Parts one to three, as listed above deal with the MHC-I aspects of this project, and in general are directed to understand the molecular details of the loading of MHC-I molecules with self or antigenic peptides. Although hundreds of MHC-I and MHC-I-like three-dimensional structures have been determined, none of these is of a peptide-receptive (PR) form of the molecule. In our studies, we determined the three-dimensional structure of a peptide epitope representative of a portion of the MHC-I molecule H2-Ld that is exposed only on partially unfolded, peptide receptive (PR) MHC-I molecules. The determination of this structure was achieved by recognizing that a unique monoclonal antibody, 64-3-7, binds only to PR molecules and not to peptide loaded (PL) molecules. Using surface plasmon resonance, we carefully mapped the part of the MHC-I molecule recognized by this antibody, and then we successfully obtained crystals of the Fab protein complex with each of three different, but overlapping peptides. Synchrotron diffraction data were collected on four such crystals and all data sets were appropriately scaled. Molecular replacement solutions of the structure were obtained using homologous antibody fragments as the search model, and the structures were refined to high resolution (from 1.64 to 2.0 angstrom resolution), using several programs, including CNS, Refmac, and Phoenix. The four structures determined were essentially identical, and revealed that a portion of the 64-3-7 epitope (a sequence of seven amino acids) remains intact as a 3,10 helix in the Fab-complexed form, with several amino acid side chain adjustments. This suggests that this seven-residue peptide forms a molecular hinge that is first exposed to solvent in the PR form, and subsequently several of its side chains are then sequestered from solvent in the PL form. To visualize the structure of the entire peptide receptive, PR form of the H2-Ld molecule, the structure of the seven-residue epitope was spliced (in silico) into the known crystal structure of the complete H2-Ld molecule. To confirm the veracity of this model, and that the antibody 64-3-7 was capable of recognizing this same epitope as spliced into the context of the whole H2-Ld molecule, we performed a molecular docking simulation, using the Rosetta Dock program run on the high-performance computational resource of the Biowulf Linux cluster at the NIH. Of 10,000 docking solutions generated, the top scoring 10% were identified and of the 10 top scoring clusters, one solution containing a loop conformation with the residues of the peptide epitope in a similar conformation to those observed in the crystal structures was obtained. This then defined the PR form of the molecule, which was used as the input for molecular dynamics simulations of a fully hydrated model, using NAMD, a scalable molecular dynamics program, also run on the NIH Biowulf Cluster. This dynamics simulation has been examined extensively and provides a structural understanding of the way that MHC-I molecules change their shape from the metastable PR form to their stable PL form. Thus, for the first time, we have dynamic images of the conformational changes that accompany the transition from peptide receptive to peptide bound forms of the MHC molecule. Detailed analysis of the two structures and of the transition from PR to PL suggest a detailed mechanism of how the MHC-I molecule works. To extend our understanding of the nature of pepide-loading, we have engineered the main chaperones involved in MHC-I loading, tapasin and Erp57. In addition, we have engineered a tapasin like molecule, known as TAPBPR, which is about 20% idnetical in amino acid sequence to tapasin and have embarked on studies examining the nature of its binding to the PR form of MHC-I. Preliminary studies suggest that TAPBPR interacts with a peptide-free, peptide-receptive form of MHC-I, and that this interaction is relaxed upon peptide binding. We expect that further studies of the TAPBPR interaction and structure will provide broad insight into tapasin function as well. Further related to MHC-I/peptide interactions, we have explored the role that the anti-retroviral drug, abacavir plays in binding to MHC-I and distorting the self-peptide repertoire bound by susceptible MHC-I alleles. In particular, we have shown, by peptide sequence analysis of the self peptides bound to the susceptible MHC-I allele, HLA-B*57:01, in the presence and absence of abacavir, that this drug can change the peptides that B*57:01 binds. This provides an explanation for the severe hypersentivity reactions that are observed in a high proportion of HLA-B*57:01 individuals who receive the drug. The fourth part of this project is focused on structural and functional studies of T cell receptor recognition of antigens and how this leads to autoimmune disease. To provide a baseline for understanding antigen-specific structural changes in the TCR, we have determined the X-ray structure of a viral specific, MHC-I-restricted TCR, as well as its complex with its MHC-I/viral antigen ligand. Remarkably, although the MHC/peptide complex has a relatively rigid structure, the TCR shows great movement of its CDR3 alpha and beta loops, indicative of a fly-casting mechanism for ligand engagement. Other approaches to understanding the TCR mediated aspects of autooimmunity include include: 1) the characterization of antigenic peptides recognized by the autoimmune T cells in transgenic mouse models of autoimmune gastritis;2) the structural determination of MHC-II molecules covalently linked to their antigenic peptides;and 3) the molecular characterization of MHC-II molecules loaded with autoimmunogenic peptides. Previous work from our laboratory showed that two different T cell receptors derived from mice with autoimmune gastritis were differentially pathogenic in inducing disease. One caused acute, flagrant disease and a Th1 response, one caused indolent disease and a Th2 response. We have determined the antigenic peptides recognized in the context of the MHC-II molecule, IAd, by these two TCR, and defined the frame of binding and the crucial binding and TCR residues of the peptides. In addition we have determined the high resolution X-ray structure of IAd in complex with the Th2 peptide known as PLL, as well as the structures of two other IAd/peptide complexes in which the peptides are related to PLL, but are of higher intrinsic affinity. These structures determined at 2.5 angstrom resolution (and the others to somewhat lower resolution) reveal a previously unrecognized binding motif (exploiting residues 1,4,6,7, and 9 of the peptide) for IAd, particularly with respect to the preference of glutamic acid at position 9 of the peptide. This provides a framework for understanding interactions of autoimmune TCR with self-MHC-II/peptide complexes. These experimental structural studies permitted the modeling of another gastritis-inducing peptide, known as PIT, bound to IAd, and provide further insight into the molecular basis of autoimmune gastritis. Similarities in this theme of the relevant anchors and the topology of the peptide bound to the MHC are consistent with other MHC-II/autoantigen complexes and suggest some common features that may be specifically relevant to autoimmune antigens. In a collaborative project between Dr. Natarajan in the Molecular Biology Section, and the the laboratory of Dr. Rachel Caspi of the NEI, Dr. Natarajan has contributed his expertise in generating TCR transgenic mice to the establishment of a TCR transgenic mouse model for autoimmune uveitis.
Showing the most recent 10 out of 19 publications
