Sua Myong    Spring-Loading Mechanism of DNA Unwinding  by HCV NS3 helicase Institute of Genomic Biology, UIUC NS3 is an essential helicase for replication of Hepatitis C virus and is a model enzyme for investigating helicase function. Using single molecule fluorescence analysis, we shows that NS3 unwinds DNA in discrete steps of approximately 3-bp.  Dwell time analysis indicates that about three hidden steps are required before a 3-bp step is taken.  Combined with available structural data, we propose a spring-loading mechanism where several 1nt/ATP steps accumulate tension on the protein/DNA complex which is relieved periodically via a burst of 3bp unwinding. Jiahao Chen    Polarization and Charge-Transfer Effects in Classical Molecular Dynamics Chemistry, UIUC  Polarization and charge transfer are important electrostatic phenomena that are not described in conventional molecular models. Of the several common types of models used to model polarization, only fluctuating-charge models, a.k.a. chemical potential equilibration or electronegativity equalization methods, are able to describe both effects. Such models are known to suffer from deficiencies such as unphysical charge transfer between dissociated molecular species at infinite separation. We outline a new  model that cures some of these deficiencies and present some applications to simple chemical systems. Yuan-Hung Chien    Single Cell Studies of Adhesion Dynamics Biohemistry, UIUC The pre-steady state kinetics of cell adhesion mediated by cadherin was studied by micropipette manipulation measurements. The binding probability versus contact time curves reveal two binding states. There is a first, fast binding stage followed by a slower binding step. This biphasic profile differs from the simple rise to a limiting steady state binding probability that is predicted for molecules that interact via single binding site. Measurements with the truncated extracelluar region of cadherin which lacks the transmembrane and cytoplasmic domains also exhibit a biphasic profile. This indicates that the binding kinetics are independent of the cytoplasmic domain.  Further studies with mutants lacking different domains showed that the third domain (EC3) of the extracellular region is required for both the two-stage binding mechanism and the slow forming second binding state. Mutating the critical Trp2 residue also abolished the two stage kinetic process. These findings reveal a more complex kinetic binding process than predicted on the basis of the crystal structure. Initial cadherin binding exhibits two adhesive states that require different extracellular domains, but not the cytoplasmic region. Ionel Rata     The Protein Sequence-Structure Paradigm Starts at the Backbone Description Level Biophysics and Computational Biology, UIUC Native proteins have been optimized by evolution simultaneously for structure and sequence. The available data in the structural libraries is distributed accordingly and a statistical potential should consider the structure and sequence distribution features together, accounting for their interdependency. We show that the sequence-structure paradigm starts at the backbone level by proving that the existing correlations between the phi-psi backbone torsion angles and the local sequence have an important contribution in determining the global configuration of protein loops. Based on these correlations we propose a new statistical backbone potential that has both structure (torsion angles) and sequence (residue types) characteristics as variables. For our statistical approach we collect relevant information from selected loop data available in Protein Coil Library, because in loops the non-local interactions are randomly distributed and they average out in statistics, so local interaction can be studied separately. We organize the available data in the form of local sequence-specific phi-psi torsion angle pairs. From these data, with an improved statistical method we construct two-dimensional distribution plots in the phi-psi dimensions for all the possible neighboring residues and residue pairs. We use a probabilistic analysis to deduce how the nearest neighbor correlations encoded in the various correlation plots propagate along the loop backbone. The application of Boltzmann statistics allows us to convert the resultant probability to a potential of mean force that is able to score a loop by its backbone structure and sequence. Our potential is able to discern with high accuracy the native structure of a loop with a given sequence among possible alternative conformations from sets of well-constructed decoys. Reversely, our potential can also be used for sequence prediction problems and is shown to accurately select the native sequence of a given (fixed) loop structure among the best scored of all possible alternative sequences. These results highlight the application of our approach at the basis of both structural prediction and sequence design problems.