Polymer Motors: Pushing out theFront and Pulling up the BackAlex Mogilner1and George Oster2Mechanical work in cells is performed by specializedmotor proteins that operate in a continuousmechanochemical cycle. Less complex, but stillefficient, ‘one-shot’ motors evolved based on theassembly and disassembly of polymers. We reviewthe mechanisms of pushing and pulling by actin andmicrotubule filaments and the organizationalprinciples of actin networks. We show how thesepolymer force generators are used for thepropulsion of intracellular pathogens, protrusion oflamellipodia and mitotic movements. We discussseveral examples of cellular forces generated by theassembly and disassembly of polymer gels.IntroductionMore than 20 years ago, Abercrombie noted thatcrawling cells move in three stages [1,2]. First, theypush out their leading edge; then they strengthen theiradhesions at the leading edge and weaken them atthe trailing edge; finally, they pull up their rear. Protru-sion involves generating pushing forces at the front,and pulling up the back involves contractile forces.Successes in cell biology, genetics, biophysics andmodeling over the past few decades have dissectedthe cell migration phenomena so that the processappears less mysterious than it did in Abercrombie’sday [3]. Many molecular details remain vague,however, leaving questions unanswered and hypothe-ses unconfirmed. Here we discuss but one aspect ofcell migration: the physical origins of the forces thatdrive cell movements. Broader and more particularaspects of cell motility are discussed in several recentreviews and books [2,4–10].The first things that comes to mind when discussingforce generation are the motor proteins that convert thechemical energy of nucleotide hydrolysis into mechan-ical work [11,12]. Myosin motors drive contraction ofactin networks, and kinesin and dynein motors trans-port organelles and vesicles along microtubules. Thesemolecular motors have evolved mechanochemicalcycles that enable them to move at hundreds ofnanometers per second along their polymer tracks and,if stalled, generate forces in the range of piconewtons(pN). There are many other hydrolysis-driven proteinmotors that are not involved in cell motility (see, forexample, [13]). In this review, we shall focus on muchsimpler force generators that may have appearedearlier during evolution. Rather than using complexallosteric conformational changes, these motors rely onthe relatively simpler processes of polymerization anddepolymerization of dynamic biopolymers and on thegelation and solation of cytoskeletal gels. These motorsdo not operate in a cyclic fashion, undergoing a numberof steps that correspond to changes in conformationand in chemical state and eventually resetting them-selves to their initial configuration. Rather, they are‘one-shot’ engines that do their increment of work andthen are disassembled and reassembled at anotherplace in the cell.How Individual Filaments PushAll molecular motors work on the same generalprinciple: short-range molecular attractions capture‘favorable’ Brownian fluctuations (see Box 1 and[12–14]). How they accomplish this depends on theirprotein geometry, the diversity of which gives rise tothe wide variety of protein motors. Unlike most proteinmotors, the filament motors we discuss here have acomparatively simple geometry and so their principleof operation is fairly well understood.Hill and Kirschner [15] used thermodynamics todemonstrate that a polymerizing filament would notnecessarily stop growing when it runs into anobstacle. Rather, the resistance from the obstaclewould slow its growth, thus demonstrating that poly-merization can generate a force. Subsequently, actinpolymerization was observed to develop enough forceto deform lipid vesicles [16–18], suggesting that actinpolymerization may be responsible for pushing out thefront of motile cells [19].The thermodynamic approach had severe limitations,for it says nothing about the molecular mechanism, andtells only what force could be generated in the limit ofvery slow (‘quasi-static’) polymerization rates when thesystem is very close to equilibrium, which is not thecase in a live cell. A mechanistic theory was needed tocompute the force actually generated under cellularconditions. Macroscopically, the mechanism by whichpolymerization can generate force is not obvious, for itwould appear that, when a growing filament’s tip‘bumps’ into an object, the growth would simply ceasebecause there is no room between the tip and theobject for a monomer to squeeze in. The key differenceon the molecular scale is Brownian motion.Peskin et al. [20] formulated a mechanistic theory toaccount for the force generated by polymerizationwhen the polymers are rigid. They called this model the‘Brownian ratchet’ to distinguish it from the ‘ratchetand pawl’ model famously discussed by Feynman [21]to illustrate the thermodynamic impossibility of obtain-ing work from an isolated isothermal system. Becausethey are small, cells are isothermal, but they are notisolated, and they contain abundant energy sources.ReviewCurrent Biology, Vol. 13, R721–R733, September 16, 2003, ©2003 Elsevier Science Ltd. All rights reserved. DOI 10.1016/S0960-9822(03)00652-31Department of Mathematics and Center for Genetics andDevelopment, University of California, Davis, California 95616,USA. E-mail: [email protected] 2Departments ofMolecular & Cellular Biology and ESPM, University ofCalifornia, Berkeley, California 94720-3112, USA. E-mail: [email protected] convert thermal fluctuations into mechanical force in avariety of ways. However, two prototypical strategies are generallyreferred to as power strokes and Brownian ratchets [12,22,106]. In apower stroke, the binding reaction is mechanically coupled tomovement and generation of force. For example, if the chemicalreaction of a monomer binding to a filament tip triggered aconformational change in the monomer that elongated it, then such‘stroke’ would directly drive an object in front of the polymer tip. In aBrownian ratchet, the role of monomer binding reaction is to preventbackward fluctuations of the load, rather than to apply a mechanicalforce directly to it (Figure 1). That is, the load is driven by its ownBrownian fluctuations, and the chemistry provides the energy torectify its diffusive motion [11,12].This nomenclature is somewhat misleading, as the two mechanismsrepresent
View Full Document
Unlocking...