q-bio0501024/Nandy.tex
1: 
2: \documentclass[aps,preprint,pre,array,epsfig,eqsecnum]{revtex4}
3: 
4: \usepackage{amsmath}
5: \usepackage{amsfonts}
6: \usepackage[dvips]{graphicx}
7: \usepackage{epsfig}
8: \usepackage[]{subfigure}
9: \usepackage[dvips]{color}
10: 
11: 
12: \addtolength{\hoffset}{-1cm}
13: \addtolength{\textwidth}{2cm}
14: 
15: %% \topmargin 0.0cm
16: %%\textheight 21.5cm
17: %%\oddsidemargin 0.5cm
18: %%\textwidth 15cm
19: 
20: 
21:  
22: %_____________________________________________________________________
23: 
24: 
25: \begin{document}   
26: \setlength{\parindent}{0pt}
27: 
28: 
29: \title{Competing Polymerization of Actin Skeleton explains Relation between
30: Network Polarity and Cell Movements}
31: 
32: 
33: \author{B. Nandy and A. Baumgaertner}
34: 
35: \affiliation{
36: Institut f\"ur Festk\"orperforschung, 
37: Forschungszentrum J\"ulich, 52425 J\"ulich, Germany}
38: 
39: 
40: \date{\today}
41:  
42: 
43: \begin{abstract} 
44: 
45: Based on experimental observations it is known that various biological
46: cells exhibit a persistent random walk during migration on flat substrates.
47: The persistent random walk is characterized by `stop-and-go' movements :
48: unidirectional motions over distances of the
49: order of several cell diameter are separated by localized short time
50: erratic movements.
51: Using computer simulations the reasons for this phenomena had been unveiled and
52: shown to be attributed to two antagonistic nucleation processes during
53: the polymerization of the cell's actin cytoskeleton : the (ordinary) spontaneous
54: nucleation and the dendritic nucleation processes.
55: Whereas spontaneous nucleations generate actin filaments growing in different 
56: directions and hence create motions in random directions, 
57: dendritic nucleations provide a unidirectional growth.
58: Since dendritic growth exhibits stochastic fluctuations,
59: spontaneous nucleation may eventually compete or even dominate, 
60: which results in a reorientation of filament growth and hence 
61: a new direction of cell motion.
62: The event of reorientation takes place at instants of vanishing
63: polarity of the actin skeleton.
64: 
65: \end{abstract}
66: 
67: \pacs{87.17, 05.40, 82.35}
68:       
69: 
70: \maketitle
71: 
72: %__________________________________________________________________________
73: \section{INTRODUCTION}
74: 
75: Most animal cell types are `motile' and 
76: possess the capacity to move over or through 
77: a substrate, and cell migration is a normal occurrence in both normal
78: physiological cases as in the case of diseases.
79: %
80: Unlike the phenomena related to `mobility', where the collision among molecules
81: determine their random movements, the random motion of motile cells
82: requires an energy-consuming mechanism. 
83: %
84: It is now widely accepted that the basic engine for gliding or 
85: crawling locomotion of many living cells is the ATP-supported 
86: polymerization of the actin cytoskeleton.
87: %
88: Rapidly moving cells 
89: can often move over the substrate without microtubules (keratocytes,
90: neutrophils), but crawling motility always requires actin. 
91: 
92: One of the crucial 
93: factors for cell movement is actin polarity \cite{Weiner02}, 
94: which correlates with a 
95: persistent random walk during migration.
96: The persistent random walk is characterized by `stop-and-go' movements :
97: unidirectional motions over distances of the
98: order of several cell diameters are separated by localized short time
99: erratic movements.
100: The main focus of the present study 
101: is the cause behind the persistent random walk of a cell, and hence the
102: cause behind the correlated spontaneous change of polarity.
103: 
104: 
105: It has been seen that cell motility and chemotactic 
106: migration are related to specific rearrangements of the actin cytoskeleton.
107: % 
108: Actin 
109: polymerizes into new filaments in regions of the cell that are specified by 
110: activated signalling at the plasma membrane
111: \cite{Bray92,Lauffenburger96,PollardBorisy03,Carlier03}. These signals give
112: rise to cell polarity and directional motility. In the absence
113: of chemotactic stimuli, cells exhibit a persistent random walk.
114: 
115: In this work we present results of our investigations on the
116: relation between cytoskeletal actin polymerization and the 
117: persistence random walk of a cell.
118: The results are based on the analysis of
119: Monte Carlo simulations of a simple model cell. 
120: 
121: %_________________________________________________________________________
122: \section{MODEL AND SIMULATION TECHNIQUE}
123: 
124: Due to the limitation of information and computational complexity it 
125: is not possible to construct the model cell and it's motility including all 
126: regulatory proteins with their exact concentration for all 
127: types of cells. So we focus on one of the simplest cell which is the 
128: keratocyte residing in the epidermis, the most outer layer of the skin.
129: The keratocyte is responsible for the formation of tissue and wound 
130: healing, both requiring the cell's ability to autonomously migrate 
131: within skin tissue.
132: It is in fact one of the fastest moving cells \cite{Alberts98}
133: with a speed of 0.5 $\mu$m/s. The advantage of choosing keratocyte is that
134: the crescent-like shape of the cell 
135: is almost constant as it moves \cite{Mitchison91,Lee93}.
136: The motility of the keratocyte is assumed to be based largely on a
137: continuously remodeling actin network. 
138: In our model we neglect the cell body, i.e., the
139: nucleus and other organelles. This cell type is known as a cytoplast
140: \cite{Borisy99a}.
141: 
142: 
143: 
144: \noindent
145: \subsection{The description of Membrane and Actin Models}
146: 
147: {\bf Membrane.}~~
148: The plasma membrane of a biological cell is a highly complex surface 
149: consisting of a lipid bilayer. The complexity of the cell membrane 
150: cannot be captured in our simple model membrane. Therefore, similar to the 
151: previous successful studies \cite{Leibler89,Sambeth01b,Satya04}, the 
152: cell membrane is designed as an elastic two-dimensional ring.
153: Our model membrane is a flexible 
154: closed ring embedded on the square lattice. The ring is 
155: non-self-avoiding
156: and exibit the usual random walk characteristics.
157: Conformational changes 
158: of the ring are achieved by Monte Carlo methods, where a randomly chosen pair 
159: of two successive segments of the chain perform a kink 
160: jump or hair pin jump to one of the neighbouring sites.
161: 
162: 
163: {\bf Globular Actin (G-actin).}~
164: The model membrane encloses a fixed number, $N$, of actin molecules. 
165: Each actin molecule has a size \cite{Alberts98}
166: of about 5 nm $\times$ 5 nm
167: and is located at any of the vertices of the square lattice. 
168: The G-actin molecules  diffuse freely from one lattice point to another. 
169: No excluded volume condition is imposed among G-actin molecules.
170: But it is imposed between membrane and G-actins. So the membrane is 
171: impermeable for all actin molecules.
172: 
173: 
174: {\bf Filament Actin (F-actin).}~
175: The G-actin monomers form a rigid filament by associating with each 
176: other. Within a filament the actin  monomers are called F-actin.
177: According to experimental results \cite{Horwitz02,DeMali03a},
178: the filaments are assumed to be chemically coupled to the
179: underlying substrate via membrane proteins. 
180: This attachment to the substrate provides the necessary traction forces
181: for the advancement of the cell.
182: Hence, in our model the actin filaments are assumed to be immobile.
183: %
184: Excluded volume effects between filaments and diffusing G-actin molecule 
185: are neglected, whereas excluded volume effects among filaments and between
186: filaments and membrane are included.
187: 
188: 
189: {\bf Actin-associated Proteins.}~
190: A group of actin associated proteins help the remodelling
191: %% treadmilling 
192: of the actin network, and hence control membrane protrusion.
193: Proteins of the WASP/Scar families activate Arp2/3 protein complexes.
194: These complexes nucleate new actin filaments at the sides of existing filaments.
195: The nascent filaments in the network arise in the form of 
196: branches from preexisting filaments. 
197: The newly formed filaments elongate at their ends, pushing the 
198: membrane at the leading edge forward until they are capped by specific proteins.
199: For example,  
200: ADF/Cofilin creates free ends by severing preexisting filaments and 
201: promoting depolymerization of free filaments 
202: at their opposite ends, and 
203: CapZ and gelsolin cap the fast growing  ends.
204: 
205: Based on experimental observations it is known that Arp2/3 
206: is activated only close to the membrane. Therefore Arp2/3-induced
207: branching processes must be expected to happen only near the membrane.
208: %
209: The Arp2/3 molecule is not explicitly taken into account in our model
210: because of its large physiological concentration, in particular near the
211: cell membrane where it becomes activated. In our model, the width of the range 
212: of activation is taken to be 10 lattice sites. 
213: %
214: Accordingly, in our model 
215: we perfom branching from an existing filament
216: only if the filament extends to the range of activation near the membrane. 
217: 
218: A typical snapshot of the simulated cell 
219: is depicted in Fig.1. 
220: 
221: 
222: \begin{figure}
223: \begin{center}
224:  \subfigure[]{\includegraphics [width=0.70\textwidth,angle=0]
225: {Nandy2a.eps}}\\
226: \subfigure[]{\includegraphics [width=0.70\textwidth,angle=0]
227: {Nandy1a.eps}}
228:  \label{fig:snapshot}
229:  \caption{
230:  Two snapshot ($a$ and $b$) of the model cell. The open circles represent
231:     G-actin, the filled triangles represent F-actin molecules.     
232:     The length
233:     of the membrane is L=200, the total number of actin molecules is N=500.
234:    }
235:   \end{center}
236:   \end{figure}
237: 
238: 
239: %________________________________________________________________________
240: \noindent
241: \subsection{Probabilities and Reaction Rates}
242: 
243: \vskip0.2cm\noindent
244: {\bf Polymerization.}~
245: During the polymerization step the association of G-actin molecules 
246: to an existing F-actin filament occurs at certain rates at both ends.
247: We adopt the same reaction rates for polymerization and depolymerization
248: as used in a previous work \cite{Satya04}, and are listed in Table I.
249: %
250: Due to the structural rotational symmetry of a single G-actin molecule,
251: both ends of the F-actin filament are distinguishable : the `barbed'
252: (or synonymously, `plus' or `fast growing') end, and the `pointed'
253: (or synonymously, `minus' or `slow growing') end.
254: %
255: The association rates of ATP-bound G-actin to both ends of a filament
256: are different \cite{Bray92,Alberts98,Wegner76,Wang85}, and we denote
257: the rate constants $k^+_b$ and
258: $k^+_p$ for `barbed' and `pointed' ends, respectively. 
259: The same holds for the
260: depolymerization rates $k^-_b$ and $k^-_p$, albeit their difference
261: is smaller.
262: Since both ends of a filament are distinguishable, 
263: each F-actin filament possesses an intrinsic polarity.
264: 
265: 
266: \begin{table}\label{tab:model}
267: {\footnotesize
268: \begin{tabular}{|l|c|}
269: \hline
270:     Quantity                &  Value         \\ \hline
271:   lattice constant $a$ &  5 nm                         \\
272:   typical cell size & 100 $\times$ 100 lattice \\
273: %%  Actin density  C   & 1 $\mu$M = 0.75 molecules        \\
274:   Monte Carlo step $\tau$  &  0.875 $\mu$s  \\
275:   $W_n$                  & 0.01   \\
276:   $W_{br}^+$               & 0.1   \\
277:   $W_{br}^-$               & 0.6  \\
278:   $W_b^+$                & 1    \\
279:   $W_b^-$                & 0.0012     \\ 
280:   $W_p^+$                & 0.11   \\
281:   $W_p^-$                & 0.02    \\ \hline   
282: \end{tabular}
283: \caption{Model parameters and reaction probabilities.}  
284: }
285: \end{table}
286: 
287: 
288: 
289: 
290: 
291: \vskip0.2cm\noindent
292: {\bf Nucleation.}~
293: We consider two types of nucleation for F-actin.
294: When nucleation is formed by two G-actin molecules,  it is called `spontaneous' 
295: nucleation which occurs with probability $W_n$. When
296: the `branching' nucleation takes place then a new filament is formed
297: as a branch from the side of an existing filament.
298: The branching nucleation is modeled
299: following the `dendritic nucleaction' model based on experimental observations
300: \cite{Borisy99b,Pollard01,Pollard02}.
301: It is known 
302: that the activated protein complex Arp2/3 
303: can associate with an existing
304: filament and can  nucleate there a new filament
305: as a branch from the mother filament at an angle of about 70$^{\circ}$. 
306: This leads to the formation of a branched network. 
307: In our model on the square lattice, the branching process is implemented
308: as follows. 
309: If a G-actin is found to be on the adjacent row to an 
310: existing filament, a new daughter 
311: filament is created at this site with probability $W^{+}_{br}$. 
312: If there 
313: is already a filament in that row, then the nucleation attempt is rejected.
314: Based on the experimental facts \cite{Borisy99b,Pollard01,Pollard02}, the
315: tip of the daughter filament is a plus end, 
316: and hence its
317: direction of growth must be the same as the plus end of the
318: mother filament.
319: This implies, that the polarity of a filament, i.e., the vector connecting
320: plus and minus ends, is determined by the spontaneous nucleation
321: process and cannot be changed by dendritic nucleation.
322: In our model,
323: for the sake of simplicity, the spontaneous nucleation creates
324: new filaments only parallel to the Y-axis, pointing at random in
325: one of the two directions.
326: 
327: 
328: \vskip0.2cm\noindent
329: {\bf Reaction Rules.}~
330: Regarding the reaction rules we follow our earlier work \cite{Satya04}. In brief, 
331: the reaction rules are as follows. First we randomly choose one actin 
332: molecule. Then for the selected actin molecule, a random choice with 
333: a equal probability is made between an association or dissociation 
334: process. Then four cases may happen.
335: 1) If  for a G-actin molecule the dissiocation process is selected, the step is 
336: stopped.
337: 2) If the accociation process is selected for G-actin molecules, then reaction 
338: may happen provided the neighbour G or F-actin exist and the 
339: probability $W_n$ or $W^{+}_{br}$  or $W^{+}_p$ or $W^{+}_b$  fullfill the 
340: condition $W > \eta$, where $\eta$ is a random number $0 < \eta < 1$.
341: 3) If the association process for an F-actin is chosen, then a successful 
342: reaction may happen provided a G-actin as a neighbor exists. 
343: 4) If a dissociation process of a F-actin molecule is selected, then the 
344: process will take place with probability $W^{-}_b$ or $W^{-}_p$ or $W^{-}_{br}$.  .
345: 
346: 
347: 
348: 
349: \vskip0.2cm\noindent
350: {\bf Reaction Probabilities and Monte Carlo steps.}~
351: Since the experimental reaction rates $k^{\pm}_{b,p}$ cannot be used directly 
352: in simulations, 
353: but instead reaction probabilities $W^{\pm}_{b,p}$,
354: one must establish a relation between them.
355: In order to calculate these probabilities, 
356: we define $W^{+}_{b}$ = 1 for barbed end polymerization 
357: and calculate $W^{+}_{p}$
358: for the pointed end using the relation
359: \begin{equation}
360: \label{eq:Wk}
361: \frac{W^+_p}{W^+_b} = \frac{k^+_p}{k^+_b}~,
362: \end{equation}
363: %
364: The same can be done for the depolymerization rates $W^{-}_{b,p}$, which we
365: choose $W^-_p$=0.02 and calculate $W^-_b$ by using the corresponding relation
366: to Eq.(\ref{eq:Wk}).
367: The value of $W^-_b$ has been choosen in order to ensure that
368: the filaments are much shorter than the cell size during the simulations.
369: A basic question is how many Monte Carlo steps, $m$, for Brownian motion 
370: membrane and G-actins
371: have to be performed between two sucessive chemical rection attempts.
372: Based on the experimentally known reaction rates and the diffusion coefficient
373: of G-actins, a detailed analysis 
374: \cite{Satya04} have shown that $m$ =  5 is an appropriate value.
375: All model parameters are summarized in Table I.
376: 
377: 
378: 
379: %_____________________________________________________________________________
380: \section{Results and Discussions}
381: 
382: 
383: %__________________________________
384: \subsection{Persistent Random walk}
385: 
386: From experimental observations \cite{Bray92,Lauffenburger96,Shreiber03} and
387: from simulation studies \cite{Sambeth01a,Satya04}, it is known that
388: Arp2/3-induced dendritic nucleation is essential for the appearance of
389: a persistence random walk (PRW) of certain cells. 
390: %
391: The PRW is characterized by `stop-and-go' movements :
392: unidirectional motions with almost constant velocity
393: over distances of the
394: order of several cell diameters are separated by localized short time
395: erratic movements. After each interruption of the ballistic motion, the
396: cell continues to move in a different direction. 
397: %
398: A typical trajectory of the center of our model cell is
399: shown in Fig.\ref{fig:prw_d1} as an example from our simulations.
400: %
401: The persistency of the cell's random walk can be deduced 
402: from the time-dependent mean square displacement of the cell 
403: %
404: \begin{equation}
405:      R^2(t) = \langle [Y(t) - Y(0)]^2 \rangle
406: \end{equation}
407: %
408: which is shown in Fig.\ref{fig:msd_d2}. 
409: The dynamics at short and long times are 
410: governed by diffusion, whereas for intermediate times, 
411: the cell motion exhibits a unidirectional drift during a certain
412: persistence time.
413: %
414: The unidirectional motion has been explained as a consequence of 
415: the autocatalytic nature \cite{Carlier00,Sambeth01a} of 
416: the dendritic polymerization kinetics. 
417: The branching processes induce
418: an explosive growth of filaments in a certain direction which 
419: leads to membrane protrusions and concomitant cell movements in the
420: same direction.
421: %
422: %______________________________________________
423: %\noindent
424: \begin{figure}[h] %%[b!]
425:   \begin{center}
426:     %\hspace*{-0.4cm}
427:     \includegraphics[width=0.70\textwidth,angle=0]
428: {Nandy2.eps}    
429:     \caption{The trajectory $Y(t)$ (full line, left scale) 
430:     and polarity $P(t)$ (dotted line, right scale)
431:     of a cell exhibiting a persistent random walk, as function
432:     of time $t$ in units of Monte Carlo steps, MCS. } 
433:     \label{fig:prw_d1}
434:   \end{center}
435: \end{figure}
436: %______________________________________________
437: %
438: %______________________________________________
439: %\noindent
440: \begin{figure}[h] %%[b!]
441:   \begin{center}
442:     %\hspace*{-0.4cm}
443:     \includegraphics[width=0.75\textwidth, angle=0]
444: {Nandy3.eps}    
445:     \caption{Typical mean square displacement of a cell 
446:     exhibiting a persistent random walk.} 
447:     \label{fig:msd_d2}
448:   \end{center}
449: \end{figure}
450: %______________________________________________
451: %
452: This motion lasts only for a certain persistence time, then
453: the direction of motion changes. 
454: 
455: In the following subsection we discuss 
456: the reasons for this dynamic instability.
457: It should be noted that insights to the physical basis of the cell's
458: persistency 
459: would support, in particular, our
460: understanding of the physical basis of chemotaxis, i.e., the prolongation
461: of the persistency of the cell's motion by an external signal. 
462: 
463: 
464: 
465: 
466: 
467: 
468: 
469: 
470: 
471: %_______________________________
472: \subsection{Dynamic Instability}
473: 
474: The fundamental quantity which governs the migration of the cell is its
475: polarity and its mechanism of self-perpetuation. 
476: %
477: The cell polarity is 
478: macroscopically and dynamically characterized by the
479: formation of leading and trailing membrane edges 
480: which are a consequence of the cell's coupling to the polarity of the 
481: enclosed actin cytoskeleton. 
482: The polarity of the cytoskeleton itself is the average orientation
483: of its filaments, where each actin filament possesses a polarity due to
484: its plus and minus ends.
485: %
486: This spontaneous internal polarity of the actin network
487: determines the direction of cell motion
488: and is maintained without external signals (e.g. chemotaxis) only for
489: a certain time (persistent random walk).
490: %
491: Specifically, the formation and regulation of cell polarity is achieved by
492: a complex protein signaling network with positive and negative feedback
493: loops. 
494: We do not discuss the details here which are 
495: beyond the scope of this paper. We note that different signals converge,
496: among others, on
497: the activation of an Arp2/3 protein complex, which leads to branching and 
498: the autocatalytic polymerization.
499: 
500: In our model, we define the polarity of the cell by the polarity of the total
501: actin network, i.e., 
502: by the difference 
503: in the number of filaments pointing in opposite directions,
504: \begin{equation}
505:      P = \frac{(F_{up}-F_{dn})}{F_{total}}     %%\nonumber
506: \end{equation}
507: where $F_{up}$ is the number of filaments pointing in 
508: positive Y-direction, $F_{dn}$ in negative Y-direction, and
509: $F_{total}$ is the total number of filaments.
510: The range of the polarity parameter is  $-1\le P \le 1$.
511: %
512: The trajectory of this quantity is shown
513: in Fig. \ref{fig:prw_d1} (dotted line, right scale). 
514: From this result one observes  
515: that at certain times, where $P(t) = 0$, 
516: the cell changes its direction of motion to the opposite as
517: indicated by the corresponding displacements $Y(t)$. 
518: 
519: After the analysis of several simulations under
520: different conditions of the model cell, it must be concluded that the  
521: cause behind this phenomena is the existence of two antagonistic processes, 
522: the spontaneous and the dendritic nucleation processes.
523: 
524: During a spontaneous
525: nucleation process, where two G-actin monomers form a new
526: F-actin filament, the orientation of the new filament is determined at random.
527: %
528: This is in contrast to the dendritic
529: nucleation process, where the daughter
530: filament branches off the mother filament and therefore
531: always adopts the same growth direction as the mother filament.
532: %
533: Hence, spontaneous nucleation changes the sign of polarity, whereas
534: dendritic nucleation increases the absolute value of polarity.
535: The first leads to random motion, whereas the latter prolongs 
536: unidirectional motion.
537: %
538: Since the polarity fluctuates in time, the spontaneous nucleation
539: may eventually dominate and causes a change of polarity, and hence
540: a change of vectorial motion.
541: 
542: This view is corroborated by simulations of various limiting cases, which are
543: shown in Fig.\ref{fig:msd}.
544: %______________________________________________
545: %\noindent
546: \begin{figure}[h] %%[b!]
547:   \begin{center}
548:     %\hspace*{-0.4cm}
549:     \includegraphics[width=0.76\textwidth, angle=0]
550: {Nandy4.eps}    
551:     \caption{Mean square displacements of a cell under different 
552:     conditions and at different probablilities of spontaneous nucleation, $wn$.
553:      `s' represents spontaneous nucleation, `s+d' represents both 
554:      spontaneous and dendritic nucleation, and `d' represents  dendritic
555:      nucleation only.}  
556:     \label{fig:msd}
557:   \end{center}
558: \end{figure}
559: %______________________________________________
560: %
561: 
562: In one limiting case, the cell motion is caused solely by spontaneous
563: nucleations.
564: The corresponding curve 
565: is shown in Fig.\ref{fig:msd} by the full lines with label `s'.
566: At two different nucleation probabilities, $W_n$, the mean square
567: displacements indicate
568: ordinary diffusion, $R^2(t)/t \approx const.$, at almost all times.
569: This is expected due to the random orientation of newly established
570: filaments. A very weak persistency may be expected at small nucleation rates.
571: 
572: 
573: 
574: In the other limiting case, the cell motion is caused solely by dendritic 
575: nucleation. Since in this case the orientation of the filaments is preserved,
576: the corresponding curve (dotted line with label `d') has to
577: exhibit a clear drift, $R^2(t)/t \sim t$, which takes place  for
578: times larger than a characteristic time $\tau_d \approx$ 7 $\times$ 10$^3$. 
579: %
580: The characteristic time scale $\tau_d$ separates periodes of cell 
581: advancements and localized cell displacements (`resting').
582: %
583: Since the activation range of Arp2/3 extends only a few lattice sites from
584: the membrane, the rapid advancement of the leading edge of the membrane  
585: together with the rapid autocatalytic processes of branching near
586: the leading edge deprive this area of available G-actin molecules which are
587: necessary for continuation of these processes.
588: Therefore, actin network and cell remain temporarily, during $t < \tau_d$,
589: essentially at rest.
590: %
591: While recruiting G-actins by diffusion
592: from the minus to the plus ends of the filaments near the activation zone, 
593: the membrane of the cell performes random displacements 
594: around the immobile filament network, which leads to time-independent
595: displacements of the cell, $R^2(t)/t \sim 1/t$ at $t < \tau_d$. 
596: %
597: It is still unknown whether this result from simulations
598: is related to the experimentally observed \cite{Vicker00}
599: excitation waves of F-actin assembly near the membrane and the correlated
600: cell advancements. 
601: 
602: The interesting intermediate case, when a few events of spontaneous nucleation 
603: interfere with dendritic nucleations, is shown in Fig.\ref{fig:msd}
604: by the broken line with label `d+s'.
605: At a nucleation ratio of $W_n/W_{br}^+ = $0.015, 
606: again a drift of the cell is observed albeit only up to
607: a typical persistence time $\tau_p \approx 10^6$.
608: At larger times, ordinary diffusion $R^2(t)/t = const.$ is observed
609: indicating the dominance of spontaneous nucleation processes.
610: For $t < \tau_p$, the mean square displacements are qualitative the same as
611: for the dendritic case `d'.
612: 
613: %_____________________________________________________________________________
614: \section{Summary and Conclusions}
615: 
616: 
617: The dynamics of a motile cell with unidirectional and bidirectional branching
618: is studied with treadmilling actin polymerization kinetics. The filaments 
619: growing on both sides with treadmilling rates and branching in both directions
620: leads to a persistent random walk whose 
621: persistency becomes infinite and leads to unidirectional drift
622: in the case of one side branching.
623: The dynamics are governed by  
624: diffusion at short times followed by drift at longer times. 
625: 
626: The phenomena is 
627: shown to be attributed to two antagonistic nucleation processes during
628: the polymerization of the cell's actin cytoskeleton : the (ordinary) spontaneous
629: nucleation  and the dendritic nucleation  processes.
630: Whereas spontaneous nucleations generate actin filaments growing in different 
631: directions and hence create motions in random directions, 
632: dendritic nucleations provides a unidirectional growth.
633: Since dendritic growth exhibits stochastic fluctuations,
634: spontaneous nucleations may eventually compete or even dominate, 
635: which results in a reorientation of filament growth and hence 
636: a new direction of cell motion.
637: The event of reorientation takes place at instants of vanishing
638: polarity of the actin skeleton.
639: 
640: 
641: %_______________________________________________________________________
642: \section*{ACKNOWLEDGEMENTS}
643: B. Nandy would like to thank IBM Research for providing a doctoral
644: fellowship and
645: the Institut f\"ur Festk\"orperforschung
646: at the Forschungzentrum J\"ulich for hospitality.
647: 
648: 
649: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
650: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
651: %%%%%%%%%%%%%%%%%                           %%%%%%%%%%%%%%%% 
652: %%%%%%%%%%%%%%%%%       REFERENCES          %%%%%%%%%%%%%%%%
653: %%%%%%%%%%%%%%%%%                           %%%%%%%%%%%%%%%%
654: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
655: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
656: 
657: 
658: \bibliographystyle{phaip}
659: \bibliography{Nandy} 
660:  
661: 
662: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
663: \end{document}
664: \bye
665: