Session Y47: Soft Matter Physics of Biological Systems

Antiviral activity of squalamine: Role of electrostatic membrane binding

Recent work\footnote{M. Zasloff \emph{et al.}, Proc. Nat. Acad. Sci. (USA) \textbf{108}, 15978 (2011).} has demonstrated that squalamine, a molecule found in the liver of sharks, exhibits broad-spectrum antiviral properties. It has been proposed that this activity results from the charge-density matching of squalamine and phospholipid membranes, causing squalamine to bind to membranes and displace proteins such as Rac1 that are crucial for the viral replication cycle. Here we investigate this hypothesis by numerical simulation of a coarse-grained model for the competition between Rac1 and squalamine in binding affinity to a flat lipid bilayer. We perform free-energy calculations to test the ability of squalamine to condense stacked bilayer systems and thereby displace bulkier Rac1 molecules. We directly compare our findings to small-angle x-ray scattering results for the same setup.   

Simulating ligand receptor binding at a membrane interface with graphics processing accelerated coarse-grained molecular dynamics

Motivated by a deeper understanding of the immunological synapse, we develop a molecular-based model to understand receptor-polymer/ligand binding at a membrane interface. ~ We examine the case of weak ligand binding in the limit of confined polymer chains as a function of chain length, binding constant, and system size. ~We utilize a coarse-grained (CG) model of poly(ethylene oxide) and dimyristoylphosphatidylcholine (DMPC) previously developed by the Klein group and mimic weak binding with a sticky potential. ~This work employs graphics processing units (GPU) to accelerate the CG-MD simulations, where each simulation is run with multiple random-walker replicas to enhance sampling and facilitate statistical convergence of physical observables. ~Our results demonstrate that such an aggressive combination of GPU acceleration with CG modeling can yield accurate and precise data on polymer-DMPC binding, and, more importantly, hints at the mechanism behind empirical data of polymer binding to a T-cell receptor protein.   

Stabilization of concentration fluctuations in mixed membranes by hybrid lipids

Finite-size domains have been observed at the surface of cells. These lipids ``rafts'' are stable nanodomains enriched in saturated lipids and cholesterol. While line tension favors macrodomains, one explanation for raft stabilization suggests that the membrane composition is tuned close to a spinodal temperature. From this point of view, rafts are long-lived concentration fluctuations in the mixed phase. We propose a ternary mixture model for the cell membrane that includes hybrid lipids which have one saturated and one unsaturated hydrocarbon chain. Finite amount of hybrid lipids reduces the packing incompatibility at the saturated/unsaturated lipid interface and stabilizes the concentration fluctuations. Hybrid-Hybrid interactions are included in the model and further increase the life-time of the rafts and decrease their length-scales. Moreover, the hybrid has extra orientational degrees of freedom that may lead to modulated phases.   

Cooperativity in cholesterol-induced demixing of saturated and unsaturated lipids

The ternary DPPC/DOPC/cholesterol system exhibits a phase separation between liquid-ordered and liquid-disordered domains that may capture some features of demixing believed to occur in biomembranes. Using semi-grand canonical ensemble Monte Carlo approaches on a model with atomistic detail, we have investigated the degree of non-ideality of mixing of DPPC and DOPC as influenced by cholesterol. While the signature of phase separation is observed, more surprising is that in the region of composition space characterized by low DPPC and low cholesterol content, the mixing is approximately ideal. Through simple models we have shown that this behavior is inconsistent with a simple nearest-neighbor description of the interactions between these membrane components.   

Steric Repulsion and Compressibility of Protein Resistant Brushes

End-grafted polymer brushes, such as polyethylene glycol, have become commonplace as biocompatibilizers for medical devices and diagnostic surfaces. Key to their non-fouling character is the brush's steric repulsion towards biomolecules and cells. By functionalizing the substrate with small (order 10 nm) bioadhesive features around which the brush is placed, we gain insight into the repulsion between biomolecules and cells with the brush itself. While approximately exponential compression profiles are to be expected, some features of protein interactions with these brushes are unexpected, especially for small proteins whose dimensions are within a factor of 2 or 3 of the brush persistence length. The scaling of the compressive force, for example inferred from series of studies that vary the amounts and spacings of the adhesive elements, is weakly dependent on protein size, while one might expect a proportionality between this force and the effective protein footprint. These results are consistent with entropically inexpensive chain reconfigurations around the smaller proteins and the penetration of these proteins at least partially into the brush.   

Topological defects induced by the retina's curvature improve vision

The theory of disclinations and dislocations on curved surfaces predicts the length and density of grain boundary scars on a sphere. These predictions were successfully tested with colloids on droplets for systems satisfying $5\le R/a\le20$ ($R$- sphere radius, $a$- lattice constant). The foveal cone mosaic is another realization of this problem, for which $R/a\sim10^{4}$. New theories are needed to extend current predictions for scar length and density to this regime. We present a method of introducing the effect of irregularities that changes the prediction in the relevant regime. We do so by deriving a noise induced disclination density which truncates the scars: the cone density is mapped to an effective displacement $h_{eff}$ from the sphere; then the deviation from the constant curvature is computed to first order in $h_{eff}$; and finally the effective curvature is compared to a threshold above which noise induced disclinations appear. We compare stimuli projected on mosaics and on jittered lattices and show that the curvature induced correlations in the mosaics reduce aliasing by a factor of up to 50. This reduction increases with spatial frequencies, meaning that anti-aliasing is maximal in the visual acuity limit.   

Exploring the low temperature thermodynamics of lattice proteins and polymers with chain lengths $> 1000$

Coarse-grained (lattice-) models have a long tradition in aiding to decipher the physical or biological complexity of polymers and proteins. Despite their simplicity however, numerical simulations of such models are often computationally very demanding and the quest for efficient algorithms is as old as the models themselves. I present a computational method based on Wang-Landau sampling in combination with suitable trial move sets which is particularly effective to study models such as the hydrophobic-polar (HP) lattice model of protein folding or its counterpart in polymer physics, the interactive self-avoiding walk (ISAW) at low temperatures. The approach provides a versatile and powerful mean for both the ground state search and the determination of the entire energy density of states (DOS) yielding reliable estimates of thermodynamic quantities for chain lengths $> 4000$ (ISAW) even in the very dense collapsed phase. The appearance of multiple low temperature pseudo-transitions for ISAWs will be elucidated. Further methodological improvements will be discussed.   

Long-range mechanical force in epithelial tubule self assembly

In vivo, epithelial cells can respond to extracellular matrix (ECM) molecules, type I collagen (COL), and switch their morphology from a lobular lumen (100-200 micron) to a tubular lumen (1mm-1cm). However, the mechanism is unclear. Using a temporal control of cell-ECM interaction, we find that epithelial cells, in response to a fine-tuned percentage of COL in ECM, develop various linear patterns. Remarkably, these patterns allow cells to self-assemble into a tubule of length $\sim $ 1cm and diameter $\sim $ 400 micron in the liquid phase. In contrast with conventional thought, the linear patterns arise through bi-directional transmission of traction force, but not through diffusible biochemical factors secreted by cells. In turn, the transmission of force evokes a long-range ($\sim $ 600 micron) intercellular mechanical interaction. A feedback effect is encountered when the mechanical interaction modifies cell positioning and COL alignment. Micro-patterning experiments further reveal that such a feedback is a novel cell-number-dependent, rich-get-richer process, which allows cells to integrate mechanical interactions into long-range ($>$ 1mm) linear coordination.   

Bottom-up study of flaw tolerance properties of protein networks

We study the material properties of an intermediate filament proten network by computational modeling using a bottom-up approach. We start with an atomic model of each filament's and obtain the mechanical behavior of them. We then use these parameters in setting up a mesoscale model of the network material at scales of micrometers. Using this multi-scale method, we report a detailed analysis of the associated deformation and failure mechanisms of this hierarchical material. Our modeling reveals that a structure transition that occurs at the proteins' secondary structure level is crucial for the networks' flaw tolerance property, which implies that the material retains its mechanical function despite the existence of large defects. We also examine the effect of crosslink strength on the failure properties. We discover that relatively weaker crosslinks lead to a more flaw tolerant network that is 23{\%} stronger. This unexpected behavior is caused by that the crosslink strength functions as a switch to alter the failure mechanism. Weak crosslinks are able to efficiently diffuse the stress around the crack tip, making the crack more difficult to propagate. We compare our results to that of elastic and softening materials and find that the effect of crosslink strength is much smaller in those systems. These findings imply that the mechanical properties of both the filaments and interfaces among filaments are critical for bioinspired material designs, challenging the conventional paradigm in engineering design.   

From Nano to Micro: Importance of Structure and Architecture in Spider Silk Adhesives

Spiders have developed outstanding adhesives over millions of years of evolution for prey capture and locomotion. We show that the structure and architecture of these adhesives play an important role in the adhesion. The adhesive produced by orb-weaving spiders to capture prey (viscid glue) is laid on a pair of silk fibers as micron-size glue drops composed of salts and glycoproteins. By stretching single drops, we show that viscid glue behaves like a viscoelastic solid and that elasticity is critical in enhancing adhesion caused by specific adhesive ligands by over 100 times. Comparing viscid glue with gumfoot glue, the glue produced by cob-weavers, the evolutionary descendants of orb-weavers, showed that, in spite of being produced in homologous aggregate glands, gumfoot glue behaves like a viscoelastic liquid. Moreover, gumfoot glue is humidity-resistant and viscid glue is humidity-sensitive. We use a synthetic strategy to spin beads-on-a-string (BOAS) architecture to mimic the adhesive properties of spider silk. Using these mimic threads, we show that the BOAS structure adheres more than a cylindrical structure during contact (collision of prey) and during separation (escape attempt of prey). These results inspire design of novel tunable adhesives.   

Light scattering studies of human crystallin proteins and loss of transparency in cataracts

The human lens derives its transparency and refractive index from the interactions between crystallin proteins ($\alpha$-, $\beta$-, $\gamma$-crystallin). When the ordering of these crystallins is perturbed, insoluble macromolecular aggregates of crystalline proteins can occur resulting in cataracts. Using dynamic light scattering (DLS) and fast protein liquid chromatography (FPLC), we have conducted a detailed study of the formation of these aggregates. Our DLS results on $\gamma$-crystallin solutions exhibit the occurrence of slow and fast modes demonstrating the spontaneous formation of aggregates (hydrodynamic radius, Rh $\sim$ 200 nm) in equilibrium with monomeric proteins (Rh $\sim$ 3 nm). On the other hand, DLS results on $\alpha$-crystallin solutions clearly demonstrate that $\alpha$-crystallin molecules exist as a single population (Rh $\sim$ 18 nm). Our results on mixtures of $\alpha$- and $\gamma$-crystallin solutions show that the $\alpha$-crystallin tends to demolish the clumps of $\gamma$-crystallin. Our exploration of environmental effecs (temperature, pH, salt concentration) has revealed the macromolecular mechanism of dissolution of crystallin aggregates, providing a strategy for cataract prevention and insight into protein-protein interactions.   

Quantifying the abnormal hemodynamics of sickle cell anemia

Sickle red blood cells (SS-RBC) exhibit heterogeneous morphologies and abnormal hemodynamics in deoxygenated states. A multi-scale model for SS-RBC is developed based on the Dissipative Particle Dynamics (DPD) method. Different cell morphologies (sickle, granular, elongated shapes) typically observed in deoxygenated states are constructed and quantified by the Asphericity and Elliptical shape factors. The hemodynamics of SS-RBC suspensions is studied in both shear and pipe flow systems. The flow resistance obtained from both systems exhibits a larger value than the healthy blood flow due to the abnormal cell properties. Moreover, SS-RBCs exhibit abnormal adhesive interactions with both the vessel endothelium cells and the leukocytes. The effect of the abnormal adhesive interactions on the hemodynamics of sickle blood is investigated using the current model. It is found that both the SS-RBC - endothelium and the SS-RBC - leukocytes interactions, can potentially trigger the vicious ``sickling and entrapment'' cycles, resulting in vaso-occlusion phenomena widely observed in micro-circulation experiments.   

The energetics of tightly bent DNA: a composite elastica model including local melting

Melting transitions are well-known to be affected by the application of mechanical stress. Motivated by the experiments of Zocchi and collaborators (Qu and Zocchi 2011, EPL \textbf{94} 18003), we explore the effect of the application of mechanical stress on DNA melting in a particular composite of a stiff double stranded piece of DNA (dsDNA), shorter than its own persistence length, whose ends are linked by a flexible single stranded piece of DNA (ssDNA). The flexible ssDNA acts as a Gaussian polymer coil bending the stiff dsDNA through an elastic force that is controllable by the length of the ssDNA chain. In this talk we present theoretical predictions for two experimentally accessible features: the degree of local dsDNA melting and the local elastic energy of the dsDNA/ssDNA construct both as a function of the length of the attached ssDNA. We also address the effect of introducing a nick (broken covalent bond) in the dsDNA backbone on these results and discuss the implications of such data on the relative importance of backbone elasticity versus base stacking and base pairing interactions in determining the elasticity of dsDNA. This work also addresses open questions in the nonlinear elasticity of DNA in tightly bent curves.   

Dispersion-relation phase spectroscopy of neuron transport

Molecular motors move materials along prescribed biopolymer tracks. This sort of active transport is required to rapidly move products over large distances within the cell, where passive diffusion is too slow. We examine intracellular traffic patterns using a new application of spatial light interference microscopy (SLIM) and measure the dispersion relation, i.e. decay rate vs. spatial mode, associated with mass transport in live cells. This approach applies equally well to both discrete and continuous mass distributions without the need for particle tracking. From the quadratic experimental curve specific to diffusion, we extracted the diffusion coefficient as the only fitting parameter. The linear portion of the dispersion relation reveals the deterministic component of the intracellular transport. Our data show a universal behavior where the intracellular transport is diffusive at small scales and deterministic at large scales. We further applied this method to studying transport in neurons and are able to use SLIM to map the changes in index of refraction across the neuron and its extended processes. We found that in dendrites and axons, the transport is mostly active, i.e., diffusion is subdominant.