physics of cellular movements - supplementary info

Upload: alienced

Post on 05-Apr-2018

215 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    1/22

    Physics of CellularMovements

    Erich Sackmann,1,2,

    Felix Keber,2 andDoris Heinrich2

    1Physics Department, Institute for Biophysics E22,Technische Universita t Mu nchen, D-85748 Garching,

    Germany; email: [email protected] t fu r Physik and Center for NanoScience (CeNS),

    Ludwig-Maximilians-Universita t Mu nchen, D-80539 Munich,Germany; email: [email protected],

    [email protected]

    Annu. Rev. Condens. Matter Phys. 2010. 1:25776

    First published online as a Review in Advance onMay 21, 2010

    The Annual Review of Condensed Matter Physics is

    online at conmatphys.annualreviews.org

    This articles doi:10.1146/annurev-conmatphys-070909-104105

    Copyright 2010 by Annual Reviews.All rights reserved

    1947-5454/10/0810-0257$20.00

    Corresponding author.

    Key Words

    cell adhesion, microtubule-actin-crosstalk, actin polymerizationwaves, cell locomotion, intracellular transport

    Abstract

    The survival of cells depends on perpetual active motions, includ-

    ing (a) bending excitations of the soft cell envelopes, (b) the bidi-

    rectional transport of materials and organelles between the cell

    center and the periphery, and (c) the ongoing restructuring of

    the intracellular macromolecular scaffolds mediating global cell

    changes associated with cell adhesion locomotion and phagocyto-

    sis. Central questions addressed are the following: How can this

    bustling motion of extremely complex soft structures be character-

    ized and measured? What are the major driving forces? Further

    topics include (a) the active dynamic control of global shape

    changes by the interactive coupling of the aster-like soft scaffold of

    microtubules and the network of actin filaments associated withthe cell envelope (the actin cortex) and (b) the generation of pro-

    pulsion forces by solitary actin gelation waves propagating within

    the actin cortex.

    Annu.R

    ev.Condens.MatterPhys.2

    010.1:257-276.Downloaded

    fromwww.annualreviews.org

    byIndianInstituteofScience-Bangaloreon05/15/12.F

    orpersonaluseonly.

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    2/22

    INTRODUCTION

    It is fascinating to look at our blood cells through a phase contrast microscope and see with

    our own eyes that cells are really living. In the large white blood cells, we see intercellular

    organelles moving around in the cytoplasmic space in a seemingly irregular fashion, remi-

    niscent of random walks. The nucleus-free erythrocytes (simple cell envelopes filled with

    hemoglobin solution) show pronounced thickness fluctuation as if driven by magic forces.

    Some white blood cells crawl over the surface as shown in the famous movie Crawling

    neutrophil chasing bacteria that can be downloaded from YouTube.

    Until a decade ago, physicists were mainly studying the static physical properties of the

    erythrocyte. Much of our present knowledge on the physical basis of the self-organization

    of cell membranes is due to comparative studies of erythrocyte envelopes and model mem-

    branes. These studies led to the curvature elasticity concept of cell shape changes (reviewedin Reference 1), and the thermomechanical control of microstructures in cell envelopes and

    of cell adhesion (reviewed in References 1 and 2).

    In the past decade, nucleated cells became a topic of interest for physicists. Most studies

    concentrated on the mechanics and microviscoelasticity of cells. Comparative studies of

    cells and of in vitro models of intracellular macromolecular scaffolds, made up of semi-

    flexible actin filaments and microtubules (MTs), provided valuable insight into correla-

    tions between the molecular architecture of cells and their mechanical properties (3, 4).

    Experiments provided evidence that the rheological properties of cells show typical univer-sal features of soft glassy materials, which is reflected by similar behavior of in vitro

    models of actin networks (reviewed in References 3, 5). The glass-like behavior can explain

    the astonishing robustness of cells against removal of proteins by mutations. The physics of

    cells was further stimulated by recent studies showing that cells can sense forces (68), and

    thereby adjust their mechanical impedance to that of the environment by reorganization of

    the actin cortex. Moreover, the mechanical properties of the tissue or bones can stimulate

    the differentiation of stem cells into specific cells, such as muscle cells. Mechanical forces

    also control the development of nerve connections by synapses during the development of

    embryos (9).

    During the past few years, physicists became interested in the tremendously complex

    problem of cell dynamics and its key role in numerous vital processes, such as cell migra-

    tion, immunological processes (10, 11), or the engulfment of bacteria (discussed below).

    The elucidation of the hierarchy of motions in complex soft materials, such as cells, is a

    great future challenge for physicists working in the field of soft materials.

    In this article, we describe the intrinsic motions of cells driven by thermal and activerandom and directed forces mediated by molecular motors, solitary actin gelation (or

    polymerization) waves (SAGWs), and the dynamic coupling of the soft star-like scaffold

    of MTs to the actin cortex of the composite cell envelope.

    We focus on three topics: (a) the optimized exploration of the cytoplasmatic space by

    internalized objects (endosomes) through alternating random and directed local forces,

    (b) the dynamic control of the global mechanical stability of cells via actin-microtubule

    crosstalk, and (c) the quasi-random crawling of cells over surfaces by nucleation and

    growth of actin networks.Special emphasis is placed on the control of the structure and the physical properties

    and the function of the composite cell envelopes by numerous actin-binding proteins

    acting as actuators (effectors), the activity of which is controlled by biochemical switches

    Cell shape changes:changes of the globalshape of cells, such asdiscocyte-to-sphere

    transitions of redblood cells(erythrocytes)

    MT: microtubule

    Solitary actin gelation

    (or polymerization)wave (SAGW):propagating

    accumulation of actingel mediated bypolymerisation atthe front anddecomposition atthe end

    Solitary wave:a self-reinforcingsingle wave packet

    propagating withconstant velocitywhile maintainingits shape

    Composite cellenvelope: stratifiedouter shell ofeukaryotic cells

    composed of thelipid-protein bilayer(PM), the glycocalix,and the actin cortex

    Actin-microtubulecrosstalk: interactionof microtubules withthe actin cortexresulting in the

    interactiverestructuring ofthe actin cortexA

    nnu.R

    ev.Condens.MatterPhys.2

    010.1:257-276.Downloaded

    fromwww.annualreviews.org

    byIndianInstituteofScience-Bangaloreon05/15/12.F

    orpersonaluseonly.

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    3/22

    actuated by external signals, local forces, and cell adhesion. The physical basis of the

    membrane-bending undulations and their biological role is summarized in Supplemental

    Appendix G (follow the Supplemental Material link from the Annual Reviews home page

    at http://www.annualreviews.org).

    We focus on studies of the amoeboid cells of the slime mold Dictyostelium discoideum(abbreviated as Dictyostelium cells). It is a prototype of a highly mobile cell that can be

    observed under the microscope for days. It is extremely robust against the knockout of

    proteins by mutation. Comparative studies showed that Dictyostelium cells and animal

    cells share many common features and differ mostly in the time scale of dynamic processes.

    The main difference between these cells is that the animal cells possess a third type of

    macromolecular scaffold that is composed of semiflexible macromolecules called interme-

    diate filaments.

    MICROANATOMY AND DYNAMICS OFTHE COMPOSITE CELL ENVELOPE

    Because many motional processes and mechanical properties are controlled by cell enve-

    lopes, we first summarize the microanatomy of this stratified soft shell (see Figure 1). It

    consists of two subshells: the plasma membrane (PM) and the associated macromolecular

    network. The PM is a multicomponent lipid-protein alloy containing a stunning multitudeof functional proteins, including ion channels, enzymes, hormone amplifiers, and several

    families of proteins mediating cell-adhesion, called cell adhesion molecules (CAMs) (1, 2).

    The simplest composite cell envelope is that of the erythrocyte. Its intracellular scaffold

    consists of a two-dimensional macromolecular network, and the microanatomy and

    dynamics of this shell has been extensively reviewed (2).

    In nucleated cells (such as the Dictyostelium cells or white blood cells), the mem-

    brane-associated network is formed by the actin cortex. It consists of a 0.20.5-mm-thick

    shell made up of slightly cross-linked actin filaments. These semiflexible macromoleculesare locally coupled to the intracellular domains of integral membrane proteins by spe-

    cific actin-membrane linkers. A frequent linker is talin, which anchors actin to CAMs

    (as shown in Figure 1c). The attachment of phosphate groups to tyrosine side chains of

    the protein (called tyrosine phosphorylation) triggers the binding of talin (12). This is

    mediated by specific enzymes (called tyrosine kinases), which can be switched on by

    biochemical signals, external forces, or cell adhesion. In this way, the structure and the

    elasticity of the cell envelope can be modified within seconds by external forces or

    chemical signals, such as hormones or chemoattractants (see Figure 1b and Supplemen-

    tal Appendix B).

    As shown in Figure 1d, the actin cortex can undergo a series of phase transitions, from

    homogeneous networks to heterogels composed of tightly coupled bundles or branched

    networks (12, 13). Transitions between different states of the actin cortex can be controlled

    by activation of specific actin cross-linkers and the density of the actin network, which is

    essential for the dynamic reorganization of the actin cortex (14, 15).

    The lipid-protein bilayer is extremely soft with respect to shearing and bending, but it isnearly incompressible with respect to lateral extensions. The softness is most impressively

    demonstrated by the pronounced bending excitations of many cell types, resulting in a

    dynamic roughness of cell surfaces of 10 nm (1618). The entropic disjoining pressure

    mediated by the dynamic surface roughness plays a key role for the dynamics of cell

    PM: plasmamembrane; lipidbilayer of the cellenvelope

    Cell adhesion mole-cule (CAM): cellsurface receptorsmediating celladhesion

    Supplemental Material

    Annu.R

    ev.Condens.MatterPhys.2

    010.1:257-276.Downloaded

    fromwww.annualreviews.o

    rg

    byIndianInstituteofScience-Bangaloreon05/15/12.F

    orpersonaluseonly.

    http://arjournals.annualreviews.org/article/suppl/10.1146/annurev-conmatphys-070909-104105?file=co-1-sackman.pdfhttp://arjournals.annualreviews.org/article/suppl/10.1146/annurev-conmatphys-070909-104105?file=co-1-sackman.pdf
  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    4/22

    Receptorb

    Activation(kinase)

    CAM activated CAM inactive

    Inhibition

    (phosphatase)

    PLinker

    P

    dc

    cc

    I

    II

    III

    IV

    CAM cluster

    P-3,4,5-P3-coupled protein

    Talin

    Actin

    Plasma

    membrane:

    4nm

    +

    Arp2/3

    Arp2/3

    ActinActin

    aPS CAM

    CAM

    Glycocalix:40nm

    Actin

    cortex

    :400nm

    GL

    Cross-linker

    Cross-linker+++

    +

    =

    Actin flament

    Filamin

    Myosin

    Figure 1

    (a) Diagram of the three-layered composite shell of nucleated cells, composed of the central lipid-protein bilayer (called plasmamembrane, PM), the glycocalix (G) facing the extracellular space, and the actin cortex. The glycocalix is formed by the head groupsof cell adhesion molecules (CAMs), receptors binding polysaccharides (PSs) of the extracellular matrix, and glycolipids (GLs). Notethat the actin filaments (diameters $8 nm) are slightly thicker than the lipid-protein bilayer ($5 nm) (not shown to scale).(b) Illustration of the activation of the actin-membrane linker by phosphorylation. The activation and deactivation are mediated

    by the antagonistic pair of enzymes: the phosphate-coupler (kinase) and the phosphate-decoupler (phosphatase) (see SupplementalMaterial A, follow the Supplemental Material link from the Annual Reviews home page at http://www.annualreviews.org).(c) Simplified model of coupling between subshells of nucleated cells. Actin filaments are coupled to intracellular domains of celladhesion molecules (such as integrins) via talin after activation of this actin-membrane linker by phosphorylation. Note that CAMsthat are not clustered and linked to talin (such as that on the right side) are only very weakly binding to tissues. The network canextend in the third dimension by coupling of other actin filaments via cross-linkers. The right side of the image shows the coupling ofextrinsic proteins to the membrane through the lipid anchor PI-4,5-P2, assisted by electrostatic binding of positively charged aminoacids to acidic lipids. This coupling plays a key role for the dynamic restructuring of the actin cortex (see Figure 5). (d) Manifold ofgel states in actin networks. With increasing density of cross-linkers (increasing ratio e Dcc/x of the distance between cross-linkersDcc and the mesh sizex), the homogeneous network in the center can undergo transitions into a manifold of heterogeneous gel states.These include condensed networks interconnected by thin bundles (I); actin bundles generated by myosin microbundles, which formpercolated networks (II); and branched networks formed by the cross-linker Arp2/3 (III). The heterogels may coexist with theslightly linked homogeneous networks. Such coexisting networks are formed for instance by the cross-linker filamin (IV). The typeof network formed depends on the structure and the binding kinetics of the cross-linkers. The branched network is only formed bythe cross-linker Arp2/3. Bundles may be generated by molecular motors of the myosin family, generating micromotors.

    Annu.R

    ev.Condens.MatterPhys.2

    010.1:257-276.Downloaded

    fromwww.annualreviews.o

    rg

    byIndianInstituteofScien

    ce-Bangaloreon05/15/12.F

    orpersonaluseonly.

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    5/22

    adhesion, as well as the adhesion-induced formation of microdomains (19), which is

    discussed in more detail in Supplemental Appendix G (also see Reference 1).

    THE GLOBAL MECHANICAL DESIGN OF CELLSThe global shape and mechanical stability of the simple shell of erythrocytes and of the

    envelopes of intracellular organelles are determined by the minimum elastic energy of the

    shells. In many cases, the shape is determined by the minimum bending energy. It can

    therefore be controlled by varying the area-to-volume ratio of the shell or by introducing

    spontaneous curvatures (2, 20). More complex and asymmetric shapes of the soft shells

    can be generated by lateral phase separation within the plasma membrane or by introduc-

    ing shear elasticity. The shape is then determined by the ratio of the shear elastic modulusto the bending modulus of the shell (21, 22).

    The situation is much more complex for nucleated cells, as shown in Figure 2 for

    Dictyostelium cells. The mechanical stability of the cytoplasmic space is determined by

    Fex(t)

    FR2

    F R2

    F R1

    Talin

    Bead

    CAM

    Actin-cortex

    MT-actincoupler

    +

    +

    +

    C

    MT

    End

    PM

    Figure 2

    Projected image of an adhering cell showing its mechanical stabilization by the fraction of tensedmicrotubules (MTs) that are fixed to the centrosome (C) with their minus end, and to the viscoelasticactin cortex with their plus end. Note that some MTs may also form thin bundles. The large arrowpointing to the left indicates a strong external force Fex(t) acting on a bead bound to an MT, which isbalanced by the tensil forces (FRX) evoked in the MT. The forces FRX are balanced by shear stresses (s)in the actin cortex, which are coupled to the plasma membrane (PM) by talin (as shown in the inset andFigure 1a). Inset, upper left: Model of coupling of the MT plus-end to the intracellular side of the actincortex, which can be mediated by various coupling proteins, including dynactin, a 37-nm-long rodcarrying Dynein, a minus-end-directed motor (see Supplemental Material B). The experiments leadingto this model of soft cells are summarized in Supplemental Material G.

    Supplemental Material

    Annu.R

    ev.Condens.MatterPhys.2

    010.1:257-276.Downloaded

    fromwww.annualreviews.o

    rg

    byIndianInstituteofScien

    ce-Bangaloreon05/15/12.F

    orpersonaluseonly.

    http://arjournals.annualreviews.org/article/suppl/10.1146/annurev-conmatphys-070909-104105?file=co-1-sackman.pdfhttp://arjournals.annualreviews.org/article/suppl/10.1146/annurev-conmatphys-070909-104105?file=co-1-sackman.pdf
  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    6/22

    the aster-shaped scaffold of MTs. These 25-nm-thick nanotubes are fixed with their minus

    ends to the centrosome, an organelle located near the nucleus that acts as global cell

    organizer. Some MTs are anchored with their plus end at the actin cortex by specific

    linkers, thus mediating a mechanical link between the two networks (23, 24).

    The buckling forces of the MTs (with a bending stiffness of B % 5 1028 Nm) are inthe piconewton range. Therefore, these filaments can only transmit tensions, and therefore

    the cytoplasmic space is viscoplastic. Strong external forces have to be balanced by inter-

    play of the traction forces in the MT network and the shear stress in the viscoelastic actin

    cortex exhibiting shear moduli of approximately 103 Pa (4).

    It is important to emphasize that only a fraction ($10%) of the MTs is fixed to theactin cortex, whereas the rest is mobile and subjected to continuous shrinking and

    growth, resulting in the turnover of all MTs within one hour (23, 24). In fact, free

    (nonadhering) cells exhibit mostly MTs with free dangling plus ends, whereas the frac-

    tion of fixed filaments increases when the cells adhere to surfaces or move on surfaces.

    The stable MT fraction may also help to maintain the global mechanical tension in long

    axons, which has been shown to be required for the accumulation of presynaptic vesicles

    at the axon endings (9).

    The above model on the smart mechanical design of cells is suggested by two groups of

    experiments: first, by analysis of the quasi-random motions of colloidal probes within the

    cell (25, 26) and, second, by magnetic tweezer microrheology studies of the cell cytoplasm(27). In the latter type of experiment, superparamagnetic microparticles (DynabeadsW) are

    transferred into the cells by phagocytosis. The engulfed beads are wrapped by the PM,

    which exposes its inner leaflet. The beads are treated by the cell as intracellular compart-

    ments and are shuffled around within the cytoplasmic space. Therefore, engulfed colloidal

    beads are ideal probes for systematic studies of the intracellular dynamics or for measuring

    the viscoelastic impedances or active forces.

    OPTIMIZATION OF EXPLORATION OF THE INTRACELLULAR SPACEBY INTERMITTENT RANDOM AND DIRECTED MOTION OFORGANELLES AND ENDOSOMES

    Colloidal probes engulfed by cells exhibit four types of motions in the cytoplasmic space:

    (a) random walks, (b) unidirectional local motions (with velocities $0.5 to 1.5 mm s1)along the MT that are mediated by MT-associated motors, (c) quasi-random sweeping

    (or flagella-like) motions of the MTs (with v % 0.5 mm s1

    ) attributed to the active movementof the plus end of the MTs parallel to the actin cortex, and (d) occasional very fast deflections

    over micrometer distances with ultrafast velocities of up to $10 mm s1.Because the movements of the endosomes consist of random and directed motions, it is

    useful to characterize the motions in terms of local velocity distributions [P(v)] and the

    conventional mean square displacement (MSD) as a function of time. Systematic measure-

    ments of the local velocities showed that they can be best characterized in terms of very

    broad log normal distributions of the random variable (v):

    Pv 1vffiffiffiffiffiffi

    2pp

    s2exp log v

    2

    2s2

    n o, 1:

    comprising velocities between 0.005 (the lower limit of measurement) and 2 mm s1, with a

    maximum at v % 0.3 mm s1. The low-velocity regime of the distribution is attributed to

    Annu.R

    ev.Condens.MatterPhys.2

    010.1:257-276.Downloaded

    fromwww.annualreviews.o

    rg

    byIndianInstituteofScien

    ce-Bangaloreon05/15/12.F

    orpersonaluseonly.

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    7/22

    the effective velocity of the random walk. The center of P(v) is dominated by the MT

    angular motion (with typical velocities of 0.5 mm s1) and the regime from the center of

    P(v) to velocities of 2 mm s1 by the active transport of the beads along the MT.

    Superimposed on these motions are long velocity tails ranging up to 10 mm s1. These

    ultrafast motions are rare events. They are attributed to the sudden release of strain energystored in microtubules, which are prestressed by the actin cortex. These abrupt motions are

    reminiscent of earthquakes driven by mechanical instabilities of Earths crust.

    The velocity distributions provide valuable insights into subtle changes in the intracel-

    lular transport caused by modifications of changes to the cytoskeleton composition due to

    mutations, hormones, or drugs. Examples are the decomposition of the MT network by

    benomyl (a fungi-killing poison) and of the actin cortex by latrunculin A (a toxin that

    sequesters actin monomers and inhibits its polymerization). In both cases, the fast motions

    with v > 1 mm s1 are abolished. Removal of myosin II motors results in a decrease of the

    maximum of P(v) corresponding to an apparent increase in the viscosity of the cytoplasm.

    However, the ultrafast motions are not affected (27), and are thus not driven by molecular

    motors. Both experiments are consistent with the assumption that the ultrafast motions are

    caused by the release of prestress due to the coupling of the microtubules to the actin

    cortex.

    The conventional way of analyzing the intracellular motions of the pseudo endosomes is

    to measure the MSD of the colloidal probes (25, 2830):

    DRt2 DRt t Rt2E: 2:

    Microrheological experiments show that the MSD in the intracellular space of cells follows

    a power law of the form

    DRt2 const ta: 3:Here, the exponent a is larger than 1 in all cells studied (Dictyostelium, fibroblasts) anddepends on the bead radius (R) and the measuring time. It varies from a % 1.3 for R %0.5 mm to a % 1.5 for R % 2 mm in fibroblasts (29) and approaches the value for randomwalks ofa % 1 with decreasing size of the particles.

    The subdiffusive behavior is a consequence of the general Einstein law for the MSD in

    complex fluids (25, 31),

    Dxt

    2E 2Dt 2kBT

    zt t 4:

    with a time-dependent friction coefficient z / tb. This deviation from the random walkbehavior is thus a consequence of the viscoelasticity of the cytoplasmic space and has been

    used extensively to measure viscoelastic impedance of the cytoplasm (25, 26).

    More detailed information is obtained by measuring the directional persistence of the

    bead motion. It is defined as the MSD of the fluctuation of the velocity vector according to

    Dft2

    jft t ftj2D E

    , 5:

    where f is the angular direction of the velocity vector. Df(t) varies between 0 for directed

    and 2p for random motion. Below we summarize some pertinent results characteristic for

    Dictyostelium cells (27).

    Intracellular transport:actively and passivelydriven motion ofinternalized objects or

    intracellularorganelles in thecytoplasm

    Annu.R

    ev.Condens.MatterPhys.2

    010.1:257-276.Downloaded

    fromwww.annualreviews.o

    rg

    byIndianInstituteofScien

    ce-Bangaloreon05/15/12.Forpersonaluseonly.

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    8/22

    1. By measuring the mean square values DR(t)2 and Df(t)2 as a function oft and averag-

    ing over variable time intervals, one can distinguish between random motions [with

    a % 1 and Df(t)2 $ 2p] and directed motions [with Df(t)2 % 0 and exponents varyingbetween a % 1.8 and a % 2.0]. The latter are attributed to the directed active motionsalong the MTs. The velocity distribution is asymmetric, with a maximum at v $0.3 mm s

    1

    (see Figure 3b). This could be interpreted in terms of the action of different numbers of

    molecular motors or local variations of the cytoplasmic viscosity. In rheological studies

    (27), we measured a power law z % 104 ffiffitp and an effective cytoplasmic viscosity ofZ $10 Pas. This would correspond to forces of f% 50 pN, requiring the concerted actionof up to 10 motors of the kinesin family.

    2. The duration of the straight paths obeys an exponential distribution with a decay

    time of t

    %0.65 s, whereas the duration of the random walks obeys a log normal

    distribution.

    F(t)

    Start

    Endy-coordinate(m)

    8

    12

    16

    x-coordinate (m)

    8 12 16

    0.12

    0.08

    0.04

    0.00

    Pro

    bability

    1.00.60.2

    v (ms1)

    50

    30

    10

    1

    0.3

    0.2

    0.1

    0.03.02.01.00.0

    Time (s)

    P(D)x103

    D (mms1)

    P()

    a

    b

    c

    600

    450

    300

    150

    0

    Time (s) 102

    10

    4

    Figure 3

    (a) Example of quasi-random transport of colloidal probes in a Dictyostelium cell observed for about 10 min. The time course ofthe bead motion is indicated by a rainbow color code starting with red and ending with blue. The color-time relationship isshown in the inset to the right. Note that the motion consists of local, apparently random, walks and straight paths. In somecases, the bead moves very fast, exhibiting very large velocities of up to 5 mm s1. These earthquake-like motions are attributed tothe relaxation of prestressed microtubules. At some positions, strong force pulses (of 100 to 700 nN) were applied. Sometrajectories of the bead evoked by the force pulses are indicated by small arrows. Note that the bead is not deflected in thedirection of the external force (which is always directed toward the left, as shown in the inset to the left) but perpendicular ornearly opposite to it. (The trajectory was adopted from Reference 27.) (b) Distribution of velocities of straight bead motionsfulfilling the condition Df(t)2 % 0 and a % 2.0. Inset: Distribution of diffusion coefficients P(D) of the random walks. P(D) isplotted as a function of the logarithm of D. The Gauss-like distribution indicates a log normal distribution of D. ( c) DistributionP(t) of durations while the bead moves along nearly straight trajectories attributed to the transport along microtubules.

    Annu.R

    ev.Condens.MatterPhys.2

    010.1:257-276.Downloaded

    fromwww.annualreviews.o

    rg

    byIndianInstituteofScien

    ce-Bangaloreon05/15/12.Forpersonaluseonly.

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    9/22

    3. The diffusion coefficients exhibit a broad distribution. It can be well represented

    by a log-normal distribution showing that the random walks are not simple

    Brownian motions. Such behavior is typical for diffusion in complex fluids with fractal

    dimensionality.

    There is another interesting interpretation. The average value of the diffusivity D %102 mm2 s1 would correspond to a viscosity of Z% kBT/6pRD % 101 Pas, which issmaller by two orders of magnitude than the measured values ($10 Pas). This leads to theconclusion that the random motion is driven by quasi-random forces mediated by the

    actively driven motions in the cytoplasm, which could be interpreted in terms of an excess

    temperature Tex, which is much higher than the physiological value (see also Reference 16).

    In the cytoplasm, the combination of random and directed active motion exhibits two

    advantages. By active motion along the MT alone, the endosomes would be mainlytransported between the periphery and the cell center in the radial direction. Random

    walks alone could explore the whole space but would be very slow considering the high

    viscosity of the cytoplasm. Because the two motions are independent, the MSD for a

    straight motion and a subsequent random walk from one MT to a second one (in the

    distance dd) would be of the order

    hr2

    i %6Dt

    f D

    kBT koff

    2

    , 6:

    where f is the active force and the friction coefficient z has been replaced by D kBT/z.Equation 6 shows that the time t to reach a certain position at another MT would be t /6Dt and that the search for a specific target in the cytoplasm would be strongly

    reduced by the combined modes of motion. A second advantage of the search strategy is

    that the transition between two adjacent MTs is accelerated by the active sweeping

    motions of the MTs (corresponding to a higher effective diffusion coefficient). The higher

    efficiency of combined directed and random motions in cells has also been predicted bytheoretical studies (32, 33).

    THE CROSSTALK BETWEEN THE ACTIN CORTEX ANDTHE MT-ASTER CONTROLS THE GLOBAL MECHANICALSTABILITY OF CELLS

    A promising physical strategy to explore the crosstalk between the MT and the actin

    network is magnetic tweezer microrheology. For this purpose, superparamagnetic beads(of 2.8 mm diameter) are transferred into the cells, and changes in their quasi-random

    motion evoked by force pulses of $1 s duration are analyzed. The evaluation of severalhundred creep response functions obtained in Dictyostelium cells by force pulses of up to

    1 nN provided the following insights (see Reference 27 and Supplemental Appendix F):

    1. In most cases, the beads respond to forces below 100 pN in an unpredictable manner.

    The deflections are delayed and often consist of several steps. Moreover, the direction of

    deflection is seldom parallel to the force direction. This shows that the cytoplasmicspace of the cells does not behave as a passive viscoelastic system but rather as an active

    viscoplastic body.

    2. Above a threshold force of Fex ! 100 pN, the beads are deflected very fast (>5 mm s1)in a direction opposite or perpendicular to the applied force direction. The response

    Supplemental Material

    Annu.R

    ev.Condens.MatterPhys.2

    010.1:257-276.Downloaded

    fromwww.annualreviews.o

    rg

    byIndianInstituteofScien

    ce-Bangaloreon05/15/12.Forpersonaluseonly.

    http://arjournals.annualreviews.org/article/suppl/10.1146/annurev-conmatphys-070909-104105?file=co-1-sackman.pdfhttp://arjournals.annualreviews.org/article/suppl/10.1146/annurev-conmatphys-070909-104105?file=co-1-sackman.pdf
  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    10/22

    time is tR % 0.2 s or smaller. Simultaneously, the centrosome moves in the samedirection, and its motion is correlated with that of the bead.

    3. The nonlinear active response is also observed after knockout of the muscle type

    myosin II, showing that this motor is not responsible for the active viscoplastic behavior

    (see Figure 3b). It is abolished after dismantling either the actin network withlatrunculin A or the MT-aster with benomyl.

    4. After the very fast response, the plus end of the MTs (and the centrosome) often return

    close to the original position by diffusive motion with velocities of about 1 mm s1 or

    smaller (see Reference 27 and Supplemental Figure 6). This behavior suggests that the

    cells exhibit some mechanical short-term shape memory. It is lost when the cell starts to

    crawl in a new direction by formation of a new pseudopod. The transient shape mem-

    ory is determined by the balance of the tensile forces (FRX in Figure 2) in the MTand the

    shearing (s in Figure 2) in the actin cortex.

    The mechanical model can be considered as a special case of the famous Buckminster

    Fuller design concept of tensional integrity (tensegrity) structures composed of alternating

    arrangements of ropes and struts that are stabilized by appropriate prestresses [as postu-

    lated by Ingber and coworkers (34)]. The cell is unique in the sense that the tension-bearing

    elements are arranged as spikes in a wheel, whereas the compressions are balanced by the

    hoop of the wheel (34).

    The actin-MT connections play a key role for the separation of the two new sets ofchromosomes during cell division (cytokinesis) (35). Each set of chromosomes is coupled

    to the plus ends of the MT emanating from one side of the centrosomes, whereas those

    emanating from the opposite side couple to the actin cortex. The chromosomes are then

    separated by movement of the centrosomes to opposite sides. This process is driven by the

    movement of the two actin-MT assemblies to opposite poles of the cell until gobal mechan-

    ical equilibrium between the two halves of the cell is reached. The underlying mechanism

    of this process is not completely understood yet. One explanation is that it is driven by

    actin-myosin II micromuscles because the two sets of chromosomes are not positioned

    properly in mutants devoid of the motor protein (35, 36). The separation of the spindles

    can, however, be driven by the SAGWs as described below. The actin-MT crosstalk may

    also be responsible for the quasi-periodic oscillations of the centrosome observed occasion-

    ally during the spindle positioning (36).

    GLOBAL CELL MOTIONS AND CELLULAR SHAPE CHANGES

    Cells Crawl on Surfaces by Pseudopod Spreading, thereby TransmittingPushing Forces Via Adhesion Domains

    The most dramatic large-scale motions of cells occur during their crawling on surfaces by

    cyclic spreading of broad protrusions (pseudopods) at the front followed by the active

    retraction of the opposite end. In gradients of chemotactic signal molecules (e.g., c-AMP

    in the case of Dictyostelium), the generation of pseudopods is triggered by binding of thechemoattractants to cell surface receptors. The cells polarize and generate protrusions in

    the direction of the gradients. Elegant microfluidic experiments showed that the chemotac-

    tic speed increases with the steepness of the gradient between a threshold of 103 nM mm1

    and an upper limit of 10 nM mm1 (37).

    Tensional integrity(tensegrity) structures:

    mechanical structurescomposed of elements

    that can either onlybalance tensions (suchas the MT) orextensional loads(such as the cellenevelope)

    Annu.R

    ev.Condens.MatterPhys.2010.1:257-276.Downloadedfromwww.annualreviews.o

    rg

    byIndianInstituteofScien

    ce-Bangaloreon05/15/12.Forpersonaluseonly.

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    11/22

    However, the spontaneous generation of pseudopods at random sites is a basic activity

    of vital cells even in the absence of external stimulations. The cells perform a specific kind

    of random motion. It consists of persisting zig-zag-like motions for 1 to 2 min (over

    distances of about 20 mm), which are followed by a change in direction (see Figure 4a and

    References 38, 39).This saltatory motion is consistent with the dynamics of pseudopod spreading. They

    protrude with constant velocity for about 5 mm and stop abruptly. Then the rear side is

    retracted, and the cells unbind partially from the substrate, resulting in a reduction of the

    contact area. Thereafter, a new direction is chosen. Thus, the locomotion is accompanied

    by a cyclic variation of the contact area (e.g., between 20 and 200 mm2 on mica) with a

    periodicity of 12 min. The velocity of pseudopod spreading varies from v % 0.1 mm s1 onsoft polymer cushions (38) to v

    0.16 mm s1 on albumin-coated glass and 0.5 mm s1 on

    freshly cleaved mica. This shows that the spreading velocity depends on the adhesionstrength.

    In order to crawl, cells have to push the front of the pseudopod forward. Following

    Reference 40, these forces can be determined by measuring the deformation of soft

    substrates in front of the pseudopods by colloidal force microscopy, as shown in

    Figure 4a. The forces depend on the stiffness of the substrate. For example, f % 1 nNfor a polymer film with an elastic modulus of 200 Nm2 on which the pseudopods

    proceed at v 0.13m

    m s

    1

    . To fulfill Newtons third law, the protrusion forces on thecell are balanced by opposite forces on the substrate at the rear of the protrusions, which

    requires that the cell adheres on the substrate. The experiments also suggest that the

    mechanical stability is maintained by a local dipolar stress field (see Figure 4a) as pre-

    dicted theoretically (7).

    The local pinning occurs through small ($1 mm diameter) adhesion domains, whichare formed by lateral segregation of CAMs binding to the surface of other cells or tissue

    (as shown in Figure 3b). They are stabilized by anchoring the actin filaments to the

    intracellular domains of the CAMs, which is mediated by talin or other actin-membranelinkers. The adhesion domains can become very large and are then called focal adhesion

    complexes (41).

    Model membrane studies suggest that cell adhesion occurs in two steps: First, the fast

    generation of the domains of tight adhesion by lateral segregation of bound receptors, and

    second, the stabilization of the domains by assembly of actin gels to the intracellular side of

    the adhesion domains (2, 42, 43). The initial process exhibits typical features of first-order

    wetting transitions, resulting in the formation of small adhesion domains enabling strong

    adhesion with a small number of receptors (43). The adhesion strength of the domainsdepends on the bending stiffness and the lateral tension of the composite cell envelope and

    is thus strongly increased by binding of the actin cortex to the adhesion domains (44).

    The kinetics of the first stage of adhesion is determined by the pronounced thermally

    and actively driven bending excitations, resulting in a dynamic surface roughness of about

    10 nm, generating an entropic disjoining pressure pdisj (summarized in Supplementary

    Appendix G and References 16 and 17). The dynamic surface roughness controls the

    adhesion dynamics in two ways: It impedes adhesion, but it also provides the pushingforces that bring the receptors and ligands close enough for bond formation (18). There-

    fore, the adhesion is initiated by a nucleation and growth process, and the adhesion

    kinetics depend critically on the binding strength of the receptor-ligand-pairs (which varies

    between 5 and 20 kBT) and the lateral density of the CAMs.

    Supplemental Material

    Annu.Rev.Condens.MatterPhys.2010.1:257-276.Downloadedfromwww.annualreviews.o

    rg

    byIndianInstituteofScien

    ce-Bangaloreon05/15/12.Forpersonaluseonly.

    http://arjournals.annualreviews.org/article/suppl/10.1146/annurev-conmatphys-070909-104105?file=co-1-sackman.pdfhttp://arjournals.annualreviews.org/article/suppl/10.1146/annurev-conmatphys-070909-104105?file=co-1-sackman.pdf
  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    12/22

    CAM

    CAM(integrin)

    Myosina

    c

    b

    Signal

    VASP

    ATP-Actin

    Molecular switch (activated)

    Molecularswitch

    (inactive)

    Arp2/3

    ADP-Actin

    P

    Cappingprotein

    Formin

    Strainprobes

    Soft tissue

    x

    y

    z

    +

    20 m

    60

    120

    180

    240300

    0

    ATP

    Proflin

    Activated actin

    F-actin

    Filipod

    ium

    Mem

    brane

    InactiveRho X Active

    +

    Bead

    Activation o scafolding protein

    Mechanism o actin growth by ormin

    Figure 4

    (a) Quasi-random walk of a Dictyostelium cell crawling on freshly cleaved mica (on which cells adhere weakly). Thecell position was recorded every 10 s. The cell moves in one direction in a saltatory manner and changes the directionof motion about every 60 s by spreading a pseudopod in a new direction. (b) Measurement of protrusion forcesgenerated by polar actin growth at the front of the protrusion by analyzing the shear deformation field of softsubstrates using colloidal strain probes (following Reference 40). Protrusion forces are balanced by the shear deforma-tion of the substrate behind the adhesion domains mediating the mechanical coupling between the substrate and the cell

    envelope (see Supplemental Appendix E). (c) Simplified treadmilling model of formation of two types of protrusions.Sharp protrusions (filipodia) generated by activated growth of actin bundles mediated by the actin growth promoterformin, and broad protrusions (pseudopods) generated by growth of branched network mediated by the cross-linkerArp2/3. The growth of the broad pseudopods is initiated by a large protein complex, such as vasolidator-stimulatedphosphoprotein (VASP), shown in the top right box. It acts as scaffolding proteins, which recruit the proteins requiredfor the polar growth of the actin gel, including the cross-linker Arp2/3 and ATP-actin to the front of the pseudopods.The activated ATP-actin monomer is provided by unbinding from the filaments at the minus ends, a process mediatedby the actin binding proteins cofilin (or coronin, see also Figure 5). The thick arrow indicates the flow of activatedactin monomers to the front of the pseudopod. The scaffolding protein is transformed from a sleeping state into anactive state. This step is triggered by a molecular switch of the GTPase family, such as Cdc42 (see Reference 47 andSupplemental Material B). In the bottom right box, the mechanism of actin growth by formin is illustrated (followingReferences 51 and 52).

    Annu.Rev.Condens.MatterPhys.2010.1:257-276.Downloadedfromwww.annualreviews.o

    rg

    byIndianInstituteofScien

    ce-Bangaloreon05/15/12.Forpersonaluseonly.

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    13/22

    The adhesion domains ofDictyostelium cells can be visualized by fluorescence micros-

    copy using labeled actin-membrane coupler talin (45). In mammalian cells, adhesion

    is often mediated by receptors of the integrin family, a dimer of membrane-spanning chains

    a and b, whereby the b-chain binds to talin. Recently, a CAM (spanning the membrane

    with 9 helices) sharing common features with the integrin b-chain was discovered inDictyostelium cells (46). Thus, the adhesion of amoebae and mammalian cells may share

    common features.

    A Treadmilling Process Propels Pseudopod Spreading

    A widely accepted model of the growth of pseudopods (illustrated in Figure 3b) suggests

    that it is controlled by three processes: (a) unbinding of G-actin with bound ADP (ADP-

    actin) from the minus end of matured actin filaments through the G-actin providerscofilin; (b) the activation of G-actin by ADPATP exchange mediated by profilin; and(c) the growth of actin gels (bundles or branched networks) at the front, pushing the

    pseudopod forward. The last process is mediated by two types of actin growth pro-

    moters together with actin cross-linkers: Pseudopod formation is accelerated by a scaf-

    folding protein [such as vasolidator-stimulated phosphoprotein (VASP)], which helps to

    recruit the proteins required for the actin growth, including the cross-linker Arp2/3 (see

    Figure 3b and Supplemental Appendix D). A second class of growth promotors, namely

    formin (and mDia 1), together with the cross-linker fascin, generates tube-like protru-

    sions, called filipodia.

    An impressive proof of the treadmilling process as force generator has been pro-

    vided by the observation that bacteria propel themselves through cells by polar

    growth of actin gels. These bacteria exhibit at their surface a specific protein (Act A)

    that mimicks the function of VASP as scaffolding protein (47). It mediates the activa-

    tion of Arp2/3, resulting in the generation of comet-like gels that drive the bacteria

    forward with velocities of 0.07 mm s1

    . The process can be reproduced in vitro bycovering colloidal beads with VASP and suspending them in solution containing

    (a) actin filaments, (b) cofilin (which decouples G-actin from the F-actin), and (c) Arp2/3

    and ATP (48, 49). Evidence has been provided that the motor acts as a Brownian

    ratchet (50).

    Another mechanism for generating the protrusion force that drives the propagation of

    filipodia (see Figure 4) is based on the sequential attachment of actin monomers to the tip

    of the actin bundle. The growth is mediated by the growth promoter formin. With this

    mechanism, forces of up to 10 pN per filament can be generated (49, 51, 52). The speed ofadvancement is controlled by the generation of free volume between the tip of the actin

    filaments and the membrane, suggesting that this force generator works as a Brownian

    ratchet, too (49).

    Solitary Actin Gelation Wave Model of Pseudopod Spreading

    The treadmilling process can be described in terms of a SAGW that is generated by polar

    growth of the actin network at the front and the decay of matured actin gel at the end of

    the wave. The basic idea is that the polar growth of the actin network at the front is

    mediated by localized recruitment of a supramolecular scaffolding protein complex (com-

    posed of 5 proteins) to the inner leaflet of the adhering cell envelope. This scaffolding

    Scaffolding proteins:protein complexes thatcan bind one or severaleffectors involved in

    the activation ofsignaling pathways

    Supplemental Material

    Annu.Rev.Condens.MatterPhys.2010.1:257-276.Downloadedfromwww.annualreviews.org

    byIndianInstituteofScience-Bangaloreon05/15/12.Forpersonaluseonly.

    http://arjournals.annualreviews.org/article/suppl/10.1146/annurev-conmatphys-070909-104105?file=co-1-sackman.pdfhttp://arjournals.annualreviews.org/article/suppl/10.1146/annurev-conmatphys-070909-104105?file=co-1-sackman.pdf
  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    14/22

    protein recruits the molecules required for the actin gelation, such as profilin, activated

    ATP-actin, and Arp2/3. It resides as a sleeping (closed) form in the cytoplasm and is

    activated by the binding of members of the Rho-family GTPases (Cdc42 in the case of

    VASP). They act as molecular switches that open the closed conformation (see Supplemen-

    tal Appendix B).The scaffold proteins of different cells share some common structural homologies but

    are named differently in the literature. WAVE/Hem-1 predominately refers to leukocytes

    (53), VASP/Ena to fibroblasts (24, 47), and CARMIL/Myosin-IB to Dictyostelium cells

    (54, 55). Hem1, Ena, and myosin-IB are supposed to be the key actuators (effectors) that

    drive actin growth. They are supposed to undergo cycles of activation (by binding to the

    scaffold protein) and deactivation (by unbinding).

    PI-3,4,5-P3

    PI-4,5-P2 gradients act as a guidance field of cell locomotion. Although

    the role of the scaffold protein and the components involved is well established, the mech-anism of triggering or propagation of the wave is still obscure. The present view is that the

    SAGW is driven by a local enrichment of the signal molecule PI-3,4,5-P3 at the adhered

    plasma membrane at the front of the cell, resulting in a gradient PI-3,4,5-P3PI-4,5-P2toward its rear end. The gradient is generated by the antagonistic pair of enzymes:

    phosphoinositide-3-kinase (PI-3K), which couples phosphate groups to the 3-position of

    the inositol head group, and the conjugate phosphatase (PI-3PH, often called PTEN),

    acting as PI-3,4,5-P3 annihilator. Both enzymes bind to the phosphoinositides via specific

    domains (called C2-domains; see Supplemental Apprendixes C and D). The gradient stim-

    ulates the actin polymerization at the front and the subsequent retraction of the rear end of

    cells during each cycle of crawling (53).

    The role of the PI-3,4,5-P3/PI-4,5-P2 gradient for the cell polarization is demon-

    strated by the following observation (56): If Dictyostelium cells are subjected to a

    local gradient of a chemoattractant (cAMP), then the PI-3,4,5-generator (PI-3K) accu-

    mulates at the front and the annihilator (PI-3PH) at the rear of the cell. The retrac-

    tion of the rear end is driven by the activation of muscle myosin II motors and theformation of actin-myosin micromuscles. Myosin-II is known to be activated by the

    GTPase Rho, whereas the actin polymerization at the front is activated by Cdc42.

    The gradient is therefore attributed to the negative feedback between the two signal-

    ing systems (53).

    Several formal models of SAGW generation have been proposed, all based on the

    idea that the waves are generated by three properties of the actin gelation machinery.

    One model states that actin polymerization is triggered by recruitment of an actin

    polymerization promoter (e.g., Hem-1 in the case of leukocytes and myosin I in the caseof Dictyostelium cells) from the cytosol to the adhering membrane (53, 54). A second

    model is that the growth rate of actin decreases due to the depletion of the pool of

    G-actin, which introduces a delay in the rate of gelation. A third model is that the

    recruitment of the promoter is cooperative, implying that its rate of binding to the

    membrane increases with the amount already bound (57, 58). This results in a pair of

    coupled differential equations for the actin growth rate, which are similar to the tradi-

    tional equations describing the propagation of flames or of action potentials along nerve

    axons. However, these models do not consider the role of the adhering composite mem-

    brane. The experiment described below suggests an important role for the adhesion-

    induced aggregation of receptor-ligand pairs mediating the adhesion of cells to the

    substrate.

    CARMIL: Cappingprotein, Arp2/3,Myosin I Linker

    PI-3,4,5-P3:

    phosphatidylinositol3,4,5-triphosphate

    PI-4,5-P2: phosphati-dylinositol-4,5-diphos-phate

    Cell locomotion:creeping motion ofcells on surfaces

    PI-3K (-PH):phosphoinositolkinase (phosphatase)

    Annu.Rev.Condens.MatterPhys.2010.1:257-276.Downloadedfromwww.annualreviews.org

    byIndianInstituteofScience-Bangaloreon05/15/12.Forpersonaluseonly.

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    15/22

    SOLITARY ACTIN GELATION WAVES AS TRAVELING FORCEFIELD MOTORS

    Frustrated Phagocytosis Driven by Solitary Actin Gelation Waves

    A paradigm of a motional process propelled by actin gelation is phagocytosis, the internal-ization of large particles by cells. This process occurs in two steps. First, the cell envelope

    spreads over the particle, with spreading velocities of $0.1 mm s1. Subsequently, theparticle is engulfed completely by fusion of the front rim of the cell lobe. The efficiency of

    the primary process depends also on the participation of microtubules. Their dismantling

    by poisons (e.g., colchicine) results in the slowing down of the process (but not in its

    abolishment).

    The experiment in Figure 5a shows that the engulfment of large particles by Dictyostelium

    cells is propelled by actin gelation. The objects consist of elongated elevations (with rect-angular cross sections) fabricated on solid substrates [composed of poly-dimethylsiloxan

    (PDMS), an inorganic polymer]. The cells spread over these objects with velocities

    of $0.04 mm s1, but cannot engulf them. We thus deal with frustrated phagocytosis.The time evolution of newly polymerized actin can be visualized by LIM-proteins labeled

    with red fluorescent protein (for a discussion of the function of LIM-proteins, see Supple-

    mental Appendix B and References 54 and 59). Simultaneously, the function of the

    actin monomer supplier coronin is observed by labeling these proteins with green fluorescent

    proteins (GFPs).

    The cells spread over one (or sometimes two) pillars, acting as phantom prey, and move

    to the next elevation. This motion is accompanied by the formation of one or several knob-

    like assemblies of actin gels, which form preferentially at the top edge and side walls of the

    pillars. They propagate along the upper edge and side walls and appear to drive the cell

    movement. In some cases, one observes the simple behavior shown in Figure 5a. Two actin

    assemblies form on opposite sides of the pillar and slide in a coordinated way in one

    direction, carrying the cell with them. The propagating actin assemblies are inwroughtwith coronin, which provides the reservoir of actin required for the actin growth.

    Model of Traveling Force Field Motor Mediated byPropagating Gradient of Signaling Lipid PI-3,4,5-P3

    Figure 5b shows a model of SAGW-driven phagocytosis, described below. The basic

    assumption is that the polar growth of the actin gel is determined by the propagating PI-

    3,4,5-P3/PI-4,5-P2 gradient in the adhering membrane. It is generated by interplay of thePI-3-kinase (a PI-3,4,5-P3 generator) and the PI-3-phosphatase (PTEN) (a PI-3,4,5-P3-

    annihilator). PI-3,4,5-P3 is accumulated in the adhesion domains by recruitment of PI-3K,

    as observed in Reference 57 (justified in Supplemental Apprendix A).

    The polar growth of the gel is mediated by the actin polymerization promoter coronin

    playing two roles (see, for example, Supplemental Appendix B). It removes small pieces of

    F-actin at the trailing end and transfers them to the growing front. Here, the unique

    capacity of coronin to bind strongly fragments of F-actin with bound ATP (and not to

    monomeric ADP-actin, such as cofilin) comes into play.

    The actin polymerization rate is controlled by CARMIL (60, 61), which recruits the

    players involved in the polar growth of actin to the adhesion domains, including Arp2/3

    and the motor protein myosin IB. The rate-limiting step is the unbinding of the capping

    LIM-protein:polypeptide exhibitingone or more cystein-rich protein motifs

    called LIM domains,which promote actinpolymerization andare characteristic foractin-networkregulating activity

    GFP (RFP): green (red)fluorescent protein

    Supplemental Material

    Annu.Rev.Condens.MatterPhys.2010.1:257-276.Downloadedfromwww.annualreviews.org

    byIndianInstituteofScience-Bangaloreon05/15/12.

    Forpersonaluseonly.

    http://arjournals.annualreviews.org/article/suppl/10.1146/annurev-conmatphys-070909-104105?file=co-1-sackman.pdfhttp://arjournals.annualreviews.org/article/suppl/10.1146/annurev-conmatphys-070909-104105?file=co-1-sackman.pdf
  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    16/22

    proteins at the plus end by CARMIL, enabling the attachment of fragments of F-actin

    bound to coronin (62, see also Supplemental Appendix D).

    The binding of CARMIL to the membrane can be mediated by myosin IB and by the

    phosphoinositides (see Figure 5b). Myosin IB can bind to the PM either directly, by

    electrostatically binding its long basic tail to the negatively charged lipid moiety of themembrane, or indirectly, by binding its tail domain to the actin filaments already anchored

    to the adhesion domains via talin (as shown Figure 4b).

    By the uncapping process via CARMIL, only assemblies of linear actin filaments can

    be generated. Branched networks can be formed in a second step by activation of Arp2/3.

    This can only occur after the coronin is phosphorylated. Evidence for the recruitment of

    a

    b

    Talin

    PI-4,5-P2

    PI-3K

    CARMIL

    Myo-IB

    K

    Cap

    Receptor(CAM)

    PI-3PH

    Arp2,3

    G-Actin

    F-Actin

    PHK K

    PI-3,4,5-P3

    PH PH

    PH

    Contour

    K

    Coronin

    Propagation

    = = = =

    =

    45 51 57 63 69

    Tissue

    PM

    RFP GFP

    Figure 5

    (a) Visualization of actin gelation waves by fluorescence microscopy of cells double labeled with red fluorescent protein (RFP)LIM and green fluorescent protein (GFP)-labeled coronin. The cell crawls over a rectangular elevation (6 21 5 mm3). Thebrownish rim defines the contour of the cell. For simplicity, only five snapshots of the contour of the cell and of the distribution offreshly polymerized actin and coronin at the time intervals (seconds) indicated are shown. Note: That the cell adheres only

    slightly on top of the pillar (where only small actin gel patches are seen), and the propagating acting gelation wave (SAGW)moves the cell along the pillar. (b) Model of the cell motion as described in the text, showing the major molecular ingredientsinvolved in the SAGW generation. Note that the scaffolding protein complex CARMIL (Capping protein, Arp2/3, Myosin ILinker) can couple to the membrane via myosin IB or by binding to phosphoinositides (PI-3,4,5-P3 or PI-4,5-P2). Also note thatthe PI-3,4,5-P3-generating kinase PI-3K is recruited to the membrane after coupling of CAMs to the tissue (64).

    Annu.Rev.Condens.MatterPhys.2010.1:257-276.Downloadedfromwww.annualreviews.org

    byIndianInstituteofScience-Bangaloreon05/15/12.

    Forpersonaluseonly.

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    17/22

    CARMIL by myosin IB and the delayed formation of branched networks is provided by the

    following observation: At the front of the SAGW, the newly synthesized actin (as indicated

    by LIM) and myosin IB appear simultaneously (e.g., within about 1 s after a sudden

    generation of a gradient of chemoattractants cAMP). In contrast, the recruitment of Arp2/

    3 to the front is delayed by 12 s (55).Concerning the propagation of the SAGW, we assume that the PI-3,4,5-P3/PI-4,5-P2

    gradient is driven by the constant recruitment of PI-3K at the front of the spreading

    pseudopod where the adhesion domains are constantly regenerated (63). At this site,

    new CAMs are recruited that activate the PI-3K to generate new signaling lipids (Refer-

    ence 64 and Supplemental Appendix A). Due to receptor clustering, the recruitment of the

    CARMIL/myosin IB complex is a cooperative process.

    The role of the MT network is still poorly understood, although it is well established

    that it plays a regulatory role for global cell movements and propagation of cellular pro-trusions. Thus, MTs are required for effective phagocytosis because the particle uptake is

    much slower if the MTs are removed. An established function of the MT is to recruit and

    activate the PI-3,4,5-P3 generator PI-3K to membranes because concentration of the sig-

    naling lipid in the membrane is lowered after MT decomposition (64). This mechanism

    could also trigger the growth of actin assemblies at the poles of the dividing cell, mediating

    the final positioning of the spindles (24) as suggested above.

    CONCLUSION

    SAGWs enable cells to generate nano-Newton forces, which propell cell locomotion or

    global shape changes (e.g., phagocytosis). The protrusion forces are not generated by the

    motion of motors but by propagating assemblies of the signal molecules PI-3,4,5-P3, which

    mediate the local recruitment of the actuators (CARMIL and myosin IB), stimulating the

    polar growth of the actin gel. With the SAGW-driven system, much larger forces can be

    generated than with intracellular linear motors, although motions can only be generatedover small distances ($5 mm). For that reason, the persistent motions ofDictyostelium cellsconsist of zig-zag-like paths of about 5-mm step length (as shown in Reference 38 and

    visualized in Figure 3b).

    The propagation of the PI-3,4,5-P3 PI-4,5-P2 gradient is a membrane-bound processand shares common features with the diffusive transport of action potentials along nerve

    axons. The action potential is generated by (electric fieldtriggered) cooperative opening of

    many ion channels, but the propagation is driven by the local tangential transport of the

    Na ions at the inner surface of the axon. The delayed opening of the K ions prevents thebackflow of the action potential. The role of Na channels appears to be played by the

    accumulation of new PI-3,4,5-P3 lipids at the newly formed adhesion domains, whereas

    the role of K ions is played by coronin, which dismantles the F-actin at the rear side.

    SUMMARY POINTS

    1. The cytoplasmic space of cells is an active viscoplastic body that is mechanically

    stabilized by interactive crosstalk between the viscoelastic shell and the aster-like

    array of microtubules embedded in the viscoplastic cytoplasmatic space.

    Supplemental Material

    Annu.Rev.Condens.MatterPhys.2010.1:257-276.Downloadedfromwww.annualreviews.org

    byIndianInstituteofScience-Bangaloreon05/15/12.

    Forpersonaluseonly.

    http://arjournals.annualreviews.org/article/suppl/10.1146/annurev-conmatphys-070909-104105?file=co-1-sackman.pdfhttp://arjournals.annualreviews.org/article/suppl/10.1146/annurev-conmatphys-070909-104105?file=co-1-sackman.pdf
  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    18/22

    2. Freely swimming and adhering cells exhibit typical features of tensegrity struc-

    tures, with microtubules acting as ropes, whereas the actin cortex plays the role of

    a compressible strut.

    3. The exploration of the intracellular space by organelles and phagosomes is opti-mized by superposition of random walks (driven mainly by nonthermal quasi-

    random intracellular forces) and active motions along the microtubules.

    4. Migrating cells explore the environment by Levy-like quasi-random walks com-

    posed of persistent (zig-zag-like) paths and local reorientational motions. The

    directed motion is driven by cyclic spreading of cell lobes (pseudopods) powered

    by solitary acting gelation waves (SAGWs), followed by retraction of the rear of

    the cell mediated by actin-myosin micromuscles.

    5. SAGWs can act as traveling force-field motors that drive cells over local obstacles.The traveling force field is guided by the propagation of gradients of the signaling

    lipids PI-3,4,5-P3 and PI-4,5-P2 (embedded in the adhering cell membrane),

    which is possibly coupled to the migration of adhesion domains. The process is

    reminiscent of the diffusive propagation of action potentials along nerve axons.

    FUTURE ISSUES1. Development of micromechanical tools is needed for the measurement of abso-

    lute viscoelastic moduli of cell envelopes and intracellular macromolecular scaf-

    folds of living cells.

    2. Investigation into cell signal processes triggered by mechanical forces from the

    molecular level (such as force-induced control of enzyme functions and protein-

    protein interaction) to the mesoscopic scale (such as the control of membrane

    interactions by undulation-induced entropic forces) will provide more detailedinsights into the physical basis of cell movement.

    3. The role of the interactive microtubule actin crosstalk in the polarization of

    moving cells, in the sensing and balancing of external forces, and in the control

    of cell-cell and cell-tissue adhesion should be clarified.

    4. Theoretical concepts must be developed for the understanding and evaluation of

    intracellular motions driven by quasi-random forces in systems far from thermo-

    dynamic equilibrium.

    DISCLOSURE STATEMENT

    The authors are not aware of any affiliations, memberships, funding, or financial holdings

    that might be perceived as affecting the objectivity of this review.

    LITERATURE CITED

    1. Lipowsky R, Sackmann E. 1995. Handbook of Biological Physics, Vol. I. Elsevier: Amsterdam2. Sackmann E. 2006. J. Phys. Condens. Matter 18:R78595

    3. Fabry B, Maksym GN, Butler JP, Glogauer M, Navajas D, Fredberg JJ. 2001. Phys. Rev. Lett.

    87:148102

    4. Feneberg W, Aepfelbacher M, Sackmann E. 2004. Biophys. J. 87:133850

    Annu.Rev.Condens.MatterPhys.

    2010.1:257-276.Downloade

    dfromwww.annualreviews.org

    byIndianInstituteofScience-Bangaloreon05/15/12.

    Forpersonaluseonly.

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    19/22

    5. Semmrich C, Storz T, Glaser J, Merkel R, Bausch A, Kroy K. 2007. Proc. Natl. Acad. Sci. USA

    104:20199203

    6. Tzima E, Irani-Tehrani M, Kiosses WB, Dejana E, Schultz DA, et al. 2005. Nature 437:42631

    7. Bischofs IB, Safran SA, Schwarz US. 2004. Phys. Rev. E Stat. Nonlin. Soft Matter Phys.

    69:021911

    8. Engler A, Sen S, Sweeney H, Discher D. 2006. Cell126:67789

    9. Siechen S, Yang S, Chiba A, Saif T. 2009. Proc. Natl. Acad. Sci. USA 106:1261116

    10. Springer TA. 1994. Cell76:30114

    11. Dustin ML, Colman D. 2002. Science 298:78589

    12. Critchley DR, Gingras A. 2008. J. Cell Sci. 121:134547

    13. Sackmann E, Bausch A, Vonna L. 2002. Physics of Bio-Molecules and Cells. Berlin: Springer

    14. Bausch A, Kroy K. 2006. Nat. Phys. 2:23138

    15. Borukhov I, Bruinsma RF, Gelbart WM, Liu AJ. 2005. Proc. Natl. Acad. Sci. USA 102:367378

    16. Auth T, Safran SA, Gov N. 2007. Phys. Rev. E Stat. Nonlin. Soft Matter Phys. 76:051910917. Zidovska A, Sackmann E. 2006. Phys. Rev. Lett. 96:0481037

    18. Pierres A, Benoliel A, Touchard D, Bongrand P. 2008. Biophys. J. 94:411422

    19. Smith AS, Sackmann E. 2009. ChemPhysChem 10:6678

    20. Boal D. 2001. Mechanics of the Cell. Cambridge, UK: Cambridge Univ. Press

    21. Mukhopadhyay R, Lim GHW, Wortis M. 2002. Biophys. J. 82:175672

    22. Noguchi H. Gompper G. 2005. Proc. Nat. Acad. Sci USA. 102:1415964

    23. Gundersen G, Bulinski J. 1988. Proc. Natl. Acad. Sci. USA 85:594650

    24. Kaverina I, Krylyshkina O, Small JV. 2002. Int. J. Biochem. Cell Biol. 34:7466125. Crocker JC, Valentine MT, Weeks ER, Gisler T, Kaplan PD, et al. 2000. Phys. Rev. Lett.

    85:88891

    26. Yamada S, Wirtz D, Kuo S. 2000. Biophys. J. 78:173647

    27. Heinrich D, Sackmann E. 2006. Acta Biomater. 2:61925

    28. Schu tz GJ, Axmann M, Freudenthaler S, Schindler H, Kandror K, et al. 2004. Microsc. Res. Tech.

    63:15967

    29. Caspi A, Granek R, Elbaum M. 2002. Phys. Rev. E Stat. Nonlin. Soft Matter Phys. 66:011916

    30. Arcizet D, Meier B, Sackmann E, Radler J, Heinrich D. 2008. Phys. Rev. Lett. 101:2481038

    31. Gittes F, Schnurr B, Olmsted PD, Mackintosh FC, Schmidt CF. 1997. Phys. Rev. Lett. 79:328689

    32. Loverdo C, Benichou O, Moreau M, Voituriez R. 2008. Nat. Phys. 4:13437

    33. Klumpp S, Lipowsky R. 2005. Phys. Rev. Lett. 95:2681027

    34. Ingber DE, Heidemann SR, Lamoureux P, Buxbaum RE. 2000. J. Appl. Physiol. 89:166378

    35. Rosenblatt J, Cramer L, Baum B, McGee K. 2004. Cell117:36172

    36. Grill SW, Howard J, Schaffer E, Stelzer EH, Hyman AA. 2003. Science 301:51821

    37. Song L, Nadkarni SM, Bo deker HU, Beta C, Bae A, et al. 2006. Eur. J. Cell Biol. 85:98189

    38. Schindl M, Wallraff B, Deubzer B, Witke W, Gerisch G, Sackmann E. 1995. Biophys. J.

    68:11779039. Li L, Nrrelykke SF, Cox EC. 2008. PLoS One 3(5):e209311

    40. Ladam G, Vonna L, Sackmann E. 2005. Acta Biomater. 1:48591

    41. Balaban NQ, Schwarz US, Riveline D, Goichberg P, Tzur G, et al. 2001. Nat. Cell Biol. 3:46672

    42. Smith A, Sengupta K, Goennenwein S, Seifert U, Sackmann E. 2008. Proc. Natl. Acad. Sci. USA

    105:690611

    43. Bruinsma R, Behrisch A, Sackmann E. 2000. Phys. Rev. E Stat. Phys. Plasmas Fluids Relat.

    Interdiscip. Topics 61:425367

    44. Simson R, Wallraff E, Faix J, Niewoehner J, Gerisch G, Sackmann E. 1998. Biophys. J. 74:51422

    45. Tsujioka M, Joshida K, Inouye K. 2004. EMBO J. 23:221625

    46. Cornillon S, Gebbie L, Benghezal M, Nair P, Keller S, et al. 2006. EMBO Rep. 7:61721

    47. Le Clainche C, Carlier MF. 2008. Physiol. Rev. 88:489563

    48. Loisel TP, Boujemaa R, Pantaloni D, Carlier MF. 1999. Nature 401:61316

    Annu.

    Rev.Condens.MatterPhys.

    2010.1:257-276.Downloade

    dfromwww.annualreviews.org

    byIndianInstituteofScience-Bangaloreon05/15/12.

    Forpersonaluseonly.

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    20/22

    49. Mogilner A, Oster G. 1996. Biophys. J. 71:303045

    50. Shaevitz J, Fletcher D. 2007. Proc. Natl. Acad. Sci. USA 104:1568892

    51. Kovar DR, Pollard TD. 2004. Proc. Natl. Acad. Sci. USA 101:1472530

    52. Shemesh T, Kozlov M. 2007. Biophys. J. 92:151221

    53. Weiner O, Marganski W, Wu L, Altschuler S, Koirschnert M, et al. 2007. PLoS Biol. 5:e221

    54. Gerisch G, Bretschneider T, Mu ller-Taubenberger A, Simmeth E, Ecke M, et al. 2004. Biophys. J.

    87:3493509

    55. Etzrodt M, Ishikawa H, Dalous J, Mueller-Taubenberger A, Bretschneider T, Gerisch G. 2006.

    FEBS Lett. 580:670713

    56. Chen L, Janetopoulos C, Huang Y, Iijima M, Borleis J, Devreotes P. 2003. Mol. Biol. Cell

    14:502837

    57. Dormann D, Weijer G, Dowler S, Weijer J. 2004. J. Cell Sci. 117:6497509

    58. Doubrovinski K, Kruse K. 2008. Euro. Phys. Letters 83:180037

    59. Khurana T, Khurana B, Noegel A. 2001. Protoplasma 219:11260. Jung G, Remmert K, Wu X, Volosky JM, Hammer JA III. 2001. J. Cell Biol. 153:147997

    61. Uruno T, Remmert K, Hammer JA. 2006. J. Biol. Chem. 281:1063550

    62. Miyoshi T, Takahiro T, Higashida C, Hertzog M, Fujita A, Narumyia S. 2006. J. Cell Biol.

    175:94755

    63. Rottner K, Behrendt B, Small JV, Wehland. 1999. Nat. Cell Biol. 1:32122

    64. Melikova S, Dylla S, Verfaillie C. 2004. Exp. Hematol. 32:105156

    Annu.

    Rev.Condens.MatterPhys.

    2010.1:257-276.Downloade

    dfromwww.annualreviews.org

    byIndianInstituteofScience-Bangaloreon05/15/12.

    Forpersonaluseonly.

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    21/22

    Annual Review of

    Condensed Matter

    Physics

    Contents

    Electron Transport in Carbon NanotubesShahal Ilani and Paul L. McEuen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

    FeAs-Based Superconductivity: A Case Study of the Effects of

    Transition Metal Doping on BaFe2As2Paul C. Canfield and Sergey L. Budko . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

    Scattering and Pairing in Cuprate Superconductors

    Louis Taillefer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

    Spintronics

    S.D. Bader and S.S.P. Parkin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71

    Characterizing Graphene, Graphite, and Carbon Nanotubes by

    Raman Spectroscopy

    M.S. Dresselhaus, A. Jorio, and R. Saito . . . . . . . . . . . . . . . . . . . . . . . . . . 89

    Single-Molecule NanomagnetsJonathan R. Friedman and Myriam P. Sarachik . . . . . . . . . . . . . . . . . . . . 109

    Fermi-Hubbard Physics with Atoms in an Optical Lattice

    Tilman Esslinger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

    Nematic Fermi Fluids in Condensed Matter Physics

    Eduardo Fradkin, Steven A. Kivelson, Michael J. Lawler,

    James P. Eisenstein, and Andrew P. Mackenzie . . . . . . . . . . . . . . . . . . . . 153

    The Coulomb Phase in Frustrated Systems

    Christopher L. Henley . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179

    First-Principles Calculations of Complex Metal-Oxide Materials

    Karin M. Rabe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211

    X-Ray Diffraction Microscopy

    Pierre Thibault and Veit Elser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237

    Volume 1, 2010

    vi

    Annu.Rev.Condens.Matte

    rPhys.2010.1:257-276.Dow

    nloadedfromwww.annualreviews.org

    byIndianInstitute

    ofScience-Bangaloreon05

    /15/12.Forpersonaluseonly

    .

  • 7/31/2019 Physics of Cellular Movements - Supplementary Info

    22/22

    Physics of Cellular Movements

    Erich Sackmann, Felix Keber, and Doris Heinrich . . . . . . . . . . . . . . . . . . 257

    Molecular Theories of Segmental Dynamics and Mechanical

    Response in Deeply Supercooled Polymer Melts and Glasses

    Kang Chen, Erica J. Saltzman, and Kenneth S. Schweizer . . . . . . . . . . . . . 277

    Rheology of Soft Materials

    Daniel T.N. Chen, Qi Wen, Paul A. Janmey, John C. Crocker, and

    Arjun G. Yodh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

    The Mechanics and Statistics of Active Matter

    Sriram Ramaswamy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323

    The Jamming Transition and the Marginally Jammed SolidAndrea J. Liu and Sidney R. Nagel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347

    Dynamics of Simple Cracks

    Eran Bouchbinder, Jay Fineberg, and M. Marder . . . . . . . . . . . . . . . . . . . 371

    Friction, Fracture, and Earthquakes

    Eric G. Daub and Jean M. Carlson . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397

    Errata

    An online log of corrections to Annual Review of Condensed Matter Physics

    articles may be found at http://conmatphys.annualreviews.org/errata.shtml

    Contents vii

    Annu.Rev.Condens.Matte

    rPhys.2010.1:257-276.Dow

    nloadedfromwww.annualreviews.org

    byIndianInstitute

    ofScience-Bangaloreon05

    /15/12.Forpersonaluseonly.