Session M45: Focus Session: Physics of the Cytoskeleton I

Filament turnover kinetics determine the mechanical relaxation of entangled F-actin solutions

The actin cytoskeleton of eukaryotic cells displays rich mechanical behaviors, which enable cells to efficiently transmit forces required for shape maintenance and tissue stability while also facilitating large shape changes required for morphogenic processes at longer time scales. The molecular processes that control mechanical relaxations of the actin cytoskeleton are poorly understood. Actin filament assembly kinetics are controlled in vivo by an assortment of regulatory proteins, which lead to a complete dissolution and re-formation of filaments on the timescale of seconds. How such ``turnover'' of filaments influences the mechanical properties of the actin cytoskeleton is less clear. To address this, we developed a system using purified actin regulatory proteins, including the severing protein ADF/cofilin and the formin nucleation/elongation factor mDia1, to tune filament turnover kinetics and measured the frequency-dependent shear modulus of entangled actin solutions via particle-tracking microrheology. We observe a tunable reduction in the terminal relaxation time when filament turnover is enhanced through severing, despite a constant mean filament length.   

Measuring actin dynamics during phagocytosis using photo-switchable fluorescence

Phagocytosis has traditionally been investigated in terms of the relevant biochemical signaling pathways. However, a growing number of studies investigating the physical aspects of phagocytosis have demonstrated that several distinct forces are exerted throughout particle ingestion. We use variations on FRAP (Fluorescence Recovery After Photobleaching) in combination with photo-switchable fluorescent protein to investigate actin dynamics as a phagocyte attempts to engulf its prey. The goal of our actin studies are to determine the recruitment and polymerization rate of actin in the forming phagosome and whether an organized \textit{contractile actin ring} is present and responsible for phagosome closure, as proposed in the literature. These experiments are ongoing and contribute to our long term effort of developing a physics based model of phagocytosis.   

Model of Yeast Actin Cable Distribution and Dynamics

The growth of fission yeast relies on the polymerization of actin filaments at the cell tips. These filaments are nucleated by formin proteins that localize at tip cortical sites. These actin filaments bundle to form actin cables that span the cell and guide the movement of vesicles toward the cell tips. Since fluorescence microscopy shows the structure and dynamics of actin cables, we are able to compare the results of the theoretical models of actin cables to experiment, thus enabling quantitative tests of the mechanisms of actin polymerization in cells. We used computer simulations to study the spatial and dynamical properties of actin cables. We simulated individual actin filaments as three-dimensional semiflexible polymer, composed of beads connected with springs. Formin polymerization was simulated as filament growth out of cortical sites located at cell tips. Actin filament severing by cofilin was simulated as filament turnover. We added attractive interactions between beads to simulate filament bundling by actin cross-linkers such as fimbrin. Comparison of the results of the model to prior experiments suggests that filament severing, nucleation and crosslinking are sufficient to describe the many features of actin cables. We found bundled and unbundled phases as cross-linking strength was varied and propose experiments to test the model predictions.   

Myosin II Dynamics during Embryo Morphogenesis

During embryonic morphogenesis, the myosin II motor protein generates forces that help to shape tissues, organs, and the overall body form. In one dramatic example in the \textit{Drosophila melanogaster} embryo, the epithelial tissue that will give rise to the body of the adult animal elongates more than two-fold along the head-to-tail axis in less than an hour. This elongation is accomplished primarily through directional rearrangements of cells within the plane of the tissue. Just prior to elongation, polarized assemblies of myosin II accumulate perpendicular to the elongation axis. The contractile forces generated by myosin activity orient cell movements along a common axis, promoting local cell rearrangements that contribute to global tissue elongation. The molecular and mechanical mechanisms by which myosin drives this massive change in embryo shape are poorly understood. To investigate these mechanisms, we generated a collection of transgenic flies expressing variants of myosin II with altered motor function and regulation. We found that variants that are predicted to have increased myosin activity cause defects in tissue elongation. Using biophysical approaches, we found that these myosin variants also have decreased turnover dynamics within cells. To explore the mechanisms by which molecular-level myosin dynamics are translated into tissue-level elongation, we are using time-lapse confocal imaging to observe cell movements in embryos with altered myosin activity. We are utilizing computational approaches to quantify the dynamics and directionality of myosin localization and cell rearrangements. These studies will help elucidate how myosin-generated forces control cell movements within tissues. \textit{This work is in collaboration with J. Zallen at the Sloan-Kettering Institute.}   

Model of Capping Protein and Arp2/3 Complex Turnover in the Lamellipodium Based on Single Molecule Statistics

Capping protein (CP) and Arp2/3 protein complex regulate actin polymerization near the leading edge of motile cells. Actin and regulatory proteins assemble near the leading edge of the cell, undergo retrograde flow, and dissociate into the cytoplasm as single subunits (monomers) or as part of multiple actin subunits (oligomers.) To better understand this cycle, we modeled the kinetics of actin CP and Arp2/3 complex near the leading edge using data from prior experiments [Miyoshi et al. JCB, 2006, 175:948]. We used the measured dissociation rates of Arp2/3 complex and CP in a Monte Carlo simulation that includes particles in association with filamentous and diffuse actin in the cytoplasm. A slowly diffusing cytoplasmic pool may account for a big fraction of CP, with diffusion coefficients as slow as 0.5 $\mu m^2/s$ [Smith et al. Biophys. J., 2011,101:1799]. Such slow diffusion coefficients are consistent with prior experiments by Kapustina et al. [Cytoskeleton, 2010, 67:525]. We also show that the single molecule data are consistent with experiments by Lai et al. [EMBO J., 2008, 28:986]. We discuss the implication of disassembly with actin oligomers and suggest experiments to distinguish among mechanisms that influence long range transport.   

Effect of surface topography on actin dynamics and receptor clustering in B cells

B cells are activated upon binding of the B cell receptor (BCR) with antigen on the surface of antigen presenting cells (APC). Activated B cells deform and spread on the surface of APCs which may comprise of complex membrane topologies. In order to model the diverse range of topographies that B cells may encounter, substrates fabricated with vertical ridges separated by gaps ranging from hundreds of nm to microns were coated with activating antigen to enable B cell spreading. Simultaneous imaging of actin and BCR shows that the organization of both depends profoundly on the ridge spacing. On smaller ridge spacing (\textless 2 microns), actin forms long filopodial structures that explore the substrate parallel to ridges while the BCR clusters accumulate linearly along the direction of the ridges with limited ability to escape these channels. Cells on larger ridge spacing (\textgreater 2 microns) exhibit central actin patches and peripheral actin waves and form semi-stable polymerization zones at ridges, while BCR distribution is more homogeneous. Our results indicate that surface topography may be a critical determinant of cytoskeletal dynamics and the spatiotemporal organization of signaling clusters.   

Biomimetic active emulsions capture cell dynamics and direct bio-inspired materials

The main biopolymers which make up the cellular cytoskeleton and provide cells with their shape are well understood, yet, how they organize into structures and set given cellular behavior remains unclear. We have reconstituted minimal networks of actin, a ubiquitous biopolymer, along with an associated motor protein myosin II to create biomimetic networks which replicate cell structure and actively contract when selectively provided with ATP. We emulsify these networks in 10-100 micron drops, provide a system to investigate strain-mediated protein interactions and network behavior in confined cell-similar volumes. These networks allow us to study strain-mediated protein-specific interactions in an actin network at a precision impossible in vivo. Using this system, we have identified strain-dependent behavior in actin cross linking proteins; mechanotransduction of signaling proteins in Filamin A, and unique catch-bond behavior in Alpha-actinin. This understanding of biopolymer self-organization to set cell mechanics, will help clarify how biology both generates and reacts to force; moreover this system provides a highly controlled platform for studying non-equilibrium materials, and creating microscopic building block for a entirely new class of active materials.   

Eukaryotic and Prokaryotic Cytoskeletons: Structure and Mechanics

The eukaryotic cytoskeleton is an assembly of filamentous proteins and a host of associated proteins that collectively serve functional needs ranging from spatial organization and transport to the production and transmission of forces. These systems can exhibit a wide variety of non-equilibrium, self-assembled phases depending on context and function. While much recent progress has been made in understanding the self-organization, rheology and nonlinear mechanical properties of such active systems, in this talk, we will concentrate on some emerging aspects of cytoskeletal physics that are promising. One such aspect is the influence of cytoskeletal network topology and its dynamics on both active and passive intracellular transport. Another aspect we will highlight is the interplay between chirality of filaments, their elasticity and their interactions with the membrane that can lead to novel conformational states with functional implications. Finally we will consider homologs of cytoskeletal proteins in bacteria, which are involved in templating cell growth, segregating genetic material and force production, which we will discuss with particular reference to contractile forces during cell division. These prokaryotic structures function in remarkably similar yet fascinatingly different ways from their eukaryotic counterparts and can enrich our understanding of cytoskeletal functioning as a whole.   

Cell shape can mediate the spatial organization of the bacterial cytoskeleton

The bacterial cytoskeleton guides the synthesis of cell wall and thus regulates cell shape. Since spatial patterning of the bacterial cytoskeleton is critical to the proper control of cell shape, it is important to ask how the cytoskeleton spatially self-organizes in the first place. In this work, we develop a quantitative model to account for the various spatial patterns adopted by bacterial cytoskeletal proteins, especially the orientation and length of cytoskeletal filaments such as FtsZ and MreB in rod-shaped cells. We show that the combined mechanical energy of membrane bending, membrane pinning, and filament bending of a membrane-attached cytoskeletal filament can be sufficient to prescribe orientation, e.g. circumferential for FtsZ or helical for MreB, with the accuracy of orientation increasing with the length of the cytoskeletal filament. Moreover, the mechanical energy can compete with the chemical energy of cytoskeletal polymerization to regulate filament length. Notably, we predict a conformational transition with increasing polymer length from smoothly curved to end-bent polymers. Finally, the mechanical energy also results in a mutual attraction among polymers on the same membrane, which could facilitate tight polymer spacing or bundling. The predictions of the model can be verified through genetic, microscopic, and microfluidic approaches.   

Elastic behavior of vimentin intermediate filament networks

A cell's response to mechanical stress is closely linked to the structure and elasticity of its cytoskeleton, which is comprised primarily of actin, microtubule, and intermediate filament (IF) networks. Vimentin is an IF found in mesenchymal cells that plays a role in anchoring organelles and contributes to overall cellular elasticity. Previous research has shown that vimentin networks behave like softly crosslinked gels in the presence of divalent cations. The linear elastic modulus, a measure of stiffness and resistance to elastic deformation, of the network is related to the degree of crosslinking, which is itself controlled by the cation concentration. Increasing the concentration of the divalent cations further results in the formation of bundles within the network, but this bundling behavior is not well understood. Here we investigate the response of in vitro reconstituted vimentin networks to applied shear in the presence of various divalent species to better understand their individual contributions to the network's elastic behavior.   

Biopolymer Networks: Simulations of Rigid Rods Connected by Wormlike Chains

The cytoskeleton of cells is a composite network of filaments ranging from stiff rod-like microtubules to semiflexible actin filaments that together play a crucial role in cell structure and mechanics. The collective dynamics of these cytoskeletal filaments with different mechanical properties are yet to be understood completely. To model such a strongly heterogeneous composite, we simulate networks of \textit{rigid} rods connected by \textit{flexible} linkers (wormlike chains). We extract elastic moduli by quasistatic deformations at varying filament densities and analyze the crossover between cross-link dominated and rod dominated regimes. In particular, we are interested in the asymptotic stress dependence of the \textit{differential modulus}. The simulations are accompanied by rheological experiments on networks of \textit{microtubules} (MTs) cross-linked by double-stranded \textit{DNA} of variable length (cf. talk Meenakshi Prabhune).