Integrins are functionally and structurally complex surface molecules involved in multiple aspects of vascular function including development, thrombosis, and inflammation. Current understanding and future work on integrins would be enormously advanced by a three-dimensional structure of a representative integrin. Integrins are large heterodimers, and therefore it is logical to both attempt to the structure of intact heterodimers, and modular units. Approaches are presented to optimize obtaining crystallizable heterodimers or truncated heterodimer fragments. Genetically engineered extracellular por5tions of integrin alpha/beta heterodimers, either complexed with ligand or not, will be designed, produced, crystallized, and subjected to structural determination. Work is piloted with the alpha5beta1 integrin; backup integrins that are low in glycosylation and have multiple ligands or few conformation states are alphaIIbbeta3 and alpha6beta4. Integrins will be stabilized in a single conformation state. High level secretion is obtained by fusion to disulfide-liked coiled-coils. Minimized heterodimers are already produced that bind ligand. These and other fragments are used to explore integrin function. In parallel, work is under way on previously and newly defined modular units within integrins. Beta subunit fragments to be tested include an internal deletion of the I-like domain that already yields crystals; integrin EGF domains that already yield NMR data; and Pactolus, a beta subunit homologue that does not associate with an alpha subunit and lacks a specificity-determining loop. The work on integrins will also be illuminated by work on structurally and functionally related domains, e.g. beta-propeller domains in nidogen or bacteria, that may be more easily isolated as modules and solved structurally. A putative cyanobacterial beta-propeller has sequence homology to the putative integrin beta-propeller domain in the extracellular protein nidogen, in complex with a fragment of its ligand laminin, will be structurally determined. A related YVTD module in a bacterial protein has already yielded atomic resolution diffraction. The long-term goal of this project is to define the ligand binding sites in integrins, how alpha and beta subunit modules come together in the head piece to bind ligand how the alpha and beta subunit stalk regions interact with one another to regulate ligand binding, and how the ligand-binding head piece modules interact with one another and with stalk modules to transduce signals from within the cell that modulate ligand binding, and to transduce signals from ligand binding into the cell. The molecular pathologies of Glanzmann's thrombasthenia and epidermolysis bullosa will be defined. Binding site structures will advance the development of drugs that antagonize integrins.
Showing the most recent 10 out of 43 publications
