1: %listX.tex
2: \documentclass[aps,twocolumn,groupedaddress]{revtex4}
3: \usepackage{graphicx}
4: \begin{document}
5: \title{Continuum model of actin-mediated bacterial propulsion}
6: \author{V.G.~Benza}
7: \affiliation{Dipartimento di Fisica e Matematica, Universita' dell'Insubria,
8: Como, Italy}
9: \begin{abstract}
10: The mechanism of propulsion of host bacteria under the action
11: of actin gel networks
12: is examined by means of a continuum model of the dynamics of
13: the F-actin concentration. The model includes the elasticity
14: of the network, its attachment
15: to the host and the polymerization at the interface with it.
16: A formula for the cruise velocity is derived wherefrom the
17: respective contributions of elasticity and polymerization
18: are made explicit. The velocity tends to elongate the
19: gel in the direction opposite to motion, and in turn
20: is proportional to the size of its active portion.
21: It is shown that the motion can start only past
22: a finite latency time: the condition for the onset of motion
23: is explicitly given.It is numerically found that at steady
24: state the motion has a pulsating behavior, with sudden decelerations
25: and subsequent recoveries.
26: \end{abstract}
27: \maketitle
28: \vskip2pc
29: \section{Introduction}
30:
31: The mechanism of actin-induced propulsion has been
32: the object of increasingly sophisticated experiments
33: in recent years \cite{sootheriot,bernheimprostsykes}.
34: The phenomenon is of interest as a paradigm of more
35: complex situations, and exhibits a variety of fascinating
36: and intriguing features, which keep stimulating new research.
37: In short, an alien host
38: (e.g. a bacterium or a polystyrene sphere)
39: can make the cytoplasmic material to react by
40: polymerizing the resident G-actin into an F-actin gel network surrounding
41: the host's surface;
42: in due time the network evolves into an asymmetric shape,
43: typically comet-like.
44: The host eventually starts moving, and its cruise
45: can last for relatively long times at almost constant velocity.
46: We will focus here on two features of the process;
47: the first one is a threshold mechanism
48: at the onset of motion:
49: the velocity of the host rises from zero to its
50: cruise value quite abruptly, while
51: the mass of the gel undergoes a smooth growth.
52: The second feature is a correlation between the cruise velocity
53: and the length of the comet \cite{sootheriot}:longer tails
54: generate higher velocities.
55: This suggests that the dominant role in propulsion
56: stems from a bulk force, rather than from an interface
57: effect, such as the polymerization rate at
58: the host's surface.
59: In fact in cases of growth over convex surfaces
60: it has been observed and proved
61: that the gel elasticity has a primary role:
62: the stress
63: can inhibit the polymerization at the
64: interface \cite{noireauxsykes} and become
65: the main propulsive factor.
66: Disk-shaped
67: hosts
68: can as well be put into motion \cite{schwartzmcgrath};
69: one could argue that since the gel coating
70: entirely surrounds
71: the host, the convex portions of its surface
72: are in all cases under the action of stress.
73: It is difficult though to believe that this
74: is an exhaustive explanation
75: when both the cruise velocity and the elongation of the
76: comets are perpendicular
77: to the flat faces of the disks.
78: One must add to the picture
79: the links connecting the gel to the host \cite{marcysykes};
80: such links terminate on the surface at
81: nucleation points , whose distribution is determined by
82: the combined action of various activator groups \cite{activator}.
83: The oversimplified digest given above
84: actually translates into a rather
85: intricate dynamical problem,where
86: the global
87: elastic action of the gel,
88: the gel-host links and the polymerization
89: rate come together into play.
90: This problem is formulated here in terms of
91: a one-dimensional model where the system configuration
92: is described by the gel concentration.
93: The gel velocity
94: is self-consistently determined by generalizing
95: to the time-dependent case the arguments presented in Ref.
96: \cite{chaikinprost}.
97: The time evolution of the gel concentration is described
98: by a deterministic equation, which includes elasticity
99: and polymerization/depolymerization;
100: the approach being of mean-field type, stochasticity
101: is modeled by a diffusion term.
102: In our formulation
103: the interaction with the host turns
104: into the boundary
105: condition for the concentration:
106: the evolution at the boundary
107: has the form of the tethered ratchet
108: model of Mogilner and Olster \cite{mogilneroster}.
109: One of the main results of the paper
110: is an explicit formula for the velocity (see below Eq. \ref{V}),
111: wherefrom its correlation with the comet size
112: becomes transparent.
113: Furthermore, the origin of
114: the threshold
115: mechanism at the onset of motion is made clear,
116: as a natural outcome of the model.
117: The numerical results
118: shown here
119: exhibit this threshold mechanism as well as
120: an intermittent behavior of the velocity at steady
121: state.
122: The system starts moving past
123: a given ``latency'' time, which depends on the rate
124: of accretion of the gel, the elasticity and the
125: nucleation rate.
126: The interface polymerization rate can have some relevance
127: in the initial stages after start, but at steady
128: state
129: the bulk contribution prevails: the resulting motion
130: has a pulsating behavior,
131: with sudden reductions of the velocity and consequent recoveries,
132: as in a system alternating ``loading'' phases at low speed
133: with ``unloading'' phases at higher speed.
134: (see below Fig.~\ref{fig1}).
135: \par
136: \section{Model}
137:
138: As anticipated, we are going to establish a
139: one-dimensional minimal model
140: for the dynamics of the actin gel network
141: during its propulsive action over some host object.
142: We take a macroscopic point of view, where
143: the concentration and the local velocity are
144: a natural choice as dynamical variables describing the gel.
145: Under fairly reasonable assumptions
146: one can get rid of the space dependence of the velocity.
147: The resulting ``average'' velocity is determined
148: by the concentration: hence
149: the system is described by this single variable.
150: In turn
151: the average velocity, which we identify with
152: the velocity of the host, appears in the evolution
153: equation for the concentration (see below Eq.\ref{EE}).
154: Let us first show how the host velocity is determined.
155: When the concentration $c(x,t)$ has a well defined peak,
156: say at $x_{max}$, one can approximate the average of the velocity
157: field $v(x,t)$ as:
158: $<v(t)> \approx v(x_{max},t)$.
159: This obvious remark is not a real progress,
160: unless one relates the unknown quantity $v(x_{max},t)$
161: with the value of the velocity field at some other point
162: $x_{0}$ where it can in principle be determined:
163: such a point is the contact
164: of the gel with the host \cite{chaikinprost}.
165: Indeed on the basis of the standard rate theories \cite{arrhenius}
166: the velocity at the interface is a function
167: of the elementary work $\Delta W$ according with the formula:
168:
169: \begin{equation}
170: \label{AR}
171: v_{A}=v_{+} \exp({- {{\Delta W} \over {k_{B}T}}}) - v_{-}
172: \end{equation}
173: where $v_{+},\, v_{-}$ are the free polimerization/depolymerization
174: velocities.
175: One has then:$\,v(x_{0},t)=v_{A}(t)$; the link between $v(x_{0},t)$
176: and the unknown value $v(x_{max},t)$
177: is provided by
178: the law of mass conservation.
179: From now on, unless otherwise stated, we are going to use
180: a reference frame comoving with the host, say
181: at velocity $v$; the mass conservation reads:
182: \begin{equation}
183: \label{M}
184: {d \over dt} c(x,t) + {d \over dx}[(v(x,t) -v)c(x,t)]=0.
185: \end{equation}
186: %G.B.:mettere derivate parziali
187: The role of symmetry breaking in the
188: onset of motion is not examined here;
189: let us then assume, e.g., that the gel has grown to the left of the host,
190: so that when the host moves $x_{0}$ drifts rightwards.
191: We integrate Eq.\ref{M} over the interval
192: $x_{b} \le x \le x_{0}$, where $x_{b}$ is a generic position
193: in the gel, and obtain:
194: \begin{equation}
195: \label{REL}
196: \int_{x_{b}}^{x_{0}} dx {d \over dt} c(x,t)+[v(x_{0},t)-v]c(x_{0},t)=
197: [v(x_{b},t)-v]c(x_{b},t)
198: \end{equation}
199: Let us finally
200: assume that the gel and the host drift rigidly together,
201: i.e. that $v \equiv v(t)=v(x_{max},t)$;
202: upon chosing in Eq. \ref{REL} the generic position
203: at the peak of the concentration $(x_{b}=x_{max})$ we
204: get the relation we were searching for:
205: \begin{equation}
206: \label{VV0}
207: v(t)=v(x_{0},t)+{1 \over c(x_{0},t)} \int_{x_{max}}^{x_{0}} dx {d \over dt}c(x,t)
208: \end{equation}
209: This formula
210: gives the velocity $v(t)$ as a nonlocal
211: function of the concentration: the evolution equation
212: for $c(x,t)$ will make our treatment self-consistent.
213: Let us now discuss how such an equation can be
214: determined.
215: The dynamical response of a
216: normal
217: gel, where the structure of the network is fixed
218: by the distribution of crosslinks and of
219: flexible chains,
220: is characterized by
221: well defined values of the coupling constants, such as, e.g.,
222: the Young modulus in the case of elasticity.
223: A network of semiflexible, rather than
224: flexible, chains is apparently more apt to describe the
225: actin filaments; whenever its structure is fixed
226: its elastic response is also defined, but differs
227: from that of a normal gel \cite{mackintosh}.
228: Apart from this difference one must be aware of the fact that
229: the actin network of the comets has no predetermined structure:
230: its couplings evolve with the configuration,
231: e.g. the elastic response
232: changes as the mass and crosslink densities
233: build up or decay: it would be appropriate
234: to consider both populations.
235: In the present ``minimal'' context, where we limit ourselves
236: to the mass density, the compression modulus
237: depends on $c(x,t)$ and
238: the role of crosslinks is included in an effective coupling constant
239: $C$; we refer to the Section Elastic Energy for the
240: explicit form of the elastic term.
241:
242: The environmental factors,
243: such as the "feeding" of polymerization,
244: the depolymerization, the intrinsic fluctuations of
245: the concentration at the microscopic scale, can be
246: assumed to be proportional to the local concentration.
247: Sure enough, the environment evolves with
248: the structure as well, but we will disregard its
249: time evolution for the time being.
250: We consider
251: a potential $U(x)$ (a growth rate),describing the
252: "feeding" and "decaying" processes, in dependence with its sign.
253: Furthermore, a diffusive term accounts
254: for the small scale disorder of the environment,
255: which perturbs a purely deterministic buildup
256: of the network.
257: We add a quadratic term describing the short range
258: self-repulsion between different portions of the structure.
259:
260:
261: Under such factors, in a static reference frame
262: the concentration would then evolve according to
263: the equation:
264: \begin{equation}
265: \label{POP}
266: {d \over dt} c = U(x-vt)c -b c^{2} +D {d^{2}\over {d^{2}x}} c
267: \end{equation}
268: where the time dependent potential accounts
269: for the motion of the "activators" which are
270: are predominantly concentrated around
271: the host.
272: If, e.g., the potential has a constant value
273: $U>0.$ in the feeding region, its action combined with the
274: repulsive term $-b c^{2}$ drives $c(x,t)$ towards
275: the value $\bar c = U/b$.
276: The host enforces the condition
277: $c(x,t)=0.\, (x>x_{0})$;
278: furthermore, the gel is attached
279: to it at the nucleation
280: points and grows against it when the actin filaments polymerize.
281: We model this along the lines
282: proposed by Mogilner and Oster \cite{mogilneroster},
283: i.e. by means of two concentrations $a(t),w(t)$
284: respectively describing the "attached" and"working" portions
285: of the gel. Their evolution determines the
286: boundary condition of $c$ at the interface:$c(x_{0},t)=a(t)+w(t)$.
287: We now go over and write the evolution equation
288: including the elastic term; this term
289: is proportional to the deviation of the concentration
290: with respect to a reference "background" configuration \cite{brgel}, and
291: is opposite to it.
292: If, e .g., an excess of mass has accumulated in some region
293: during the process of buildup of the network, the elastic force
294: acts as in a spring compressed in the $x$ direction,
295: and elongates the excess of mass along $x$ (see the Section
296: Elastic Energy).
297: The frame comoving with the host has coordinates
298: $x'=x-vt,t'=t$; in these variables the evolution equation
299: has the following form:
300: \begin{eqnarray}
301: \label{EE}
302: {d \over dt}c(x,t)& =& D {d^{2} \over {d^{2} x}} c(x,t) +v {d \over dx}c(x,t)\nonumber\\
303: + U(x) c(x,t) &-&b c^{2}(x,t) -\Gamma {{\delta E_{el}} \over {\delta c(x,t)}}
304: \end{eqnarray}
305: where we have dropped the primes;
306: $E_{el}$ is the elastic energy and
307: $\Gamma$ is the relative decay rate.
308: Notice that a drift term has appeared, in going to the
309: moving frame: it
310: has the effect of
311: elongating the profile of
312: $c(x,t)$ in the direction opposite to $v$, as if
313: the concentration experienced a backward wind.
314: The higher the velocity, the stronger the effect:
315: the elongation of the profile "measures" the
316: velocity.
317: The velocity and
318: the elongation of the profile actually have a link which goes beyond
319: this purely hydrodynamic effect.
320: In fact, according with Eq.\ref{VV0}
321: the velocity $v$ is in turn $c$-dependent, and
322: upon inserting the l.h.s. of Eq.\ref{EE} into Eq.\ref{VV0} we finally get:
323: \begin{eqnarray}
324: \label{V}
325: v={{v(x_{0}) c(x_{0})}\over {c(x_{max})}}&+& {1\over c(x_{max})} \cdot \lbrace D ({{ dc}\over{dx}})_{x_{0}}+ \nonumber\\
326: \int_{x_{max}}^{x_{0}}dy[U(y)c(y)&-& b c^{2}(y)-\Gamma{{\delta E_{el}}\over{\delta c(y)}}]\rbrace\\
327: \nonumber
328: \end{eqnarray}
329: where $x_{max}<x_{0}$ and the time dependence of the arguments is understood.
330: In the time independent case only the first term survives,
331: and the formula simply states the conservation of stationary
332: fluxes \cite{chaikinprost}; the boundary term $D({{dc} \over {dx}})_{x_{0}}$
333: is negative, as one easily verifies: it averages
334: over the region $x_{max}<x<x_{0}$
335: the loss in propulsion originating from network's disorder.
336:
337: \section{Numerical Results}
338: The formula \ref{V} is one of the primary results of the paper;
339: it shows how
340: the velocity gets contributions
341: both from the interface and from the bulk.
342: Notice that $v$ goes as the inverse of the maximal concentration;
343: the elastic force must counter the formation
344: of exceedingly high concentration peaks
345: in order to have higher velocities.
346: The bulk contribution becomes increasingly relevant
347: as the peak at $x_{max}$ moves
348: off $x_{0}$ in the backward direction.
349: In the feeding region, close to the interface,
350: we have $U(x)=U>0.$: the network grows and contributes
351: positively to the propulsion.
352: The diffusion constant D $(D=5. 10^{-3}(\mu m)^{2}/sec)$ has a minor effect:
353: its value has been chosen two orders of magnitude smaller
354: than the diffusion constant of monomers within the gel
355: estimated in Ref.\cite{softlisteria}.
356: The elastic and self-repulsion terms
357: operate positively or negatively dependending on
358: the configuration. They give rise to
359: an intermittent behavior of the velocity even at steady state;
360: typically the velocity 'jumps' between two (or more) values,
361: staying most of the time in one of them, which can
362: be identified with the cruise velocity of the host.
363: Among the temporary, intermittent velocities
364: we found negative values as well: this happens, e.g.,
365: when the rate
366: $\Gamma$ is half the value $\Gamma =2.\, (pN \cdot sec)^{-1} (nm)^{-2}$
367: used in the simulation presented here.
368: At steady state the motion can be depicted as a sequence of very
369: short ``charging''steps, where the velocity
370: is small or negative, followed by longer
371: advancement steps.
372: A typical behavior is shown in Fig.~\ref{fig1},where $U=1. (sec)^{-1}$:
373: the cruise velocity is $v \approx 8.\, nm/sec$,
374: and on average $x_{0}-x_{max} =250.\,nm$.
375: The velocity is
376: quite near to $10. nm/sec$, established
377: in Ref.\cite{sootheriot} as a minimum value for
378: classifying an object as being in motion.
379: At a closer inspection
380: the concentration profile reveals time dependent
381: small size
382: deformations tuned with the steps of the velocity. In particular
383: the peak position $x_{max}$ evolves with an intermittent behavior:
384: temporary compressions are subsequently released.
385: \begin{figure}
386: \includegraphics[width=3.0in]{./fig1.eps}
387: \caption{Typical behavior of the velocity (in units of
388: 10.\, nm/sec) versus time (in seconds); notice
389: the discontinuity at the onset of motion, the transient
390: before reaching steady state and the intermittent behavior at steady state.
391: The values of the parameters are reported in the text}
392: \label{fig1}
393: \end{figure}
394: Let us call active region the portion of space
395: going from the peak to the interface; in the example
396: shown in Fig.\ref{fig1} this region goes slightly beyond
397: the feeding region, which has size $200.\,nm$ (see Fig.\ref{fig3}).
398: The formula for the velocity Eq. \ref{V} shows that the bulk contribution
399: is proportional to the size of the active region.
400: A linear relation between the length of the comet
401: and the cruise velocity has been observed in
402: very accurate kinetics experiments by Soo and Theriot \cite{sootheriot}:
403: we believe that our results are consistent with
404: such observations.
405: Actually there are two concomitant effects:
406: the drift term elongates
407: the profile by a factor proportional to $v$,
408: in turn a more elongated profile generates a larger $v$.
409: Let us examine the interface contribution.
410: According to Eq. \ref{AR}, $v(x_{0})$
411: is a decaying function of the elementary work $\Delta W$:
412: \begin{eqnarray}
413: \label{W}
414: \Delta W &=& \delta l \cdot F(x_{0})\equiv \delta l \cdot \sigma_{x,x}(x_{0})\nonumber\\
415: &=&\delta l \cdot c(x_{0}){a\over w}C_{a}\int_{x_{max}}^{x_{0}} dy [c(y)-c(x_{0})]
416: \end{eqnarray}
417: where $\delta l$ is the elementary displacement
418: having the order of the size of a monomer $\delta l=2.2\, nm$,
419: $ {a \over w}\,\,$ the ratio of attached
420: versus working gel and $C_{a}=2.\,(pN \cdot nm)$ the elastic coupling
421: of the attached portion of the gel.
422: The above expression follows from the equilibrium condition
423: between working forces and load \cite{mogilneroster},
424: the load here being uniquely given by the attached links
425: (see the Section Elastic Energy).
426: Let us examine the order of magnitude of the force
427: at steady state in the case reported in Fig. \ref{fig1}.
428: There the ratio $a/w$ is close to
429: one (see Fig. \ref{fig4}), the active region
430: has size $x_{0}-x_{max}=250.\, nm$, the peak
431: is $c(x_{max})\approx 10.\,$ (the concentration has units $(10.\, nm)^{-1}$)
432: the deviation $c(x_{max})-c(x_{0})$ is of order
433: $10^{-1}$ so that one has $F(x_{0})\approx 5.\, pN$.
434: Taking into account that the Boltzmann factor is
435: $k_{B}T \approx 4.1 pN \cdot nm$ one can verify that
436: the contribution to $v$ arising from $v(x_{0})$ is quite
437: small as compared with the bulk contribution.
438:
439: Assuming a free polymerization
440: velocity $v_{+}=10.\, nm/sec$ and $v_{-}=0.$ we always obtained that
441: the bulk velocity
442: has the dominant role in propulsion, with the exception
443: of the initial transient following the onset of the motion.
444: Our simulations start with zero concentration.
445: A given nucleation rate drives the attached gel $(a(t))$,
446: its decay activates $w(t)$ \cite{mogilneroster}.
447: As $a(t)$ and $w(t)$ start increasing,
448: the boundary condition $c(x_{0},t)=a(t)+w(t)$
449: initiates the evolution of $c(x,t)$.
450: The velocity does not follow smoothly
451: the gel growth,
452: but rather jumps abruptly from zero to a finite value:
453: the onset of motion has a threshold.
454: The motion can start
455: provided that the gel generates a positive propulsion:
456: this happens when the maximal concentration
457: rises above $c(x_{0},t)$: at this point an excess
458: of mass has been created on the back of the host,
459: and the elastic response to it translates into motion.
460: If the gel growth rate is too small,
461: the maximal concentration stays locked at the interface,
462: and the motion cannot start.
463:
464: \begin{figure}
465: \includegraphics[width=3.0in]{./fig2.eps}
466: \caption{Latency time (in secs) versus the
467: growth rate $U$ (in units $(sec)^{-1}$);the size of the feeding region
468: is $x_{f}=200.\,nm$; notice that when $U < \bar U \approx 0.6$
469: the motion cannot start;
470: these results where obtained with $\Gamma =0.1 \,(nm)^{-2}\cdot (pN sec)^{-1}$,
471: $C_{a}=1.\,pN \cdot nm$, $C_{w}=0.5\, pN \cdot nm$; all the other parameters
472: have values as reported in the text}
473: \label{fig2}
474: \end{figure}
475: If the growth rate is sufficiently high the maximum of $c$ stays at $x_{0}$,
476: but only for a finite ``latency'' time.
477: This time is strongly dependent on
478: the rate $U$ in the feeding region; in Fig.~\ref{fig2}
479: it is shown that it diverges as $U$ decreases towards a
480: finite value $\bar U \approx 0.6$.
481: Similarly, one can verify that shorter feeding regions generate longer
482: latency times.
483: As the peak unlocks from $x_{0}$, the host is instantly put
484: into motion, and its velocity
485: $v$ abruptly departs from zero, with a value
486: depending on the mass accumulated during the
487: latency.
488: In the evolution equation Eq. \ref{EE} the drift term is
489: simultaneously activated: the profile
490: is faced with an abrupt backward wind, and
491: must adapt its shape to it.
492: \begin{figure}
493: \includegraphics[width=3.0in]{./fig3.eps}
494: \caption{From bottom to top: potential U(x); profile c(x) at t=2.sec,
495: right before
496: the onset of motion (dotted line);
497: c(x) at steady state (t=15.sec)(continuous line).
498: The abscissa has space unit $\bar a=10.nm$. The vertical axis is in
499: units $(sec)^{-1}$
500: (U(x)), and $(10.\, nm)^{-1}$ (c(x))}
501: \label{fig3}
502: \end{figure}
503: A transient follows, during which
504: the active region modifies its size with strong
505: fluctuations, until the system
506: reaches a steady state profile.
507: In Fig. ~\ref{fig3} we compare the profile during
508: latency, but close to start, with the steady state
509: profile; the potential is superimposed [$U(x)=1.$ in the
510: feeding region, $U(x)=-0.2$ in the decay region].
511: \begin{figure}
512: \includegraphics[width=3.0in]{./fig4.eps}
513: \caption{Time behavior of the variables a (continuous line) and w
514: (dotted line), representing
515: the attached and working part of the gel at the interface;
516: same parameters as in Fig. \ref{fig1};the vertical axis has units
517: $(10.\, nm)^{-1}$, the time unit is $1.\,sec$. At t=1.
518: $ w$ is still sensibly smaller than $a$. Notice that at some
519: time
520: $t\approx 2.$ $w$ grows above $a$: this event is close to the
521: onset of motion }
522: \label{fig4}
523: \end{figure}
524: The steady state motion results from
525: the combined action of the drift
526: term and of the elastic forces; the
527: drift term in particular marks the shape of
528: the steady state profile, which appears smoothly elongated away
529: from the interface as compared with its form at the
530: onset of motion.
531: A stationary transfer of mass from the feeding region
532: to the queue maintains the steady state profile.
533:
534:
535: \section{Elastic energy}
536: We write the elastic energy as a quadratic form
537: \begin{equation}
538: \label{EL}
539: E_{el}= {C \over 2}\int dx \int dy \,\delta c(x,t)\cdot G(x,y)\cdot\delta c(y,t)
540: \end{equation}
541: where $\delta c(x,t)= c(x,t) -\bar c(x,t)$ is the deviation
542: from a "background" configuration $\bar c(x,t)$, $C$
543: is an energy whose value must account of
544: the effect of crosslinks.
545: The kernel $G(x,y)$ has the role of bounding
546: the range of elastic response within a finite
547: length $\lambda$:
548: $G(x,y)=\exp({-|x-y| \over \lambda})$.
549: We define background or reference configuration
550: a configuration which is in static
551: equilibrium but only over a suitable time scale.
552: New configurations generated at shorter time scales
553: are attracted towards it.
554: In the present paper we give a simplified treatment
555: where the elastic interaction is limited to
556: the active region: we believe that this region
557: has the dominant role in the propulsion.
558: In the bulk of the network the elasticity
559: operates against changes in the concentration profile.
560: The background configuration results from the past
561: hystory of the structure.
562: For the time being we disregard memory effects and
563: simply mimic the rigidity of the system with the
564: self-repulsion term included in Equation \ref{EE};
565: the relative coupling constant has value $b= 0.5\, nm/sec$.
566: Let us focus now on the active region:
567: there a natural choice for the background configuration
568: is given by the concentration at the interface.
569: This is because at the host's surface the concentration
570: is fixed by the boundary condition
571: $c(x_{0},t)=a(t)+w(t)$, so that in the adjacent region
572: the profile must be predominantly attracted towards
573: $c(x_{0},t)$.
574: In specializing the energy Eq. \ref{EL} to the active region
575: we let then $\bar c(x,t)=c(x_{0},t)$;it is
576: understood that the integrations are
577: extended over the interval $x_{max}<x<x_{0}$,
578: and we assume $\lambda>>(x_{0}-x_{max})$.
579: About the form of the coupling $C$,
580: one must consider
581: that the attached and working portions operate
582: in opposite ways and with different strength.
583: The variable $a(t)$
584: is related with links connecting the gel to the
585: host: their role is of the same nature as that of the
586: gel-gel crosslinks, which determine the bulk's elastic
587: response.
588: Hence $a(t)$ attracts the system towards the background
589: configuration while $w(t)$ acts in the opposite direction.
590: The specific form of the coupling $C$ must then be:
591: $C= {a\over c(x_{0})} C_{a} - {w \over c(x_{0})} C_{w}$
592: where with $C_{a}$ and $C_{w}$ we have separated the
593: elastic response of the
594: attached links from that of the working filaments
595: (we always assume that $C_{a} > C_{w}$; in particular in
596: the simulations presented here we have $C_{a}=2. C_{w};
597: C_{w}=1.\,pN \cdot nm)$).
598: As we assumed that the length $\lambda$ is much larger than the
599: active region, we average over the space dependence of the
600: deviation:
601: $\delta c(x,t) \approx \delta c \equiv {1\over2}( c(x_{max},t)-c(x_{0},t))$.
602: The total elastic force, proportional to $\delta c$,
603: stabilizes the gel provided that the contribution from
604: the gel-host links is the dominant one, i.e. under the condition
605: $C_{a}\cdot a -C_{w}\cdot w > 0.$
606: Let us add some comment on the dynamics of $a(t)$ and $w(t)$:
607: their evolution equations
608: are slightly modified with respect to equations (1) and (2)
609: of Ref. \cite{mogilneroster}. Here $w(t)$ gets a contribution not only
610: from the detached links, but also from the
611: adjacent gel $(\bar c\equiv c(\bar x);\,\, (x_{0}- \bar x)<<1)$:
612: \begin{eqnarray}
613: \label{MO}
614: \dot a &=& j - \delta(v) a\nonumber\\
615: \dot w &=& \delta(v) a -d_{c} w +D (\bar c -2\,w)
616: \end{eqnarray}
617: here $\delta(v)=\delta_{0}(1.+ {v \over {\bar v}})$
618: is the velocity-dependent detachment rate, $d_{c}$
619: is the capping rate and $j$ is the nucleation rate
620: per space unit.
621: We assigned to these variables the following
622: values:$j=0.5\, (nm \cdot sec)^{-1}$ $\delta_{0}=1.\, (sec)^{-1}$
623: $\bar v = 100.\, nm/sec$, $d_{c}=0.8\, (sec)^{-1}$.
624: It is necessary to connect this
625: formalism with the usual treatment of
626: elasticity: this is done by assuming
627: that the system is incompressible, i.e. that the elementary mass
628: contained in a generic interval $\Delta x$ is conserved
629: under deformation: $\delta (c(x) \Delta x)=0.$
630: The condition above gives:
631: \begin{equation}
632: \delta c(x) = - c(x) {d \over dx} \Delta x = - c(x) u_{x,x}(x)
633: \end{equation}
634: where $u_{x,x}(x)$ is the strain tensor in one dimension.
635: Eq. \ref{EL} then reads:
636: \begin{equation}
637: E_{el}= {C\over 2} \int dx \int dy [c(x) u_{x,x}(x)]G(x,y) [c(y) u_{y,y}(y)]
638: \end{equation}
639: this should be compared with the standard form of the elastic
640: energy:
641: \begin{equation}
642: E_{el}=B \int dx (u_{x,x}(x))^{2}
643: \end{equation}
644:
645: The compression modulus $B$, which in one dimension
646: is an
647: energy/length, in our case is explicitly concentration-dependent.
648: In fact,
649: the stress tensor $\sigma_{x,x}(x)$, which here has the dimension
650: of a force, is given by
651: \begin{eqnarray}
652: \sigma_{x,x}(x)&=& {{\delta E_{el}}\over {\delta u_{x,x}(x)}}\nonumber\\
653: &=&
654: c(x){{\delta E_{el}}\over{\delta c(x)}}\approx C \lambda_{eff} c^{2}(x) u_{x,x}(x)
655: \end{eqnarray}
656: where $\lambda_{eff}$ is the effective range of the interaction.
657: We have then:
658: $B \to C \lambda_{eff} c^{2}(x)$.
659: In the case of Fig.\ref{fig3} one would have
660: $B \approx C_{a}\cdot (x_{0}-x_{max})\cdot c^{2}(x_{max})\approx 0.5\, nN.$
661: Correspondingly the elastic stiffness can be estimated as
662: the ratio $k_{el}=B/(x_{0}-x_{max})$:
663: we obtain then $k_{el}= 10^{-3}\, nN/nm$, quite small as compared
664: with the value $k_{el}= 0.17\, nN/nm$,
665: reported in Ref. \cite{theriotelastic}.
666:
667: The value of $C_{a}$ was
668: chosen by requiring a
669: $\Delta W$ of order $k_{B}T$.
670: In the equation of motion the elastic coupling $C$
671: multiplies $\Gamma$, whose value was chosen in order
672: to make the elastic term strong enough to counter
673: an unlimited growth in the feeding region.
674:
675: From Eq.\ref{W} one can argue that a steady state with a ratio
676: $a/w << 1.$ would allow for
677: higher values of $C_{a}$ and consequently would give more
678: realistic values of stiffness;
679: the rate $\Gamma$ should be accordingly reduced, in order
680: to keep the same strength of the elastic term in the equation of motion.
681:
682:
683: \section{Conclusions}
684:
685: We presented a continuum model of actin-mediated
686: bacterial propagation.The model is a minimal one, in that
687: it describes the system in terms of a single variable:
688: the concentration $c(x,t)$.
689: The time evolution of $c(x,t)$ includes
690: the interaction with the environment and with
691: the host as well as the elastic
692: interaction in the region adjacent to the host's
693: surface.
694: A formula for the velocity is obtained,
695: where the interface and bulk contributions are made explicit.
696: This formula, together with the hydrodynamic effect
697: acting on the profile, makes the correlation
698: between the velocity and the length of the actin network explicit:
699: this result agrees with the behavior observed in Ref. \cite{sootheriot},
700: and indicates that the bulk, rather than the interface, has the major role
701: in propulsion.
702: The model furthermore motivates on theoretical grounds the existence
703: of an intrinsic threshold mechanism underlying the onset of the motion.
704: A relevant feature of a minimal model is that in the
705: numerical simulations
706: the number of adjustable parameters
707: is accordingly quite small. The key parameters
708: regulating the dynamics of the concentration in the bulk
709: are the growth rate $U$,the self-repulsion coupling $b$, the elastic
710: coupling $C$ and the rate $\Gamma$; the interface
711: dynamics is controlled by the nucleation, detachment and
712: capping rates $(j,\,\delta(v),\,d_{c})$.
713: Preliminary numerical results exhibit the threshold
714: mechanism
715: and its dependence on the growth rate (see Figs.\ref{fig1}, \ref{fig2}).
716: The ``cruising'' steady state is characterized by a
717: pulsating behavior, where the motion undergoes sudden
718: reductions and subsequent recoveries.\\
719:
720:
721:
722: Acknowledgements.
723:
724:
725: The Author wishes to thank M.Cosentino Lagomarsino an B.Bassetti,
726: who got him acquainted with comets and listeria and encouraged the
727: accomplishment of this work.
728:
729:
730: \begin{thebibliography}{1}
731:
732: \bibitem{sootheriot}
733: Soo F S and Theriot J A 2005 Large-scale quantitative analysis of sources
734: of variation in the actin polymerization-based movement of
735: {\it Listeria monocytogenes}\,
736: {\it Biophys.J.} {\bf 89} 703-723
737:
738: \bibitem{bernheimprostsykes}
739: Bernheim-Groswasser A, Prost J and Sykes C 2005
740: Mechanism of actin-based motility: a dynamic state diagram
741: {\it Biophys. J.} {\bf 89} 1411-1419
742:
743:
744: \bibitem{noireauxsykes}
745: Noireaux V , Golsteyn R M, Friederich E,Prost J,
746: Antony C, Louvard D and Sykes C 2000
747: Growing an actin gel on spherical surfaces {\it Biophys. J.} {\bf 278} 1643-1654\\
748:
749:
750: \bibitem{schwartzmcgrath}
751: Schwartz I M,Ehrenberg M,Bindschadler M and McGrath J L 2004
752: The role of substrate curvature in actin-based pushing forces
753: {\it Curr.Biol.} {\bf 14} 1094-1098 \\
754:
755:
756: \bibitem{marcysykes}
757: Marcy Y,Prost J, Carlier M F, Sykes C 2004 Forces generated during
758: actin-based propulsion: a direct measurement by micromanipulation
759: {\it Proc.Natl.Acad.Sci.USA} {\bf101} 5992-5997 \\
760:
761:
762: \bibitem{activator}
763: Welch M D, Rosenblatt J, Skoble J, Portnoy D A and
764: Mitchison T J 1998 Interaction of human Arp2/3 complex
765: and the Listeria monocytogenes ActA protein in actin filament nucleation
766: {\it Science} {\bf 281} 105-108 \\
767: Cameron L A ,Svitkina T M,Vignjevic D,Theriot J A and
768: Borisy G G 2001 Dendritic organization of actin comet tails
769: {\it Curr.Biol.} {\bf 11} 130-135\\
770:
771: \bibitem{chaikinprost}
772: Gerbal F,Chaikin P,Rabin Y and Prost J 2000
773: An elastic analysis of {it Listeria monocytogenes} propulsion
774: {\it Biophys. J.} {\bf 79} 2259-2275\\
775:
776:
777: \bibitem{mogilneroster}
778: Mogilner A and Oster G 2003 Force generation by actin polymerization II:
779: the elastic ratchet and tethered filaments {\it Biophys. J.} {\bf 84}, 1591-1605\\
780:
781:
782: \bibitem{arrhenius}
783: Hill T L 1987 {\it Linear aggregation theory in cell biology} (New York: Springer-Verlag)\\
784:
785: \bibitem{mackintosh}
786: Head D A ,Levine A J and MacKintosh F C 2003
787: Deformation of cross-linked semiflexible polymer networks
788: {\it Phys.Rev.Lett.} {\bf 91} 108102-4\\
789:
790: \bibitem{brgel}
791: Read D J,Brereton M G and Mcleish T C B 1995 Theory of the order-disorder
792: phase transition in cross-linked polymer blends
793: {\it J.Phys.II} France {\bf 5} 1679-1705\\
794:
795: \bibitem{softlisteria}
796: Boukellal H,Campa\'s O,Joanny J F,Prost J and Sykes C 2004
797: Soft Listeria:actin-based propulsion of liquid drops
798: {\it Phys.Rev.} {\bf E69} 061906-1 \\
799:
800: \bibitem{theriotelastic}
801: Parekh S H,Chaudhuri O,Theriot J A and Fletcher D A 2005
802: Loading history determines the velocity of actin-network growth
803: {\it Nature Cell Biology} {\bf 7} 1219-1223\\
804: %in dell:/gel/theriot05.pdf
805:
806: \end{thebibliography}
807: \end{document}