and changes in lipid-rhodopsin interactions by 2 H NMR. The amount of MII formed after photoactivation was determined by UV/vis spectroscopy and the rate of transducin activation studied with a GTPgS-assay. At low rhodopsin concentrations (1/300 and lower) rhodopsin appears to be predominantly mo- nomeric. At rhodopsin/lipid ratios higher than 1/300, the level of oligomeriza- tion increases in a highly cooperative fashion with concentration such that at physiological concentrations rhodopsin is mostly oligomeric. Protein function correlated tightly with rhodopsin oligomerization. Data on the influence of bi- layer properties on the monomer - oligomer transition of rhodopsin and the rate of transducin activation will be presented. 2862-Plat Effects of Membrane Geometry on Voltage-Gated ion Channel Distribu- tion Studied with a Model System Sophie Aimon, Gilman Toombes, Domanov Yegor, Patricia Bassereau. Institut Curie, paris, France. Voltage-gated ion channels are inhomogeneously distributed between the highly curved axons and distal dendrites and the relatively flat soma and prox- imal dendrites. To investigate the effects of membrane geometry on channel distribution and diffusion, we developed a model system based on membrane nano-tubes connected to cell-sized Giant Unilamellar Vesicles (GUVs). KvAP, a bacterial analog of eukaryotic Kv channels [1], was purified, fluores- cently labeled and reconstituted into GUVs. Channel density and homogeneity in GUVs were quantified via confocal microscopy while patch-clamp was used to measure the activity of the reconstituted channels. To study the effect of membrane curvature, we pulled a membrane nano-tube from a GUV and could set the tube radius between 10 nm and 200 nm by varying the tension of the GUV membrane. The concentrations of channels in the tube and GUV were measured via confocal microscopy while diffusion was measured by tracking individual channels labeled with quantum dots. As the tube radius decreased, the channel concentration increased while the diffusion coefficient decreased. Results obtained with this model system should give insight into the diffusion of membrane proteins into and out of synaptic boutons. [1] Ruta et al., Nature 2003, 422 : 180-185. 2863-Plat The Role of Cardiolipin Domains in Protein Localization in Bacterial Cells Lars D. Renner, Douglas B. Weibel. University of Wisconsin Madison, Madison, WI, USA. A central question in cell biology is how the spatial organization of machinery within the cell is established, maintained, and replicated in response to external stimuli. In Eubacteria, our understanding of the spatial and temporal organiza- tion of proteins is beginning to take shape. Many proteins localize to regions of rod-shaped bacterial cells that are characterized by a high intrinsic curvature (e.g. the poles). Recent data suggested that there are geometric cues for the lo- calization of proteins and lipids in bacteria. We present data testing the hypoth- esis that membrane anisotropy at highly curved regions of the cell wall leads to protein localization. This research takes a top-down approach that focuses on a combination of in vivo and in vitro experiments with Escherichia coli cells. To study the response of lipids to the geometry of the cell wall in vitro, we have developed a technique for controlling the curvature of bacterial cells using microstructured polymers and quantitatively measuring the spatial localization of lipids in the resulting membranes. This approach allows to engineer an ‘ar- tificial’ pole with a user-defined curvature into the E. coli inner membrane and to measure the spontaneous localization of lipids and polar proteins to this re- gion of the cell. Using this approach we have determined that a critical radius of curvature of ~1.3mm-1 is required to drive the formation of cardiolipin-rich do- mains in the membrane. We have observed that the bacterial division protein MinD localizes to regions of high curvature and co-localizes with cardiolipin domains. Our data provide support for the curvature hypothesis as a general mechanism for regulating spatial organization in bacterial membranes. This re- search is expanding our understanding of Eubacteria and provides insights into the spatial and temporal dynamics of membranes relevant to cell biology. Platform BD: Protein-Ligand Interactions 2864-Plat Protein Affinity Pattern Calculations using Protein-Fragment Site Identi- fication by Ligand Competitive Saturation (SILCS) E. Prabhu Raman, Wenbo Yu, Alexander D. MacKerell Jr. University of Maryland, Baltimore, MD, USA. We demonstrate the applicability of a computational method, Site Identification by Ligand Competitive Saturation (SILCS) to identify regions on a protein sur- face with which different classes of functional groups interact. The method in- volves MD simulations of a protein in an aqueous solution of chemically diverse small molecules. In the present application, SILCS simulations are per- formed with an aqueous solution of 1 M benzene and propane to map the affin- ity pattern of the protein for aromatic and aliphatic functional groups. In addition, water hydrogens and oxygen serve as probes for hydrogen bond donor and acceptor functionality, respectively. The method is tested using a set of pro- teins for which crystal structures of complexes with several high affinity inhib- itors are known. SILCS simulations are performed for these proteins and the affinity pattern is obtained as 3D probability distributions of fragment atom types on a 3D-grid surrounding the protein called ‘‘FragMaps’’. Good agree- ment is obtained between FragMaps of each type and the positions of chemi- cally similar functional group in inhibitors as observed in the Xray crystallographic structures. For proteins for which inhibitor decoy sets are available, we demonstrate the statistical significance of the SILCS predictions by showing a significantly higher degree of overlap of the ligand atoms in the experimental conformation with the FragMaps. For a few test cases, we corre- late the extent of overlap of the ligand functional groups with FragMaps to the experimental binding affinities. SILCS is also shown to capture the subtle dif- ferences in protein affinity across homologs, information which may be of util- ity towards specificity-guided drug design. Taken together, our results suggest that SILCS can recapitulate the location of functional groups of bound inhibi- tors, suggesting that the method may be of utility for rational drug design. 2865-Plat CHARMM Additive All-Atom Force Field for O-Glycan and N-Glycan Linkages in Carbohydrate-Protein Modeling Sairam S. Mallajosyula 1 , Olgun Guvench 2 , Alexander D. MacKerell Jr. 1 1 University of Maryland School of Pharmacy, Baltimore, MD, USA, 2 University of New England College of Pharmacy, Portland, ME, USA. The O-glycosidic and N-glycosidic linkages are important protein modifications in which oligosaccharides are linked to Ser/Thr and Asn residues, respectively. These linkages involve the anomeric carbon of the carbohydrates and the alcoholic side groups of Ser/Thr or the amide group of the Asn side chain. The O- and N-glycosidic linkages are ubiquitous in biological systems including gly- coproteins like mucin, epidermal growth factor (EGF), domains of different se- rum proteins and Notch receptors, among many others, where the presence of the carbohydrate moiety is important for the biological functions of the proteins. In an ongoing effort to develop the CHARMM all-atom additive carbohydrate force field we present and validate parameters that will enable the modeling of the O- and N-glycosidic linkages. The parameters represent an extension of the existing CHARMM carbohydrate and protein force fields.1-2 The target data for the optimization process included quantum mechanical (QM) potential en- ergy scans of the torsions involved in the glycosidic linkages. Force field val- idations included comparison of the intermolecular geometries for the QM and crystal studies, comparison of the crystalline unit-cell properties and experi- mental NMR J-coupling constants. The optimized parameters were then used to rationalize the differences between the Ser and Thr O-glycan linkages using a Hamiltonian Replica Exchange protocol (HREX). We found that the solvent structure closely governs the linkage geometry due to the involvement of bridged waters between the carbohydrate and protein regions. (1) Guvench, O.; Hatcher, E. R.; Venable, R. M.; Pastor, R. W.; Mackerell, A. D. J. Chem. Theory Comput 2009, 5, 2353-2370. (2) Guvench, O.; Greene, S. N.; Kamath, G.; Brady, J. W.; Venable, R. M.; Pastor, R. W.; Mackerell, A. D. J. Comput. Chem. 2008, 29, 2543-64. 2866-Plat Kinetic Properties of the Two-State Model for Cooperativity Sargis Simonyan, Nadja Hellmann. Institute for Molecular Biophysics, Mainz, Germany. Cooperativity is a regulation mechanism of protein function which is defined by equilibrium binding properties, namely the shape of the oxygen binding curve. This sigmoid shape is the consequence of the existence of different conforma- tions which differ in ligand and effector binding affinity. Positive cooperativity also leads to characteristic behavior in the kinetics of li- gand binding and dissociation. Oxygen dissociation from hemocyanins is typi- cally ‘‘auto-catalytic’’ since the off-rate for oxygen is slower for the initial high affinity state than for the final low affinity state. The relative change of the off-rate at the beginning of the reaction compared to the final phase might serve as a measure for kinetic cooperativity. We compared the oxygen dissoci- ation kinetics of 6 different arthropod hemocyanins. The modulation of the kinet- ics by allosteric effectors in most cases is what might be expected, leading to an increased rate for negative effectors and an decreased rate for positive effectors. A rather unexpected finding was the mostly linear change in apparent rate of dissociation with decreasing saturation degree. Simulations based on the MWC-model showed that the observed relationship between off-rate and satu- ration degree is typical for relatively slow conformational transitions and not- too-large allosteric equlibrium constants. It can be shown that under the condi- tions employed here (high cooperativity) the maximal increase in the off-rate is 526a Wednesday, March 9, 2011