frequently employed to study hetero-interactions of TM domains for this reason. FRET analyses of fluorophore labelled model peptides suggested that the TM domains of Ii and MHCαcould associate. Furthermore, the oligomeric states of that association could be determined and were found to from dimers or trimers depending on the peptide:micelle molar ratio. FRET analyses also suggested that the TM domain of Ii may interact with that of MHCβbut the findings were difficult to interpret conclusively. It is possible the solubility of the MHCβpeptide in the chosen detergent DPC, is an issue for this type of analyses. Intriguingly, it was found that the FRET between Ii with MHCβcould not be enhanced by the presence of MHCα. This data, is consistent with the finding in the full length proteins that individualα-chains coisolate with Ii, whereasβ-chains exhibit only a low-affinity interaction with Ii (Neumann and Koch, 2005). Therefore, the TM domain of MHCβwould not appear to play a role in the assembly of the Ii-MHC complex. It should be noted that given the findings from thein vivoanalysis of the association of Ii and MHCβperformed using the GALLEX assay (see Section 5.2), which suggested they were interacting, we would expect to see significant interaction in the FRET analyses also. The reasons for this discrepancy are uncertain at present, but it is possible that they may reflect the differing environment in which the analyses were performed, since the GALLEX assay is performed on proteins inserted into the inner membrane of E.coli compared to peptides solubilised in detergent for FRET measurements. Further study will be required to conclusively determine if the model peptides of Ii and MHCβTM domains are interacting.

6 Developing

NMR

Methods

for

investigating protein interactions

Identifying helix-helix interactions and the structural determinants that drive α- helical membrane protein folding is a technically challenging problem due to their hydrophobicity. Of particular interest is accessing detailed structural information regarding the interacting side chains and hence enabling identification of the non- covalent bonds stabilising the association of the helices. Using current methods for studying TM interactions (e.g. TOXCAT, GALLEX, Cross-linking, FRET, AUC) it is possible to identify homo- and hetero- association between TM domains and to also determine their oligomeric state, as has been shown in the previous chapters of this work. Current studies on the hetero-association of TM domains have used the GALLEX assay, FRET experiments and immune- precipitation. Complementing these techniques, molecular modelling and mutagenesis studies can suggest residues that are important for the interactions. However, the results from these studies in our experience are often difficult to interpret conclusively. Furthermore, ultimately these methods, though powerful, do not provide information on the arrangements of atoms in these oligomers and ultimately the definitive test for any predictions made from such studies is to solve the atomic structure of the domains.

Currently the two most successful methods for solving the atomic structures of membrane proteins are X-ray crystallography and NMR spectroscopy. In particular, solution-state NMR spectroscopy has been applied to the study of transmembrane domains in isolation, which due to their small size are most amenable to analysis by this technique. We therefore sought to develop protocols for using NMR to solve the structure of TM domain oligomers and additionally designed a novel assay for determining the association of these domains that also has the potential to provide important structural information.

This section describes work carried out in this study on developing these NMR – based methodologies for identifying interacting α-helical TM domains and for determining the atomic structure of those oligomers. Preliminary work has been

performed on solving the structure of a TM domain from the well characterised E5 protein and on developing paramagnetic NMR methods for the rapid determination of TM domain structure. It is hoped that the results from these studies will facilitate future investigation of the Ii, and MHC proteins.

6.1 Towards solving the structure of the TM domain