The newly completed draft of the human genome has provided new insights into the importance of membrane proteins in all cells. In most eukaryotes, over one-third of open reading frames are predicted to be integral membrane proteins with various number of membrane-spanning segments. Membrane proteins provide numerous vital cellular functions that include cell-cell communication such as recognition, adhesion, signal transduction and membrane fusion; material exchange including transportation and detoxification; and cellular energy conservation. Structures of a limited number of membrane proteins have contributed significantly to our understanding of functions of these biological macromolecules. The demand for structural knowledge of membrane proteins has increased more than ever in light of increased number of membrane proteins for which important functions have been identified. However structural data of membrane proteins are only being obtained rather slowly, mainly due to the tremendous difficulty in obtaining sufficient quantity of membrane proteins, especially those of eukaryotic origin, and in producing diffraction quality crystals. My lab, in collaboration with both intramural and extramural laboratories, is trying to explore structure and function relationships of a few selected polytopic membrane proteins. Our efforts are concentrated on those involved in cellular multidrug resistance (MDR) such as the human P-glycoprotein (P-gp) and its homologs from different species, and in cellular respiratory chain such as the cytochrome bc1 complex (cyt. bc1 or bc1) of both mitochondria and bacteria. Our goal is to gain structural insights into functions of these proteins, which, we hope, will aid in efforts of developing reagents that have potential therapeutic value. Cyt. bc1 is an essential component of the cellular respiratory chain and the photosynthetic machinery. It catalyzes the reaction of electron transfer (ET) from quinol to cyt. c and concomitantly translocates protons across cellular membrane to generate a proton gradient for various cellular functions. Existing as a dimer in the membrane with a molecular mass near 500 kDa, the bovine mitochondrial bc1 consists of 11 different subunits and have 26 trans-membrane helices. Three subunits, cyt. b, cyt. c1, and the iron-sulfur protein (ISP), are considered important for the ET function. In collaboration with Dr. C. A. Yu's laboratory of Oklahoma State University, the following progresses have been made: First, after careful refinement of protein purification and crystallization protocols, we have successfully diffracted the bovine bc1 crystals to 2.6 ? and 2.4 ? resolution, respectively, for native and with bound famoxadone, a commercial fungicide. We subsequently refined the structures of bc1 to corresponding resolutions. Second, the refinement allowed completion of 11 subunits, modeling of bound lipids and substrate, as well as placing solvent molecules. Third, we also refined bc1 structures with various respiratory inhibitors bound (antimycin A, NOHO, UHDBT, azoxystrobin, methoxyl-acrylate stilbene, etc.), for which only lower resolution diffraction data are available, based on refined structures of high resolution. Fourth, the refined structures permit analysis in great detail of structure of each subunit, structural alterations as a result of inhibitor binding, mechanisms of inhibitor binding, correlating structural data with genetic data, and proposing novel mechanism for bc1 function. Fifth, results from these studies are being submitted for publications. We have one Biochemistry paper in press, and three more in preparation. We also obtained the four-subunit bc1 complex crystals of R. spheroidaes (RS), a photosynthetic bacteria, which diffracted X-rays to 3.5 ? resolution. Unlike its mitochondrial counterpart, the bacterial bc1allows easy genetic manipulation. We are currently searching for better cryo conditions to stabilize RS bc1 crystals, making them suitable for synchrontron data collection and for heavy atom derivatives. Human P-gp is a major contributor to the multidrug resistance of cancer chemotherapy; it is an integral membrane protein with 12 predicted membrane-spanning segments and two ATPase modules. In collaboration with Drs. Ambudkar and Gottesman at LCB, we have been striving to obtain sufficient amount of P-gp. Progress has been made over the last few years in producing enough protein in insect cell culture to initiate crystallization experiments. Several issues remain: First, consistent batches of proteins in mg quantity are needed, efforts are being sought to scale up the production of P-gp producing cells. Second, the quality of protein preparation requires more work. By using the (His)10 tag, the quality of the P-gp preparation from the one step purification procedure has improved. But SDS-page shows broadening of the P-gp band and significant number of minor bands. Further purification by applying the sample to a gel filtration column to remove small molecular contaminants and possible aggregates of P-gp is viewed necessary for producing high quality protein amenable for crystallization. Third, the solubility of P-gp preparation has to improve. Currently, P-gp precipitates at a protein concentration above 2 mg/ml. Several strategies are being employed. One is to test different detergents or detergent combinations, this approach has increased the solubility from previous 1.2 mg/ml. Another approach is to attach the protein with soluble protein fragments such as antibody fragments to enhance solubility. We have made effort to generate Fab and Fv fragments of a monoclonal antibody that binds specifically to the extracellular region of P-gp and is sensitive to P-gp conformation. Constructs of Fab and scFv have been made, but expressions in E. coli have not been optimized, leading to either low yield or large amount of inclusion bodies. Constructs using different promoters, expression under various conditions and in eukaryotic cells are now being investigated. Finally, in light of the low quantity and concentration of P-gp preparation, we are employing crystallization approaches that consume sub-ml volume of proteins for each experiment, and devising P-gp specific crystallization screen that give considerations for P-gp substrates, modulators, and ATP analogs. We have also been working on MDR homologs from other species; one successful example is the cloning and expression of bacterial LmrA of L. lactis. Large amount of protein (~20 mg) can be produced after one affinity and one gel filtration column purification. Crystallization experiments are underway. Energy-dependent intracellular protein degradation is needed for cell growth, mediating stress responses, and supervising protein quality control. ATP-dependent proteases, such as the Clp proteases and proteasomes, are composed of a proteolytic core particle, in which the active sites are compartmentalized, and an ATP-dependent regulatory particle with ATPase and chaperone activity. In E. coli ClpAP and ClpXP, the functional units are ClpP, a protease core of 14 identical subunits in two stacked heptameric rings and either ClpA or ClpX, ATP-dependent chaperones consisting of six identical subunits arranged in hexameric rings. The ATPase subunits are responsible for substrate selection, protein unfolding and translocation to the proteolytic core, and allosteric modulation of the ClpP activity; they also have stand-alone chaperone activity, catalyzing limited structural remodeling of proteins and disassembly of stable protein complexes in the absence of the proteolytic component. ATP binding and hydrolysis have distinct roles and are essential in various activities of ClpA or ClpX. In collaboration with Dr. Maurizi at LCB, we have determined and refined the crystal structure of ClpA, the regulatory component of the ClpAP complex. ClpA consists of five tandemly connected structural domains corresponding to three functional groups. The N-terminal domain represents a novel fold of a repeating motif with pseudo two-fold symmetry. The two AAA+ modules (D1 and D2) are connected head-to-tail with a 90? rotation. In the crystal, the D1 and D2 domain of ClpA subunit interact mostly with neighboring D1 and D2 domains, respectively, in which the D1-D1 interface has considerably more electrostatic contacts than the D2-D2 interface, providing structural basis for different contributions of the two AAA+ modules to hexameric assembly and ATPase activity. A symmetrical hexameric ring model of ClpA locates a ClpP-interaction loop on the distal surface of D2, reveals a large negatively charged central cavity, and identifies two narrow constrictions speculated for substrate disruption. These studies will provide a framework for further experiments in understanding sub-domains and specific amino acid side-chains involved in substrate recognition, ATP hydrolysis and the accompanying conformational changes, interactions leading to assembly of the chaperone and the protease complex, and the mechanism of protein unfolding, translocation, and degradation carried out by this and similar essential regulatory chaperones. We also crystallized and determined the structure of the N-domain of ClpA in complex with ClpS, the adaptor protein that binds specifically to ClpA and alters the substrate specificity of ClpA. The ClpS structure forms an a/b-sandwich and is topologically analogous to the C-terminal domain of the ribosomal protein L7/L12. ClpS contacts two surfaces on the N-terminal domain in both crystal forms; and the more extensive interface was shown to be favored in solution by protease protection experiments. Biochemical experiments show that ClpS lacking N-terminal 16 residues still binds to ClpA N-domain but no longer inhibits ClpA activity. A zinc binding site involving two His and one Glu residues is identified crystallographically in the N-terminal domain of ClpA. In a model of ClpS bound to hexameric ClpA, ClpS is oriented with a long N-terminal extension directed toward the distal surface of ClpA, suggesting that the N-terminal region of ClpS may affect productive substrate interactions at the apical surface or substrate entry into the ClpA translocation channel.
