Session F45: Focus Session: Physics of Proteins I

Exploring the landscape for protein folding: from function to molecular machines

Globally the energy landscape of a folding protein resembles a partially rough funnel with reduced energetic frustration. A consequence of minimizing energetic frustration is that the topology of the native fold also plays a major role in the folding mechanism. Some folding motifs are easier to design than others suggesting the possibility that evolution not only selected sequences with sufficiently small energetic frustration but also selected more easily designable native structures. The overall structures of the on-route and off-route (traps) intermediates for the folding of more complex proteins are also strongly influenced by topology. Going beyond folding, the power of reduced models to study the physics of protein assembly, protein binding and recognition, and larger biomolecular machines has also proven impressive. Since energetic frustration is sufficiently small, native structure-based models, which correspond to perfectly unfrustrated energy landscapes, have shown to be a powerful approach to explore larger proteins and protein complexes, not only folding but also function associated to large conformational motions. A discussion of how global motions control the mechanistic processes in the ribosome and molecular motors will be presented. For example, this conceptual framework is allowing us to envisage the dynamics of molecular motors and the ribosome from the structural perspective and it provides the means to make quantitative predictions that can be tested by experiments.   

Predicting folding-unfolding transitions in proteins without a priori knowledge of the folded state

The common computational method of studying folding transitions in proteins is to compare simulated conformations against the folded structure, but this method obviously requires the folded structure to be known beforehand. In the current study, we show that the use of bond orientational order parameter (BOOP) Q$_{l}$ [Steinhardt PJ, Nelson DR, Ronchetti M, Phys. Rev. B 1983, 28, 784] is a viable alternative to the commonly adopted root mean squared distance (RMSD) measure in probing conformational transitions. Replica exchange molecular dynamics simulations of the trp-cage protein (with 20 residues) in TIP-3P water were used to compare BOOP against RMSD. The results indicate that the correspondence between BOOP and RMSD time series become stronger with increasing \textit{l}. We finally show that robust linear models that incorporate different Q$_{l}$ can be parameterized from a given replica run and can be used to study other replica trajectories.   

Exploring Beta-Amyloid Protein Transmembrane Insertion Behavior and Residue-Specific Lipid Interactions in Lipid Bilayers Using Multiscale MD Simulations

Beta-amyloid (Abeta) interactions with neurons are linked to Alzheimer's. Using a multiscale MD simulation strategy that combines the high efficiency of phase space sampling of coarse-grained MD (CGD) and the high spatial resolution of Atomistic MD (AMD) simulations, we studied the Abeta insertion dynamics in cholesterol-enriched and -depleted lipid bilayers that mimic the neuronal membranes domains. Forward (AMD-CGD) and reverse (CGD-AMD) mappings were used. At the atomistic level, cholesterol promoted insertion of Abeta with high (folded) or low (unfolded) helical contents of the lipid insertion domain (Lys28-Ala42), and the insertions were stabilized by the Lys28 snorkeling and Ala42-anchoring to the polar lipid groups of the bilayer up to 200ns. After the forward mapping, the folded inserted state switched to a new extended inserted state with the Lys28 descended to the middle of the bilayer while the unfolded inserted state migrated to the membrane surface up to 4000ns. The two new states remained stable for 200ns at the atomistic scale after the reverse mapping. Our results suggested that different Abeta membrane-orientation states separated by free energy barriers can be explored by the multiscale MD more effectively than by Atomistic MD simulations alone.   

Combined copper/zinc attachment to prion protein

Misfolding of prion protein (PrP) is responsible for diseases such as ``mad-cow disease'' in cattle and Creutzfeldt-Jacob in humans. Extensive experimental investigation has established that this protein strongly interacts with copper ions, and this ability has been linked to its still unknown function. Attachment of other metal ions (zinc, iron, manganese) have been demonstrated as well, but none of them could outcompete copper. Recent finding, however, indicates that at intermediate concentrations both copper and zinc ions can attach to the PrP at the octarepeat region, which contains high affinity metal binding sites. Based on this evidence, we have performed density functional theory simulations to investigate the combined Cu/Zn attachment. We consider all previously reported binding modes of copper at the octarepeat region and examine a possibility simultaneous Cu/Zn attachment. We find that this can indeed occur for only one of the known binding sites, when copper changes its coordination mode to allow for attachment of zinc ion. The implications of the simultaneous attachment on neural function remain to be explored.   

Direct observation of apolipoprotein B refolding at single molecule level by ultra sensitive fluorescence microscopy and solution transmission electron microscopy

Apolipoprotein (apo) B is the only protein of low-density lipoprotein (LDL). The huge size and extreme hydrophobicity of apoB make examination of its lipidation process an experimental challenge. In this study, we showed that apoB lipidation and its intermediates could be observed at single molecule level by an on-path folding process. When carboxyl-terminal-truncated mutants apoB-29 and apoB-48, representing the amino-terminal 29{\%} and 48{\%}, respectively, of the full-length apoB (apoB-100), were used for comparison, we observed that the refolded apoB-100 resembled both native LDL and VLDL precursors. Thus the process of lipidation recapitulates that of pre-VLDL assembly, \textit{in vitro}. These results suggest that the assembly of mature VLDL requires involvement of factors in addition to apoB-100 and lipids. Using solution transmission electron microscopy (TEM), we were able to detect incorporation of hydrophobic super-paramagnetic iron oxide nanoparticles into apoB-100 particles at the initial, but not final, stage of refolding. The current study thus demonstrates that VLDL assembly can be monitored at single molecule level, too.   

The strength of side chain hydrogen bonds in the plasma membrane

There are no direct quantitative measurements of hydrogen bond strengths in membrane proteins residing in their native cellular environment. To address this knowledge gap, here we use fluorescence resonance energy transfer (FRET) to measure the impact of hydrogen bonds on the stability of a membrane protein dimer in vesicles derived from eukaryotic plasma membranes, and we compare these results to previous measurements of hydrogen bond strengths in model lipid bilayers. We demonstrate that FRET measurements of membrane protein interactions in plasma membrane vesicles have the requisite sensitivity to quantify the strength of hydrogen bonds. We find that the hydrogen bond-mediated stabilization in the plasma membrane is small, only -0.7 kcal/mole. It is the same as in model lipid bilayers, despite the different nature and dielectric properties of the two environments.   

Molecular Dynamics Simulations of Hydrophobic Residues

Molecular recognition and protein-protein interactions are involved in important biological processes. However, despite recent improvements in computational methods for protein design, we still lack a predictive understanding of protein structure and interactions. To begin to address these shortcomings, we performed molecular dynamics simulations of hydrophobic residues modeled as hard spheres with stereo-chemical constraints initially at high temperature, and then quenched to low temperature to obtain local energy minima. We find that there is a range of quench rates over which the probabilities of side-chain dihedral angles for hydrophobic residues match the probabilities obtained for known protein structures. In addition, we predict the side-chain dihedral angle propensities in the core region of the proteins T4, ROP, and several mutants. These studies serve as a first step in developing the ability to quantitatively rank the energies of designed protein constructs. The success of these studies suggests that only hard-sphere dynamics with geometrical constraints are needed for accurate protein structure prediction in hydrophobic cavities and binding interfaces.   

Solvation Free Energy and Classical Density Functional Theory

The cell environment in which proteins fold and function is crowded with biological molecules, at densities of $\sim$300g/L. Treating these molecules explicitly in a MD simulation introduces enormous computational cost, so accurate ways of modelling their contribution to protein behaviour is desirable. I will discuss existing models and propose a new approach, which uses classical density functional theory to calculate the effect of these solutes on protein folding. I will discuss implementing this approach as both an implicit solvent and a post-processing method, and discuss some general conclusions we can derive from it.   

Hydrogen Bonding in the Electronic Excited State

Here, I will give a talk on our recent advances in electronic excited-state hydrogen-bonding dynamics and the significant role of excited-state hydrogen bonding on internal conversion (IC), electronic spectral shifts (ESS), photoinduced electron transfer (PET), fluorescence quenching (FQ), intramolecular charge transfer (ICT), and metal-to-ligand charge transfer (MLCT). The combination of various spectroscopic experiments with theoretical calculations has led to tremendous progress in excited-state hydrogen-bonding research. We first demonstrated that intermolecular hydrogen bond in excited state can be greatly strengthened or weakened for many chromophores. We have also clarified that intermolecular hydrogen-bond strengthening and weakening correspond to red-shifts and blue-shifts, respectively, in the electronic spectra. Moreover, radiationless deactivations (via IC, PET, ICT, MLCT, and so on) can be dramatically influenced by excited-state hydrogen bonding. References: [1] Guang-Jiu Zhao, and Ke-Li Han, \textit{Hydrogen Bonding in the Electronic Excited State, } \textit{Accounts of Chemical Research} \quad 45, 404--413 \quad (\textbf{2012}). http://pubs.acs.org/doi/pdf/10.1021/ar200135h [2] Book: \textit{Hydrogen Bonding and Transfer in the Excited State}, Editors: Ke-Li Han and Guang-Jiu Zhao, ISBN: 978-0-470-66677-7, \textit{John Wiley {\&} Sons Ltd}, Chichester, UK (\textbf{2011}). http://onlinelibrary.wiley.com/book/10.1002/9780470669143   

Intermediate Resolution Models and Protein Folding and Allostery

Intermediate Resolution Models (IRM)s model proteins with nearly all atom precision but consider solvent implicitly, and treat electrostatics as short-range interactions. In this talk, we describe a new IRM. We discuss its differences from other existing IRMs and test it again a set of 13 proteins. The model successfully folds 12 of them into its native state, starting from a random configuration. The stability of the native state versus other states with different topologies (arrangement of the secondary structure) is also discussed. Implications for general protein motion are also presented.   

Using extremely halophilic bacteria to understand the role of surface charge and surface hydration in protein evolution, folding, and function

Halophilic Archaea accumulate molar concentrations of KCl in their cytoplasm as an osmoprotectant, and have evolved highly acidic proteomes that only function at high salinity. We examine osmoprotection in the photosynthetic Proteobacteria Halorhodospira halophila. We find that H. halophila has an acidic proteome and accumulates molar concentrations of KCl when grown in high salt media. Upon growth of H. halophila in low salt media, its cytoplasmic K$+$ content matches that of Escherichia coli, revealing an acidic proteome that can function in the absence of high cytoplasmic salt concentrations. These findings necessitate a reassessment of two central aspects of theories for understanding extreme halophiles. We conclude that proteome acidity is not driven by stabilizing interactions between K$+$ ions and acidic side chains, but by the need for maintaining sufficient solvation and hydration of the protein surface at high salinity through strongly hydrated carboxylates. We propose that obligate protein halophilicity is a non-adaptive property resulting from genetic drift in which constructive neutral evolution progressively incorporates weakly stabilizing K$+$ binding sites on an increasingly acidic protein surface.   

Infrared Structural Biology of Proteins: Development of Vibrational Structural Markers for Probing the Structural Dynamics of COO- of Asp/Glu in Proteins

Asp and Glu often play critical roles in the active sites of proteins. Probing the structural dynamics of functionally important Asp and/or Glu provides crucial information for protein functionality. Time-resolved infrared structural biology offers strong advantages for its high structural sensitivity and broad dynamic range (ps to ks). In order to connect the vibrational frequencies to specific structures of COO- groups, such as the number, type, and geometry of hydrogen bond interactions, we develop two vibrational structural markers (VSM), built on the symmetric and asymmetric COO- stretching frequencies. Extensive quantum physics (density functional theory) based computational studies, combined with 13C isotopic editing of Asp/Glu and experimental FTIR data on Asp/Glu in proteins, are used to establish a unique correlation between the symmetric and asymmetric COO- vibrations with more than 10 types of hydrogen bonding interactions. Development of the COO- VSM markers enhances the power of time-resolved infrared structural biology for the study of functionally important structural dynamics of COO- in proteins, including rhodopsin for biological signaling, bacteriorhodopsin for proton transfer, photosystem II for energy transformation, and HIV protease for enzymatic catalysis.   

Controlling allosteric networks in proteins

We present a novel methodology based on graph theory and discrete molecular dynamics simulations for delineating allosteric pathways in proteins. We use this methodology to uncover the structural mechanisms responsible for coupling of distal sites on proteins and utilize it for allosteric modulation of proteins. We will present examples where inference of allosteric networks and its rewiring allows us to ``rescue'' cystic fibrosis transmembrane conductance regulator (CFTR), a protein associated with fatal genetic disease cystic fibrosis. We also use our methodology to control protein function allosterically. We design a novel protein domain that can be inserted into identified allosteric site of target protein. Using a drug that binds to our domain, we alter the function of the target protein. We successfully tested this methodology \textit{in vitro}, in living cells and in zebrafish. We further demonstrate transferability of our allosteric modulation methodology to other systems and extend it to become ligh-activatable.